Health Policy A custom publication of the Disease Control Priorities Project ©2006 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved The findings, interpretations, and conclusions expressed herein do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent, The World Health Organization, or the Fogarty International Center, U.S. National Institutes of Health. The World Bank, the World Health Organization, and the Fogarty International Center, U.S. National Institutes of Health do not guarantee the accuracy of the data included herein. The boundaries, colors, denominations, and other information shown on any map do not imply any judgment on the part of The World Bank, the World Health Organization, or the Fogarty International Center, U.S. National Institutes of Health concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is copyrighted. Copying and/or transmitting portions or all of this work without permission other than for non-commercial, educational and scholarly purposes may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to reproduce, photocopy or reprint other than for non-commercial, educational and scholarly purposes, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-7508400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. The chapters in this publication were originally published in the following two books, and any citations should include the complete information provided: 1) Dean T. Jamison, Joel G. Breman, Anthony R. Measham, George Alleyne, Mariam Claeson, David B. Evans, Prabhat Jha, Anne Mills, and Philip Musgrove, eds. 2006. Disease Control Priorities in Developing Countries, 2nd ed. New York: Oxford University Press. 2) Alan D. Lopez, Colin D. Mathers, Majid Ezzati, Dean T. Jamison, and Christopher J. L. Murray, eds. 2006. Global Burden of Disease and Risk Factors. New York: Oxford University Press. Contents 1 Millennium Development Goals for Health: What Will It Take to Accelerate Progress? ............................................... 1 Disease Control Priorities in Developing Countries Adam Wagstaff, Mariam Claeson, Robert M. Hecht, Pablo Gottret, and Qiu Fang 2 Fiscal Policies for Health Promotion and Disease Prevention ......................................................................................... 15 Disease Control Priorities in Developing Countries Rachel Nugent and Felicia Knaul 3 Indoor Air Pollution ......................................................................................................................................................... 29 Disease Control Priorities in Developing Countries Nigel Bruce, Eva Rehfuess, Sumi Mehta, Guy Hutton, and Kirk Smith 4 Prevention of Chronic Disease by Means of Diet and Lifestyle Changes ....................................................................... 53 Disease Control Priorities in Developing Countries Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, Courtenay Dusenbury, Pekka Puska, and Thomas A. Gaziano 5 Tobacco Addiction ........................................................................................................................................................... 71 Disease Control Priorities in Developing Countries Prabhat Jha, Frank J. Chaloupka, James Moore, Vendhan Gajalakshmi, Prakash C. Gupta, Richard Peck, Samira Asma, and Witold Zatonski 6 Alcohol ............................................................................................................................................................................. 89 Disease Control Priorities in Developing Countries Jürgen Rehm, Dan Chisholm, Robin Room, and Alan Lopez 7 Illicit Opiate Abuse .......................................................................................................................................................... 109 Disease Control Priorities in Developing Countries Wayne Hall, Chris Doran, Louisa Degenhardt, and Donald Shepard 8 Public Health Surveillance: A Tool for Targeting and Monitoring Interventions ........................................................... 135 Disease Control Priorities in Developing Countries Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others 9 Information to Improve Decision Making for Health ...................................................................................................... 157 Disease Control Priorities in Developing Countries Sally K. Stansfield, Julia Walsh, Ndola Prata, and others 10 School-Based Health and Nutrition Programs ................................................................................................................. 171 Disease Control Priorities in Developing Countries Donald Bundy, Sheldon Shaeffer, Matthew Jukes, and others 11 Natural Disaster Mitigation and Relief ............................................................................................................................ 189 Disease Control Priorities in Developing Countries Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio 12 Control and Eradication ................................................................................................................................................... 205 Disease Control Priorities in Developing Countries Mark Miller, Scott Barrett, and D. A. Henderson 13 Improving the Quality of Care in Developing Countries ................................................................................................. 219 Disease Control Priorities in Developing Countries John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others 14 Health Workers: Building and Motivating the Workforce .............................................................................................. 235 Disease Control Priorities in Developing Countries Charles Hongoro and Charles Normand 15 Strategic Management of Clinical Services ..................................................................................................................... 249 Disease Control Priorities in Developing Countries Alexander S. Preker, Martin McKee, Andrew Mitchell, and others 16 Measuring the Global Burden of Disease and Risk Factors,19902001 ............................................................................ Global Burden of Disease and Risk Factors Alan D. Lopez, Colin D. Mathers, Majid Ezzati, Dean T. Jamison, and Christopher J. L. Murray 263 Chapter 9 Millennium Development Goals for Health: What Will It Take to Accelerate Progress? Adam Wagstaff, Mariam Claeson, Robert M. Hecht, Pablo Gottret, and Qiu Fang The scale of the diseases and conditions that the Millennium Development Goals (MDGs) address is staggering: • Almost 11 million children died before their fifth birthday in 2000 (UNICEF 2001). Less than 1 percent of these 11 million deaths (79,000) occurred in high-income countries, compared with 42 percent in Sub-Saharan Africa, 35 percent in South Asia, and 13 percent in East Asia. • In 1998, an estimated 843 million people were considered undernourished on the basis of their food intake (FAO 2000). Of the estimated 140 million children under the age of five who were underweight, almost half (65 million) were in South Asia. • Of the 3.1 million people who died from HIV/AIDS in 2003, almost all (99 percent) were in the developing world—74 percent in Sub-Saharan Africa alone (UNAIDS 2004). Tuberculosis and malaria together killed an equal number; most of these deaths were among the poor. • In 1995, 515,000 women died during pregnancy or childbirth: 1,000 in the industrial world, contrasted with 252,000 in Sub-Saharan Africa (UNICEF 2001). This burden of death and suffering is heavily concentrated in the world’s poorest countries (Wagstaff and Claeson 2004). Death and disease matter in their own right, but they also act as a brake on poverty reduction. Nobel laureate Amartya Sen (2002) has described health as one of “the most important conditions of human life and a critically significant constituent of human capabilities which we have reason to value.” Health also matters because it influences the living standards of both households and countries. Health expenses can easily become burdensome for households. In Vietnam, they are estimated to have pushed 3 million people into poverty in 1993 (Wagstaff and van Doorslaer 2003). Beyond the direct impact of ill health on households’ living standards through out-of-pocket expenditures, it indirectly affects labor income through productivity and the number of hours that people can work. The effects of illness on income, which may take time to appear, are often long lasting. Malnourished children are less likely to attend school and less likely to learn when they do attend, reducing their productivity in later life. The devastating economic consequences of illness and death are evident at the macroeconomic level as well. The AIDS epidemic alone has been estimated to reduce rates of economic growth by 0.3 to 1.5 percentage points annually (Bell, Devarajan, and Gersbach 2003). In the 1990s, the international community recognized the importance of health in development. In a period when overall official development assistance declined, development assistance to health rose in real terms. World Bank lending for health increased, with a doubling of the share of International Development Association disbursements going to health (OECD Development Assistance Committee 2000). The 1990s saw an increased global concern over the debt in the developing world, fueled in part by a perception that interest payments were constraining government health expenditures in developing countries. The enhanced Highly Indebted Poor Country Initiative, spearheaded by the International Monetary Fund and World Bank in response to the unsustainable debt burden of the poorest countries, was explicitly geared to channel freed resources into the health and other social sectors. The 181 ©2006 The International Bank for Reconstruction and Development / The World Bank 1 Poverty Reduction Strategy Papers submitted by governments of developing countries seek debt relief or concessional (low-interest) International Development Association loans to set out their plans for fighting poverty on all fronts, including health. The 1990s also saw the development of major new global health initiatives and partnerships, including the Joint United Nations Programme on HIV/AIDS (UNAIDS); the Global Alliance for Vaccines and Immunization; the Stop TB Partnership; the Roll Back Malaria Partnership; the Global Fund to Fight AIDS, Tuberculosis, and Malaria; and the Global Alliance for Improved Nutrition. A range of new not-for-profit organizations were set up to spur the accelerated discovery and uptake in developing countries of low-cost health technologies to address the diseases of the poor; these organizations included the International AIDS Vaccine Initiative, the Medicines for Malaria Venture, the Global Alliance for Tuberculosis, and the International Trachoma Initiative. In addition, the scale of philanthropic involvement in international health increased, with the launch of the Bill & Melinda Gates Foundation and the Packard Foundation and the continued attention to global health issues by such established entities as the Rockefeller Foundation. These initiatives brought not only new resources—funds, ideas, energy, and mechanisms— but also new challenges to harmonization in the attempt to coordinate and link global goals with local actions in the fight against disease, death, and malnutrition in the developing world. As the 1990s closed, the international community decided that even more needed to be done. At the United Nations Millennium Summit in September 2001, heads of 147 states endorsed the MDGs, nearly half of which concern different aspects of health—directly or indirectly (box 9.1). Several other goals are indirectly related to health—for example, the goals on education and gender. Gender equality is considered important to promoting good health among children. Other health outcomes than those included in the MDGs measure progress on health—for example, targets related to noncommunicable diseases. These targets are referred to as the MDG plus and are included in national priority setting, especially in many middle-income countries. THE MILLENNIUM DEVELOPMENT GOALS FOR HEALTH: PROGRESS AND PROSPECTS Of the MDGs for which trend data are available or estimated, the fastest progress has been on malnutrition, whereas overall progress on under-five mortality and maternal mortality has been slower. A Mixed Score at Halftime In-depth analysis of the health-related MDGs shows a mixed score at halftime (Wagstaff and Claeson 2004): • The number of people living in on-track countries (countries that will reach the MDGs if they maintain the rate of progress they have already achieved during the period from 1990 to the present) matters. For the malnutrition target, 77 percent of the developing world’s people live in an ontrack country, but in Sub-Saharan Africa only 15 percent of the people live in an on-track country. • Different indicators show different levels of improvement. For under-five mortality, the developing world was reduced by an average of only a 2.5 percent in the 1990s, well short of the target of 4.3 percent. Box 9.1 The Health-Related Millennium Development Goals Goal 1—eradicating extreme poverty and hunger. This goal includes as a target the halving between 1990 and 2015 of the proportion of people who suffer from hunger, with progress to be measured in terms of the prevalence of underweight children under five years of age. The target implies an average annual rate of reduction of 2.7 percent. Goal 4—reducing child mortality. The target is to reduce by two-thirds between 1990 and 2015 the under-five mortality rate, equivalent to an annual rate of reduction of 4.3 percent. Goal 5—improving maternal health. The target is to reduce by three-quarters between 1990 and 2015 the maternal mortality ratio, equivalent to an annual rate of reduction of 5.4 percent. Goal 6—combating HIV/AIDS, malaria, and other diseases. The target is to halt and begin to reverse the spread of these diseases by 2015. Goal 7—ensuring environmental sustainability. This goal includes as a target the halving by 2015 of the proportion of people without sustainable access to safe drinking water. Goal 8—developing a global partnership for development. This goal includes as a target the provision of access to affordable essential drugs in developing countries. Source: United Nations Millennium Declaration, the United Nations Millennium Summit 2000. 182 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 2 • Regional differences are also pronounced, with Sub-Saharan Africa faring worse than other regions. In Africa, trends in reducing under-five mortality and underweight in children were barely above zero during the 1990s, and maternal mortality fell on average by just 1.6 percent a year compared with the annual target rate of 5.4 percent. • Evidence on how the poor are faring within countries is mixed. For malnutrition, the poorest 20 percent of the population within countries appears, on average, to have been experiencing broadly similar rates of reduction to the population as a whole. However, for under-five mortality, the rate has been falling more slowly among the poor, while better-off families are seeing faster rates of progress. Will the Second Half Go Better? As a comparison of the child mortality experiences in the 1980s and 1990s demonstrates, past performance is not necessarily a good predictor of future performance. The fact that a country is on track on the basis of its performance in the 1990s does not guarantee that it will maintain the required annual rate of reduction of malnutrition or mortality during the second half of the MDG “window” from 2000 to 2015. Countries currently off track may possibly get on track in the second half if they can combine good policies with expanded funding for programs that address both the direct and the underlying determinants of the health-related goals. Stimuli External to the Health Sector. The World Bank estimates that economic growth will fall somewhat in East Asia and the Pacific in 2000–15, turn from negative to positive in Europe and Central Asia as well as Sub-Saharan Africa, and increase somewhat in Latin America and the Caribbean, the Middle East and North Africa, and South Asia (Jones and others 2003). Primary education completion rates will probably grow faster in the new millennium as a result of the global education initiatives and partnerships on Education for All and the Fast-Track Initiative. However, higher rates of educational attainment among women of childbearing age will not be achieved until 2005 or so, and even then the first full round of effects on under-five mortality will not be felt until 2010. More relevant is the fact that gender gaps in secondary education may well narrow faster in the new millennium than in the 1990s as a result of the gender MDG (Goal 3: Eliminate gender disparity in primary and secondary education by 2005 and in all levels of education no later than 2015). To achieve parity with boys by 2015 in the proportion of the population who are age 15 and have completed secondary education, girls will have to achieve a faster growth in completion rates in the new millennium than in the 1990s in most regions, especially in South Asia and in East Asia and the Pacific. If the water MDG (ensuring that households have access to safe drinking water) is to be reached, access rates will need to grow much faster in 2000–15, especially in Sub-Saharan Africa (Wagstaff and Claeson 2004). Gender equality in school and access to clean water will have a positive effect on progress toward the health MDGs. Even with economic growth and faster progress on these nonhealth goals, however, many regions will still miss many of the health targets. The picture is bleakest for underfive mortality—and for Sub-Saharan Africa. The Goals Matter for All Countries. These goals need to be taken seriously for three main reasons: • Faster progress is important even if targets are missed. A key message of this chapter is that progress can be accelerated in all countries through a judicious mix of spending and policy and institutional reform. • The goals facilitate benchmarking and monitoring of results. Because the goals focus on a limited set of outcomes, monitoring and evaluating progress toward the MDGs can show what is achievable and where faster progress can be made. • Focusing attention on national progress, as measured by distributional analysis of the MDGs, forces countries to consider how the benefits of progress are distributed among the rich and poor within each country—the poor risk being left behind even in countries making progress overall. One limitation of the MDGs and targets is that they are national averages. However, distributional analysis of MDG trends (Wagstaff and Claeson 2004) reminds us that progress needs to be for everyone, not just the better off. Progress has been uneven, with the poorer countries lagging behind the rest, and for under-five mortality, the poor within countries are lagging behind the rest of the population. SCALING UP: DEFINING INTERVENTIONS AND REMOVING CONSTRAINTS A lack of interventions is not the primary obstacle to faster progress toward the goals, although new interventions that can be delivered by weak health systems could greatly improve progress—for example, malaria or HIV vaccines and effective vaginal microbicides to block the spread of HIV and other sexually transmitted infections. The main obstacle is the low levels of use—especially among the poor—of existing effective interventions. For example, if use of all the proven effective preventive and treatment interventions for childhood illness were to rise from current levels to reach all, the number of under-five deaths worldwide could fall by as much as 63 percent (World Bank 2003b). Array of Interventions, Programs, and Service Modalities The available interventions constitute a powerful arsenal for preventing and treating the main causes of malnutrition and death (table 9.1).1 The major diseases and conditions that the MDGs aim to prevent and control are discussed in several Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 183 ©2006 The International Bank for Reconstruction and Development / The World Bank 3 Table 9.1 Effective Interventions to Reduce Illness, Deaths, and Malnutrition MDG Preventive interventions Treatment interventions Child mortality Breastfeeding; hand washing; safe disposal of stool; latrine use; safe preparation of weaning foods; use of insecticide-treated bednets; complementary feeding; immunization; micronutrient supplementation (zinc and vitamin A); prenatal care, including steroids and tetanus toxoid; antimalarial intermittent preventive treatment in pregnancy; newborn temperature management; nevirapine and replacement feeding; antibiotics for premature rupture of membranes; clean delivery Case management with oral rehydration therapy for diarrhea; antibiotics for dysentery, pneumonia, and sepsis; antimalarials for malaria; newborn resuscitation; breastfeeding; complementary feeding during illness; micronutrient supplementation (zinc and vitamin A) Maternal mortality Family planning (lifetime risk); intermittent malaria prophylaxis; use of insecticide-treated bednets; micronutrient supplementation (iron, folic acid, calcium for those who are deficient) Antibiotics for preterm rupture of membranes, skilled attendants (especially active management of third stage of labor), basic and emergency obstetric care Nutrition Exclusive breastfeeding for 6 months, appropriate complementary child feeding for next 6–24 months, iron and folic acid supplementation for children, improved hygiene and sanitation, improved dietary intake of pregnant and lactating women, micronutrient supplementation for prevention of anemia and vitamin A deficiency for mothers and children, anthelmintic treatment in school-age children Appropriate feeding of sick child and oral rehydration therapy, control and timely treatment of infectious and parasitic diseases, treatment and monitoring of severely malnourished children, high-dose treatment of clinical signs of vitamin A deficiency HIV/AIDS Safe sex, including condom use; unused needles for drug users; treatment of sexually transmitted infections; safe, screened blood supplies; antiretrovirals in pregnancy to prevent maternal to child transmission and after occupational exposure Treatment of opportunistic infections, co-trimoxazole prophylaxis, highly active antiretroviral therapy, palliative care Tuberculosis Directly observed treatment of infectious cases to prevent transmission and emergence of drug-resistant strains and treatment of contacts, Bacillus Calmette-Guérin immunization Directly observed treatment to cure, including early identification of tuberculosis symptomatic cases Malaria Use of insecticide-treated bednets, indoor residual spraying (in epidemic-prone areas), intermittent presumptive treatment of pregnant women Rapid detection and early treatment of uncomplicated cases, treatment of complicated cases (such as cerebral malaria and severe anemia) Source: Authors. chapters (for example, see chapters 15, 19, 21–24, 28–31, 44, and 45). The most cost-effective interventions and programs are also discussed in several chapters (see chapters 59–62 and 65). The chapters dealing with health systems and service delivery issues and other constraints related to the health MDGs are found in the latter part of the book (see chapters 66–68, and 73). In the case of child mortality, for example, diarrheal diseases, pneumonia, and malaria account for 52 percent of deaths worldwide (World Bank 2003b). For each of these major causes of childhood mortality, at least one proven effective preventive intervention and at least one proven effective treatment intervention exist, capable of being delivered in a low-income setting. In most cases, several proven effective interventions exist. For diarrhea—the second-leading cause of child deaths—no fewer than five proven preventive interventions and three proven treatment interventions are available. Effective Interventions Reaching Too Few People The high rates of malnutrition and death in the developing world have several causes. First, people do not receive the effec- tive interventions that could save their lives or make them well nourished. In middle- and high-income countries, 90 percent of children are fully immunized, more than 90 percent of deliveries are assisted by a medically trained provider (that is, a doctor, nurse, or trained midwife, excluding traditional birth attendants), and more than 90 percent of pregnant women have at least one prenatal visit (UNICEF 2001). In South Asia, fewer than 50 percent of pregnant women receive a prenatal checkup, and only 20 percent of deliveries are assisted by a trained provider. The story is similar for other childhood interventions— and for interventions for other goals. Condom use to prevent transmission of HIV is low in much of Sub-Saharan Africa and South Asia, and inexpensive one-time treatment with antiretroviral medicine to prevent transmission from mother to child covers only a small fraction of at-risk pregnant women in most of the developing world. In Asia, where more than 7 million people are living with HIV/AIDS, no country has yet exceeded 5 percent antiretroviral therapy coverage among those who could benefit from it (World Bank 2003c). 184 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 4 Just as shortfalls in coverage vary across countries, so do they vary within countries, with the poor and other deprived groups consistently lagging. These groups are less likely to receive full basic immunization coverage, to have their deliveries attended by a trained provider, and to have at least one prenatal care visit to a medically trained provider. On the positive side, the poor are often making fastest progress in coverage, reflecting in part that the better off already have high coverage rates for many interventions. Underuse of Effective Interventions Costs Lives The low use of effective interventions—in the developing world in general and among the poor in particular—translates into rates of mortality, morbidity, and malnutrition that are far higher than necessary. If use of all the proven effective childhood preventive and treatment interventions, for example, were to rise from their current levels to 99 percent—95 percent for breastfeeding—the number of under-five deaths worldwide could fall by as much as 63 percent (Jones and others 2003). Deaths from malaria and measles could be all but eliminated, and deaths from diarrhea, pneumonia, and HIV/AIDS could be reduced dramatically. If coverage rates of the key maternal mortality interventions were increased from current levels to 99 percent, an estimated 391,000 maternal deaths worldwide (74 percent of current maternal deaths) might be averted (Ramana 2003). One intervention stands out as especially important: access to essential obstetric care, which accounts for more than half the maternal deaths averted. allocations. In practice, however, the amount of extra spending required would be difficult to attain on present trends and would even be prohibitively expensive. In the case of East Asia and the Pacific, for example, if economic growth proceeds as expected and the other relevant Millennium Development Targets are attained, the region would achieve the required rates of reduction of underweight and maternal mortality— assuming that economic growth is accompanied by the development of appropriate human resources for health—even if government health spending continues to grow at its current rate. However, the region would miss the under-five mortality target. To reach that target, a minimum of 5 percentage points would need to be added to the annual rate of growth of the government health share of gross domestic product (GDP). That would take the projected share of GDP spent on government health programs to 3.7 percent in 2015—more than twice what it would be if the 1990s pattern of growth continued (Wagstaff and Claeson 2004). In Sub-Saharan Africa, the situation is even starker. Even if faster economic growth materializes and the other targets are achieved, the share of government health spending in GDP would need to grow nearly sixfold over the coming decade, taking the share to 12.2 percent of GDP in 2015. This percentage compares with a 2000 figure of 1.8 percent and a 2015 forecast of 2.2 percent based on the 1990s annual growth in government spending for health. In conclusion, African countries will not be able to reach the MDGs simply by multiplying their health spending along the lines of historical expenditure patterns, because the multiples required are beyond any realistic expectation of what these governments will be able to do during the next 10 years. WHAT DO COUNTRIES NEED TO DO? If the lack of interventions is not holding countries back from achieving the goals, what is? What do countries need to do to make progress toward the MDGs? In countries with good governance, additional government health spending does reduce child mortality (Rajkumar and Swaroop 2002). Development assistance has a stronger effect in countries with strong policies and institutions than in countries with only average-quality policies and institutions—and an insignificant effect in countries where policies and institutions are weak. This assertion is also consistent with the findings of a study undertaken by the World Bank for the MDG report, The Millennium Development Goals for Health: Rising to the Challenges (Wagstaff and Claeson 2004). The study includes other outcomes with child mortality and uses the World Bank’s Country Policy and Institutional Assessment index to measure the quality of policies and institutions. In principle, well-governed countries with good policies and institutions could achieve the goals simply by scaling up their expenditures on existing programs in proportion to current What Are the Implications? Poorly governed countries cannot expect to make much progress toward the MDGs simply by scaling up their expenditures on existing programs in proportion to current allocations. Although well-governed countries could, in principle, simply scale up existing spending to reach the targets, this option is unlikely to be affordable for them or their donors. This situation has two implications: • First, targeting additional government spending to activities that will have the largest effect on the MDGs is important for both sets of countries. • Second, building good policies and institutions is important for all countries: doing so increases the productivity not just of additional spending but also of existing spending commitments. What do better policies and institutions entail in the health sector? Health systems are very broad, and weak policies and institutions can arise at several points along the pathway, from government health spending to health Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 185 ©2006 The International Bank for Reconstruction and Development / The World Bank 5 outcomes (Claeson and others 2001). Countries can do a number of things, with help from donors, to build stronger policies and institutions. Improving Expenditure Allocations and Targeting In most countries, government spending gets stuck in the cities and disproportionately accrues—in a financial sense—to people who are better off. Geographic Targeting. Resource allocation formulas can be used to reduce government spending gaps across regions and ideally to favor geographic zones that are furthest behind. These formulas have been used, for example, as part of Bolivia’s decentralization efforts since 1994 and have been associated with some large—and pro-poor—improvements in maternal and child health indicators. Targeting resources to poor regions and provinces may be most effectively implemented through nontraditional mechanisms for priority setting and implementation, such as social investment funds. In Bolivia, a recent impact evaluation concluded that such funds were responsible for a decline in under-five mortality from 88.5 to 65.6 per 1,000 live births over a five-year period (Newman and others 2002). Changing the Allocation of Spending across Care Levels. Spending on health in developing countries is characterized by a high concentration of spending on secondary and tertiary infrastructure and personnel. Some governments have tried to scale back the share of hospital spending. Tanzania, for example, reduced the share of hospital spending from 60 percent in 2000 to 43 percent in 2002. Chapter 3 deals with the issue of how to couple expenditure reallocations across levels of care with measures to improve performance at each level of the health care system. Targeting Specific Programs. Programs such as those delivering directly observed treatment short course (DOTS) for tuberculosis or integrated management of infant and childhood illness (IMCI) for child health are good examples of programs that may yield high returns to government spending at the margin. A recent World Bank study in India provides further support for the idea that the way government spending is allocated across programs makes a difference to its effect on the Millennium Development Indicators (World Bank 2003a). Successful public health programs—large-scale programs with a measurable health effect over at least a five-year period—are further discussed in chapter 8. All successful programs have several factors in common: technical innovation and stakeholder consensus, strong political leadership, coordination across agencies and management, effective use of information and financial resources, and participation of the beneficiary community. Targeting Specific Population Groups. Many countries subsidize all government health services for everyone. These blanket subsidy schemes not only fail to target interventions that give rise to externalities but also fail to disproportionately benefit the poor—despite the stronger equity case for subsidizing their care and the fact that they tend to bear a disproportionate burden of malnutrition as well as child and maternal mortality. There are many proven ways to target the poor—for example, by delivering essential services in clinics or health posts that only poor families attend or by promoting and delivering services in a way that segments the market and appeals to those in low-income households. Targeting Spending to Remove Bottlenecks. A planning and budgeting approach is to assess—for a country—the health sector impediments to faster progress, to identify ways of removing them, and to estimate both the costs of removing them and the likely effects of their removal on MDG outcomes (Soucat and others 2002). MDG analysis along these lines— referred to sometimes as marginal budgeting for bottlenecks (MBB)—has begun in several African countries and in some states of India (UNICEF and World Bank 2003). In Mali, key bottlenecks were identified for supporting home-based practices and delivering periodic and continual professional care. They included low access to affordable commodities and the need for community-based support for home-based care; low geographical access to preventive professional care (immunization, vitamin A supplementation, and prenatal care); shortages of qualified nurses and midwives; and an absence of effective third-party payment mechanisms for the poor for professional continuous care. Important health systems bottlenecks, such as human resources, drug availability, and health care management, are discussed in chapters 71–73. Improving Policies toward Households as Producers and Demanders of Care Households are at the center of any efforts to scale up; they not only demand and consume care, but they are also important producers of prevention and care. Policies to increase coverage of cost-effective interventions to reach the health MDGs, therefore, need to identify and influence the key constraints to both the production and the demand for those services at the household and community levels. Lowering Financial Barriers. Low income is a barrier to the use of most health interventions, and economic growth is an important weapon in the war against malnutrition and mortality. However, social protection programs are also important. Successful schemes aimed at households and communities are discussed in chapter 56. 186 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 6 One part of the affordability equation is price. User charges for MDG interventions are to be discouraged. Why? Many of those interventions involve benefits that spill over to people who do not receive the intervention; high coverage of immunization is a classic example. However, an equity case also can be made for reducing prices facing the poor and near poor, even where no spillovers occur. Subsidies should be targeted to services with spillovers and to the poor. In practice, subsidies are often badly targeted in at least one respect if not both. Exceptions exist. In Ifakara, Tanzania, a voucher program for mosquito nets was launched successfully for pregnant women and children under five (Schellenberg and others 2001). Some recent programs, especially in Latin America, have not only made health care affordable for the poor but have also made it profitable. Rather than simply reducing the cost of using specific interventions, these programs provide users with cash payments, which are linked to specific interventions and restricted to certain groups—often poor mothers and their children. The experience with these programs in targeting and achieving results is encouraging (Mesoamerica Nutrition Program Targeting Study Group 2002; Morris and others 2003; Palmer and others 2004). Risk aversion coupled with the unpredictability of illness provides a motivation for pooling risks through an insurance scheme. The Arab Republic of Egypt, for example, introduced a school health insurance program for all children attending school. The program resulted in larger increases in coverage among the poor and achieved considerable effect on use and out-of-pocket expenditures (Yip and Berman 2001). However, insurance in the developing world is very limited, and those who are least able to smooth consumption without insurance are the least likely to have insurance coverage (Musgrove, Zeramdini, and Carrin 2002). Another problem is that many of the schemes are small scale, and evaluations of these schemes do not generally measure health effect or effect on equity, thus resulting in limited evidence (Palmer and others 2004). Providing Information—Enhancing Knowledge. Lack of knowledge is a major factor behind poor health. It results in people not seeking care when needed, despite the absence of price barriers, and it also results in people—especially poor people—wasting limited resources on inappropriate care. Ignorance may also result in people not getting the maximum health gain out of inputs they have available to them and use. Many people do not know that hand washing confers much of the health benefit of piped water (see chapter 41). Not surprisingly, piped water has a much greater effect on the prevalence of diarrhea among the children of the better off and better educated. Better-educated women—especially those with a secondary education—achieve better health outcomes for themselves and their children not by using health-specific knowledge that they acquire at school, but by using general numeracy and literacy skills learned at school to acquire healthspecific knowledge later in life. Although better-educated girls will mean healthier women and healthier children in years to come, a shorter and more direct route to increasing healthspecific knowledge and skills is through information dissemination, health promotion, and counseling in the health sector. Several success stories exist. In Brazil, after health workers trained by IMCI provided information and counseling at health facilities and in the community, health knowledge among mothers improved, as did feeding practices (Santos and others 2001). After only 18 months, the nutritional status of children in the area improved as well. Social marketing and media campaigns—for example, malaria and social marketing of insecticide-treated nets (see chapter 21)—have also proved effective in some circumstances. Reducing Time Costs Transportation systems, road infrastructure, and geography influence the demand for care delivered by formal providers through their effect on time costs, which can be substantial. In rural communities, where the roads are poor and the transportation unreliable, the time spent waiting for the transportation is also a major cost. Time costs tend to be a major issue for maternal mortality: health centers are unable to provide essential obstetric care for a complicated delivery, and women would have to travel to distant hospitals to get such services. Road rehabilitation and other transportation projects are important here, but so are subsidies linked to the use of health services. Malaysia and Sri Lanka provide free or subsidized transportation to hospitals in emergencies (Pathmanathan and others 2003). Other options for tackling inaccessibility include using outreach and establishing partnerships between government and nongovernmental organizations (NGOs), private providers, or community organizations. Providing Access to Water and Sanitation The availability of adequate supplies of water and improved sanitation is associated with better maternal and child health outcomes, at least among the better educated, even after controlling for other influences. This result is not altogether surprising. Hand washing is easier if the household has piped water that provides readily available quantities of safe water. The safe disposal of feces is easier if the household has an improved form of sanitation. The developing world lags well behind the industrial world in both; the poorer people fare especially badly. They are less likely to be connected to a network, and the sources they rely on tend to be more costly per liter than the networked services used by the better off. The challenge from a health perspective is to get maximum health benefits from investments in access to water and sanitation infrastructure. Efforts to work across sectors on water and health, in order to influence the health MDGs, are under way in Ethiopia, Peru, and Rwanda. Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 187 ©2006 The International Bank for Reconstruction and Development / The World Bank 7 Improving Health Service Delivery Health providers—in the public and private sectors, as well as in both formal and informal sectors—should deliver interventions of relevance to the MDGs. Many are efficient, deliver high quality services, and are responsive to their patients. Many, however, are not; many are not even there to deliver any services at all. As a result, resources—public and private—are often nonexistent, underused, or wasted. Two things can make a difference. One is the quality of management. Better management means a clearer delineation of responsibilities and accountabilities inside organizations, a clearer link between performance and reward, and so on. Management means getting accountabilities right within an organization. The other thing that can make a difference is getting accountabilities right between the organization and the public (World Bank 2003d). Improving Management—Increasing Accountability within Provider Organizations. Management styles in governmentfunded and government-implemented health schemes have recently begun to change, focusing on performance—that is, on outputs and outcomes—rather than on inputs and processes. Good performance is rewarded, financially or in some other way. The focus is on clients and on the belief that an organization is ultimately accountable to its clients. A clientoriented strategy emphasizes customer choice and satisfaction. Business techniques enhance performance and are a standard part of strategic planning. This new approach is evident in several countries, and elements of the approach are visible in successful nutrition and child health programs (see chapter 56). For example, in Tamil Nadu’s Integrated Nutrition Program, community nutrition workers were given clearly defined duties. Information on outputs not only enabled the community to keep the workers accountable but also enabled the nutrition workers to see how their program was working. In Ceara’s Programa de Agentes de Saude, which is credited with a substantial reduction in child mortality (Victora and others 2000), health agents and nursesupervisors were assigned clear tasks and given clear responsibilities. The intended outcomes of the program were emphasized to health workers and members of the public, and the health agents were held accountable through communitybased monitoring and rewarded for good performance. Governance. The accountability of provider organizations to the public can be improved through enhanced governance or contracting. Having community representatives participate in the governance and oversight of providers can improve the productivity and quality of public sector providers. In Burkina Faso, participation of community representatives in public primary health care clinics increased immunization coverage, the availability of essential drugs, and the percentage of women with two or more prenatal visits. In Peru, comparisons of primary health care clinics with and without community participation in governance suggested decreases in staff absenteeism and waiting times and suggested increases in perceived quality by patients (Cotlear 1999). The approach probably works best for primary care and in situations in which strong technical and advisory support is provided to community representatives who are close to the service being delivered. Contracting. Evidence on the effect of contracting within the public sector is mixed, and the experiences are mainly based on lessons learned from middle-income countries. In several countries in Europe and Central Asia, evidence shows a positive effect from performance-based payment, but that is not necessarily the same as contracting, which can occur without performance-related pay. The best evidence relates to the use of target payments for the attainment of a given level of coverage—for example, for immunization or cervical cytology at the primary care level (Langenbrunner 2003). In Argentina and Nicaragua, social security institutes have increased productivity by establishing capitation-based payments for an integrated package of inpatient and ambulatory services (Bitran 2001). Key influences on the success of contracts within the public sector include whether the provider has the ability to respond, whether service commitments are congruent with funding levels, whether output and key components of performance expectations are easily measurable, and how far capacity strengthening of the payer or funder is addressed. Contracting with nonprofit organizations is most common in low-income countries (see chapter 12, which contains a longer discussion of contracting with NGOs). Most cases have had positive effects on target outcome or output variables. In Bangladesh, contracts with nonprofit organizations for planning and implementing an expanded program on immunization project were credited with a dramatic increase in immunization. In Haiti, contracting for a primary health care package also significantly increased immunization coverage (Eichler, Auxilia, and Pollock 2001). In Bangladesh, Madagascar, and Senegal, significant reductions in nutrition rates were attributed to contracting initiatives (Marek and others 1999). Only a few cases assess efficiency. Contracting with nonprofits works best when the contractors have wellfunctioning accountability arrangements and strong intrinsic motivation and when the government makes timely payments to the NGOs. The government needs to be capable of assessing, selecting, and managing the ongoing relationship with contractors. The methodological quality of evaluating contracting is often poor and needs to be improved. An exception is the Cambodian contracting trial that used a rigorous cluster randomized design, but the intervention groups had greater input of resources than the control communities, which may have been partly responsible for the difference in performance. 188 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 8 Results on contracting with for-profit private service providers are also mixed. Experience from the hospital sector warns that weak government contracting capacity often allows the provider to capture efficiency gains or to expand volume— not necessarily of cost-effective services—to generate more income. In Zimbabwe, the cost per service decreased, but the lack of volume control led to an increase in total cost (McPake and Hongoro 1995). Other adverse outcomes are possible. In Brazil, contracting with for-profit hospitals led to increases in access, but also increases in fraud (false billing) and creamskimming to avoid costly patients (Slack and Savedoff 2001). These problems seem less pronounced in primary health care. In Peru and El Salvador, contracting with private primary health care providers increased access, choice, and consumer satisfaction (Fiedler 1996). Contracting with for-profit providers seems to work best when the government invests in the development of capacity to manage the contracting process (Mills, Bennett, and Russell 2001); when quality is at least as high in the private sector as in the public sector; and when the services involve primary care or other relatively observable services, such as diagnostic services. Strengthening Core Public Health Functions Vulnerable populations need to be protected from risks and damages, informed, and educated. Public health regulations need to be established and enforced. Infrastructure needs to be in place to reduce the impact of emergencies and disasters on health. All this action needs to be implemented through a public health system that is transparent and accountable. Governments in developing countries generally recognize that these public health functions are important, but they often lack the capacity and financial resources to implement them. Indeed, few low-income countries invest in these public health functions. By employing public health professionals with core public health competencies, the government can develop and enforce standards; can monitor the health of communities and populations; and can emphasize health education, public information, health promotion, and disease prevention. Public action can help improve consumer knowledge and change attitudes so that private markets can operate effectively to meet the needs of the poor, for example, through social marketing of insecticidetreated bednets to reduce malaria transmission or of condoms for protection against HIV/AIDS. Government-Led Monitoring and Evaluation. Integrated disease surveillance, program assessment, and collection and analysis of demographic and vital registration data are essential if governments and donors are to ascertain whether policies and programs are positively affecting health goals. Governments can use a list of intermediate indicators and proxies for the goals that can help monitor progress, test the impact of policies, and adjust programs going forward (World Bank 2001). Such indicators should be simple, easily measurable, representative, easy to understand, scientifically robust, and ethical. They need to be assessed regularly because the MDGs themselves are difficult to collect, thus entail delays, and are therefore not useful for regular monitoring of progress. Greater investments are needed in systems to monitor these intermediate indicators and to track expenditures on public health. Although some good practices in surveillance are being developed—for example, in Brazil, China, and India—few lowincome developing countries can afford to invest in the infrastructure required for strong surveillance systems. Most rely on alternative short- to medium-term solutions for data gathering, such as intermittent household surveys, health facility surveys, and simplified facility-based routine reporting. A few countries have made special efforts to improve the surveillance of a specific intervention, such as AIDS and tuberculosis treatment or childhood immunization, whereas others have attempted to monitor progress toward a specific MDG. INDEPTH (International Network of Field Sites with Continuous Demographic Evaluation of Populations and Their Health in Developing Countries), which is supported by the Rockefeller Foundation with help from other donors, coordinates a range of surveillance sites, many of them in Africa, and the Health Metrics network aims at improving the quality of surveillance data. Some governments are explicitly developing or modifying their monitoring and evaluation framework to focus on the MDGs. Intersectoral Actions—Going Beyond the Ministry of Health. A review of the evidence base for the key determinants of the health and nutrition MDGs identifies significant potential for intersectoral synergies (Wagstaff and Claeson 2004). Transportation Although roads and transport are vital for health services, especially for reducing maternal mortality, it is not just the physical infrastructure that matters. Also important are the availability of transportation and the affordability of its use, as shown in a study in Nigeria (Eissen, Efenne, and Sabitu 1997). Transportation and roads complement health services. A 10-year study in Rajasthan, India, found that better roads and transportation helped women reach referral facilities, but many women still died because no corresponding improvements took place at household and facility levels. Working with the transportation sector is also important for reducing HIV transmission in many settings and making progress on the HIV/AIDS-related MDG. Hygiene Improved hygiene (use of hand washing) and sanitation (use of latrines and safe disposal of children’s stools) are at least as important as drinking water quality in shaping health Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 189 ©2006 The International Bank for Reconstruction and Development / The World Bank 9 outcomes, specifically in reducing diarrhea and associated child mortality (Esrey and others 1991). Constructing water supply and sanitation facilities is not enough to improve health outcomes; sustained human behavior change must accompany the infrastructure investment. By collaborating with other sectors, the health sector can develop public health promotion and education strategies and implement them in partnership with agencies that plan, develop, and manage water resources. The health sector can also work with the private sector to manufacture, distribute, and promote affordable in-home water purification solutions and safe storage vessels—and advocate for water, sanitation, and hygiene interventions in strategies to reduce poverty. Indoor Air Quality Indoor air pollution is caused by use of low-cost, traditional energy sources, such as coal and biomass for cooking and heating, the main source of energy for 3.5 billion people. Indoor air pollution is a major risk factor for pneumonia and associated deaths in children and for lung cancer in women who risk exposure during cooking (see chapter 42). Studies in China, Guatemala, and India are under way to improve access to efficient and affordable energy sources through local design, manufacturing, and dissemination of low-cost technologies, modern fuel alternatives, and renewable energy solutions. The community-based project in China was initiated by the Ministry of Health, which was troubled by the leveling off of child mortality reductions among the rural poor and was seeking ways to influence major environmental determinants of child mortality. The program combines appropriately improved stoves and ventilation with behavior-change modification; it is in an early stage of implementation, and results on outcomes are not yet available. Agricultural policies and practices influence food prices, farm incomes, diet diversity and quality, and household food security. Policies that focus on women’s access to land, training, and agricultural inputs; on their roles in production; and on their income from agriculture are more likely to have a positive effect on nutrition than interventions without a focus on women, particularly if combined with other strategies, such as women’s education and behavior change (Johnson-Welch 1999; Quisumbing 1995). The MDG agenda highlights the need not only to prioritize within health to achieve better health outcomes, but also to better inform priority setting in resource allocations between sectors, identifying intersectoral synergies and finding ways to maximize benefits for health. COSTING AND FINANCING ADDITIONAL SPENDING FOR THE MDGS Additional health spending will be required in many countries to accelerate progress toward the health goals (see chapter 12). What will it cost, and how will extra spending be financed? Cost of Achieving the MDGs Globally The global estimates of what it would cost to achieve the MDGs range from an additional US$20 billion to US$70 billion a year. A World Bank study (http://www.worldbank.org/html/extdr/ mdgassessment.pdf) estimates that the additional official development assistance required to meet the health goals is in the range of US$20 billion to US$25 billion per year, which is roughly four times the current amount of official development assistance spending for health in 2002 (US$6.5 billion) and three times all external financing, including that of foundations and loans from multilateral sources (see chapter 13). The dramatic shortfalls in resources required to achieve the MDGs were emphasized during the 2002 Monterrey Conference on Financing for Development, which brought significant attention to issues concerning the estimation of the cost of achieving the health MDGs. Another analysis conducted by the Commission on Macroeconomics and Health (2001) of the World Health Organization estimated that an additional US$40 billion to US$52 billion annually would be required until 2015 to scale up the coverage for malaria, tuberculosis, HIV/AIDS, childhood mortality, and maternal mortality (Kumaranayake, Kurowski, and Conteh 2001). A third study using the production frontiers approach estimated that between US$25 billion and US$70 billion of additional spending was needed to bring poorly performing countries up to the level of high performers (Preker and others 2003). A fourth study prepared by the World Bank for the Development Committee estimated at least US$30 billion annually in additional aid was needed to accelerate all the MDGs, including health (Development Committee 2003).Whatever the method of analysis, all global estimates show that reaching the MDGs will require significant additional resources compared with the current levels of funding for health. Cost of Achieving the MDGs in Countries Global estimates of what it costs to achieve the health MDGs are not very useful for countries wanting to plan and budget in order to reach the MDGs. The substantial range of estimates between US$20 billion and US$75 billion per year to achieve the MDGs at a global level has led to debates over the most appropriate costing method for country-specific analysis and to the development of new costing methodologies for obtaining consistent and reliable estimates to use for policy dialogue and decision making at the country levels. Some of the methods are summarized in box 9.2. Preliminary Country Cost Estimates. Table 9.2 provides a set of preliminary country-level estimates for the cost of removing bottlenecks and accelerating progress toward the health MDGs (MBB method) and for the cost of achieving the health MDGs (Millennium Project tools) in selected 190 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 10 countries. The estimates are presented for illustration of orders of magnitude and should not be used for intercountry comparison. Encouraging Risk Pooling Rather Than Out-of-Pocket Spending. Health spending can be broken down into three categories: Financing Extra Health Spending The additional resources needed to reach the MDGs are large at both country and global levels, as discussed in the previous section. The key question is how to finance the extra spending that is needed. • private (out-of-pocket expenditures and private insurance) • public (financing from general revenues and social insurance contributions) • external sources (development assistance). Box 9.2 Estimating the Cost of Scaling Up to Achieve the MDGs The following are the country-specific models for MDG cost analysis: • The MDG Needs Assessments Model developed by the United Nations Millennium Project, (Millennium Project 2004). The Millennium Project model yields total cost estimates for full coverage of the needs of a defined population with a comprehensive set of health interventions in a given year. It uses unit cost of covering one person multiplied by the total population in need in a given year to yield the direct health cost. Additional resource requirements are added (on the basis of assumptions rather than actual inputs) for, among other items, health system improvement, salary increases for human resources, administration and management, promotion of community demand, and research and development. • The Marginal Budgeting for Bottlenecks Model developed by the United Nations Children’s Fund, the World Bank, and the World Health Organization (Soucat and others 2002, 2004; UNICEF and World Bank 2003). The MBB model yields additional resources required for removing a set of health system bottlenecks that are considered to hinder the delivery of health services to the population through three delivery modes: familycommunity, outreach, and clinical levels. The MBB method also estimates the effect on outcomes (for instance, child and maternal mortality) of increased coverage and use of the health services provided. First, a set of high-impact services are selected on the basis of a country’s epidemiological needs. These services are the same as those cost-effective priority interventions identified in the relevant disease control priorities chapters. Second, health system bottlenecks hindering delivery of these services are identified. Then, strategies for removal of bottlenecks are discussed, and the inputs are identified for improving coverage, for example, in a village. Cost estimates are based on these inputs by scaling up the cost to cover the district, province, or nation. • Elasticity estimates through econometric modeling developed by the World Bank staff (Wagstaff and Claeson 2004). A few studies have used econometric techniques to analyze the effect on MDG outcomes of certain crosssector determinants (such as economic growth, water and sanitation, education, and road infrastructure) as well as government expenditures on health. Econometric analysis has been used mostly to analyze the effect of changes in government health expenditures on outcomes using cross-sectional or panel data at a global scale. But in one particular study in India, the methodology was used to estimate the marginal costs of averting a child’s death at the state level. The estimates could vary from as low as US$2.40 per child death in a lowincome state to US$160 in a middle-income state in India. • The Maquette for Multisectoral Analysis of MDGs is under development by the World Bank (Bourguignon and others 2004). The thesis for this new approach is that development aid is a key ingredient of a country’s development process, but its effectiveness has to be assessed at the country level within each country’s local implementation and macroeconomic constraints. The objective of the model is to calculate the financial needs to attain a targeted path to 2015 and determine an optimal allocation of additional funding toward different social sectors for the MDGs. This modeling framework is still at an early stage of development and will be applied later to countries. This model is anticipated to draw extensively from results of other models, such as the elasticity analysis and MBB models. Source: Millennium Project 2003, 2004; Soucat and others 2004; Bourguignon and others 2004. Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 191 ©2006 The International Bank for Reconstruction and Development / The World Bank 11 Table 9.2 Alternative Cost Estimates Using Millennium Project and Marginal Budgeting for Bottlenecks Models Country Model used Cost estimate (US$ per capita per year) Ethiopia MBB 3.56 Madagascar (Toamasina) MBB 2.38 Mali (one region) MBB 3.97 Millennium Project 32.00 Bangladesh Millennium Project 20.60 Cambodia Millennium Project 22.50 Ghana Millennium Project 24.70 Tanzania Millennium Project 34.70 Uganda Millennium Project 32.10 Source: Authors. Private spending absorbs a larger share of income in poorer countries. In low-income countries, it absorbs a larger share of GDP, on average, than domestically financed public spending. In low-income and lower-middle-income countries, it invariably means out-of-pocket expenditures rather than private insurance (Musgrove, Zeramdini, and Carrin 2002). This situation leaves many near-poor households heavily exposed to the risk of impoverishing health expenses. The risk is clearly greater the poorer the country, because poorer countries tend, on average, to have larger shares of poor people (World Bank 2000). Governments thus have a major role to play in helping shape effective risk-pooling mechanisms, in addition to increasing their own spending and targeting it to services for the poor that will have a large positive effect on the MDGs. Getting Governments to Spend What They Can Afford Government spending is an important part of the picture, and the issue is how much they can afford. Unlike private spending, government spending as a share of GDP is higher in richer countries. However, at any given per capita income, a surprising amount of variation occurs across countries in the share of GDP allocated to government health programs. Countries that appear able to spend similar shares of GDP on government health programs end up spending quite different amounts. How can extra domestic resources be mobilized if countries are spending less than they can afford? Domestically financed government health spending comes from general revenues, social insurance contributions, or both. The amount of general revenues flowing into the health sector is the product of the amount of general (tax and nontax) revenues collected by the government (the general revenue share) and the share of general revenues allocated to the health sector (the health share of government spending) (Hay 2003). Low government health spending could be attributable to either share or both shares being low. In poorer countries, both shares are typically lower than they are in richer countries. However, differences exist across countries that cannot be explained by per capita income alone. Countries need to ascertain whether their low spending is caused by unduly low general revenues or by unduly low allocations to health and explore ways of making appropriate adjustments. Bolivia managed to raise its general revenue share consistently in the 1990s as the result of a sustained reform process begun in 1983. The health sector there has been one of the beneficiaries of this growth of tax revenues: government health spending as a share of GDP grew at an annual rate of nearly 10 percent in the 1990s. Although raising domestic resources takes time, countries that can apparently afford to spend more out of their own resources should be encouraged to start the process. Development agencies have a role here—in providing technical support of tax reform, in helping develop government commitment to health in public expenditure allocations, and in giving financial assistance, both to ease the adjustment costs and to provide support while the gap is being closed between current and affordable spending. Recognizing the Limits of Development Assistance. Official development assistance tends to account for a larger share of government health spending in poorer countries. Development assistance for health is especially important in Sub-Saharan Africa. Twelve countries in Sub-Saharan Africa had external funding exceeding 35 percent of total health expenditures in 2000 (World Bank 1998). Increased development assistance is needed to achieve the MDGs. Development assistance, however, is not without its drawbacks. Many donors require that assistance be kept in parallel budgets outside the ministry of finance, which risks undermining government efforts to appropriately plan and target expenditures. Such off-budget expenditures make it difficult in some countries to properly target resources to 192 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 12 particular interventions, geographic locations, or population groups, even though such targeting may be essential for improving the effect of expenditures on outcomes and the probability of reaching the health goals. Donors often require recipient governments to maintain separate accounts and to provide separate progress reports, thereby increasing the administrative burden on weak health ministries. Most important, donor commitments of expenditures in health are short term, whereas the needs are permanent. Thus, any external financing must at some point be substituted with additional domestic revenues or expenditure reallocations. This substitution or transition to domestic sources of funding has typically been difficult to achieve, leading to a dropoff in effort in important health programs, such as immunizations and reproductive health services. Consensus on how to improve aid effectiveness is growing among development partners, and partners at the High Level Forum on Health MDGs (http://www.hlfhealthmdgs. org). This agenda includes supporting countries in developing more MDG-responsive Poverty Reduction Strategy Papers, tracking resource flows, strengthening monitoring and evaluation, and more effectively dealing with the human resources crisis in health. Effective monitoring can help ensure that increased external funds do not simply lead to reduced domestic financing (the fungibility problem) but actually boost overall spending for health. In concert with moves affecting all development assistance, donors and governments are trying to see that in the health area external funds are pooled and that ministries can use a common management and reporting format. In addition, a research agenda to support acceleration toward the health MDGs is being proposed; it needs to focus on how to translate knowledge into action and on how to remove health systems constraints to scaling up coverage of costeffective interventions that are available but do not reach those who need them (Claeson and others 2004; Task Force on Health Systems Research 2004). Bourguignon, F., M. Bussolo, H. Lofgren, H. Timmer, and D. van der Mensbrugghe. 2004. “Towards Achieving the Millennium Development Goals in Ethiopia: An Economywide Analysis of Alternative Scenarios.” World Bank, Washington, DC. Claeson, M., C. Griffin, T. Johnston, M. McLachan, A. Soucat, A. Wagstaff, and A. Yazbeck 2001. “Poverty-Reduction and the Health-Sector.” In Poverty Reduction Strategy Sourcebook. Washington, DC: World Bank. Claeson, M., A. Wagstaff, E. Bos, P. Hay, and J. Baudouy. 2004. “The Case for Mobilizing New Research behind the Health Millennium Development Goals.” In Global Forum Update on Research for Health 2005, 73–75. Geneva: Global Forum. Commission on Macroeconomics and Health. 2001. “Macroeconomics and Health: Investing in Health for Economic Development—Report of the Commission on Macroeconomics and Health.” December 20, World Health Organization, Geneva. Cotlear, D. 1999. “Peru: Improving Health Care for the Poor.” Human Development Department (Latin America and the Caribbean Human Development) Paper 57, World Bank, Washington, DC. Development Committee. 2003. “Supporting Sound Policies with Adequate and Appropriate Financing.” Discussion paper, World Bank, Washington DC. Eichler, R. 2001. “Improving Immunization Coverage in an Innovative Primary Health Care Delivery Model: Lessons from Burkina Faso’s Bottom up Planning, Oversight, and Resource Control Approach That Holds Providers Accountable for Results.” Discussion paper, World Bank, Washington, DC. Eissen, E., D. Efenne, and K. Sabitu. 1997. “Community Loan Funds and Transport Services for Obstetric Emergencies in Northern Nigeria.” International Journal of Gynecology and Obstetrics 59 (Suppl. 2): S237–44). Esrey, S. A., J. B. Potash, L. Roberts, and C. Shiff. 1991. “Effects of Improved Water Supply and Sanitation on Ascariasis, Diarrhea, Dracunculiasis, Hookworm Infection, Schistosomiasis, and Trachoma.” Bulletin of the World Health Organization 69 (5): 609–21. FAO (Food and Agriculture Organization of the United Nations). 2000. The State of Food Security in the World. Rome: FAO. Fiedler, J. L. 1996.“The Privatization of Health Care in Three Latin American Social Security Systems.” Health Policy and Planning 11 (4): 406–17. Filmer, D., and L. Pritchett. 1999.“The Impact of Public Spending on Health: Does Money Matter?” Social Science and Medicine 49, pp. 1309–23. Gwatkin, D., S. Rutstein, K. Johnson, and R. P. Pande. 2000. Socio-economic Differences in Health, Nutrition, and Population. Washington, DC: World Bank. NOTE Hay, R. 2003. “The ‘Fiscal Space’ for Publicly Financed Health Care.” Oxford Policy Institute Policy Brief, Washington, DC. 1. Intervention in this chapter refers to the direct action that leads to prevention or cure. Johnson-Welch, C. 1999. “Focusing on Women Works: Research on Improving Micronutrient Status through Food-Based Interventions.” International Center for Research on Women, Washington, DC. REFERENCES Jones, G., R. W. Steketee, R. E. Black, Z. A. Bhutta, and S. S. Morris. 2003. “How Many Child Deaths Can We Prevent this Year?’ Lancet 362 (9377): 65–71. Bell, C., S. Devarajan, and H. Gersbach. 2003. The Long-Run Economic Cost of AIDS: Theory and an Application to South Africa. Washington, DC: World Bank. Bitran, R. 2001. “Paying Health Providers through Capitation in Argentina, Nicaragua, and Thailand: Output, Spending, Organizational Impact, and Market Structure.” USAID Partners for Health Reform Project, Washington, DC. Bokhari, F., P. Gottret, and Y. Gai. Forthcoming. “Government Health Expenditures, Donor Funding and Health Outcomes.” World Bank, Washington, DC. Kumaranayake, L., Christoph Kurowski, and Lesong Conteh. 2001. “Costs of Scaling Up Priority Health Interventions in Low-Income and Selected Middle-Income Countries: Methodology and Estimates.” Commission for Macroeconomics and Health Working Paper WG5:19, World Health Organization, Geneva. Langenbrunner, J. 2003. “Resource Allocation and Purchasing in ECA Region: A Review.” Discussion paper, World Bank, Washington, DC. Marek, T., I. Diallo, B. Ndiaye, and J. Rakotosalama. 1999. “Successful Contracting of Prevention Services: Fighting Malnutrition in Senegal and Madagascar.” Health Policy and Planning 14 (4): 382–89. Millennium Development Goals for Health: What Will It Take to Accelerate Progress? | 193 ©2006 The International Bank for Reconstruction and Development / The World Bank 13 McPake, B, and C. Hongoro. 1995. “Contracting Out of Clinical Services in Zimbabwe.” Social Science and Medicine 41 (1): 13–24. Mesoamerica Nutrition Program Targeting Study Group. 2002. “Targeting Performance of Three Large-Scale, Nutrition-Oriented Programs in Central America and Mexico.” Food and Nutrition Bulletin 232 (2): 162–74. Millennium Project. 2003. “Millennium Development Goal Country Case Studies: Methodology and Preliminary Results.” October. United Nations, NY. ———. 2004. “Millennium Development Goals Needs Assessments: Country Case Studies of Bangladesh, Cambodia, Ghana, Tanzania, and Uganda.” Unpublished working paper for the Millennium Project. United Nations, NY. Mills, A., S. Bennett, and S. Russell. 2001. The Challenge of Health Sector Reform: What Must Governments Do? New York: St. Martin’s Press. Morris, S., E. Flores, P. Olinto, and J. Medina. 2003. “A Randomized Trial of Conditional Cash Transfers to Household and Peripheral Health Centres: Impact on Child Health and Demand for Health Services.” Presented at Fourth International Health Economics Association World Congress, San Francisco, CA, June. Musgrove, P., R. Zeramdini, and G. Carrin. 2002. “Basic Patterns in National Health Expenditure.” Bulletin of the World Health Organization 80 (2): 134–42. Newman, J., M. Pradhan, L. Rawlings, G. Ridder, R. Coa, and J. Evia. 2002. “An Impact Evaluation of Education, Health, and Water Supply Investments of the Bolivian Social Investment.” World Bank Economic Review 6 (2): 241–74. OECD (Organisation for Economic Co-operation and Development) Development Assistance Committee. 2000. Recent Trends in Official Development Assistance to Health. Paris: OECD. Palmer, N., D. Mueller, L. Gilson, A. Mills, and A. Haines. 2004. “Health Financing to Promote Access in Low Income Settings—How Much Do We Know?” Lancet 364: 1365–70. Pathmanathan, I., J. Liljestrand, J. M. Martins, L. C. Rajapaksa, C. Lissner, A. de Silva, and others. 2003. “Investing in Maternal Health: Learning from Malaysia and Sri Lanka.” Health, Nutrition, and Population Department, World Bank, Washington, DC. Preker, A. S., E. Suzuki, F. Bustero, A. Soucat, and J. Langenbrunner. 2003. “Costing the Millennium Development Goals.” Background paper to The Millennium Development Goals for Health: Rising to the Challenges, World Bank, Washington, DC. Quisumbing, A. R. 1995. “Gender Differences in Agricultural Productivity: A Survey of Empirical Evidence.” IFPRI Discussion Paper 5. International Food Policy Research Institute, Washington, DC. Rajkumar, A., and V. Swaroop. 2002. “Public Spending and Outcomes: Does Governance Matter?” Policy Research Working Paper 2840, World Bank, Washington, DC. Ramana, G. 2003. Background paper for The Millennium Development Goals for Health: Rising to the Challenges. World Bank, Washington, DC. Santos, I., C. G. Victora, J. Martines, H. Goncalves, D. P. Gigante, N. J. Valle, and G. Pelto. 2001. “Nutrition Counselling Increases Weight Gain among Brazilian Children.” Journal of Nutrition 131 (11): 2966–73. Schellenberg, J. R., S. Abdulla, R. Nathan, O. Mukasa, T. J. Marchant, N. Kikumbih, and others. 2001. “Effect of Large-Scale Social Marketing of Insecticide-Treated Nets on Child Survival in Rural Tanzania.” Lancet 357 (9264): 1241–47. Sen, A. 2002. “Why Health Equity?” Health Economics 11 (8): 659–66. Slack, K., and W. D. Savedoff. 2001. “Public Purchaser–Private Provider Contracting for Health Services: Examples from Latin America and the Caribbean.” Sustainable Development Department Technical Paper 111, Inter-American Development Bank, Washington, DC. Soucat,A.,W.Van Lerberghe, F. Diop, S. Nguyen, and R. Knippenberg. 2002. “Marginal Budgeting for Bottlenecks: A New Costing and Reallocation Practice to Buy Health Results.” World Bank, Washington, DC. ———. 2004. “Marginal Budgeting for Bottlenecks: A New Costing and Resource Allocation Practice to Buy Health Results—Using Health Sector’s Budget Expansion to Progress toward the Millennium Development Goals in Sub-Saharan Africa.” Unpublished paper. World Bank, Washington, DC. Task Force on Health Systems Research. 2004. “Informed Choices for Attaining the Millennium Development Goals: Towards a Cooperative Agenda for Health Systems Research.” Lancet 364: 997–1003. UNAIDS (United Nations Joint Programme on HIV/AIDS). 2004. Report on the Global AIDS Epidemic. Geneva: UNAIDS. UNICEF (United Nations Children’s Fund). 2001. Progress since the World Summit for Children: A Statistical Review. New York: UNICEF. Victora, C., F. Barros, J. Vaughan, A. Silva, and E. Tomasi. 2000. “Explaining Trends in Inequities: Evidence from Brazilian Child Health Studies.” Lancet 356 (9235):1093–38. Wagstaff, A., and M. Claeson. 2004. The Millennium Development Goals for Health: Rising to the Challenges. Washington, DC: World Bank. ———. 2005. “The Millennium Development Goals for Health: Rising to the Challenges: Appendix A, pp. 169–174.” World Bank. Washington, DC. Wagstaff, A., and E. van Doorslaer. 2003. “Catastrophe and Impoverishment in Paying for Health Care: with Applications to Vietnam 1993–1998.” Health Economics 12 (11): 921–34. World Bank. 1998. Assessing Aid: What Works, What Doesn’t, and Why. Oxford, U.K.: Oxford University Press. ———. 2000. World Development Report 2000/2001: Attacking Poverty. New York: Oxford University Press. ———. 2001. “‘Health, Nutrition, and Population Development Goals: Measuring Progress Using the Poverty Reduction Strategy Framework.” Report of a World Bank Consultation, World Bank, Washington, DC. ———. 2003a. Attaining the Millennium Development Goals in India: How Likely and What Will It Take? Washington, DC: World Bank. ———. 2003b. Global Economic Prospects and the Developing Countries. Washington, DC: World Bank. ———. 2003c. World Development Indicators 2003. Washington, DC: World Bank. ———. 2003d. World Development Report 2004: Making Services Work for Poor People. Washington, DC: World Bank. Yip, W., and P. Berman. 2001. “Targeted Health Insurance in a Low Income Country and Its Impact on Access and Equity in Access: Egypt’s School Health Insurance.” Health Economics 10 (3): 207–20. 194 | Disease Control Priorities in Developing Countries | Adam Wagstaff, Mariam Claeson, Robert M. Hecht, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 14 Chapter 11 Fiscal Policies for Health Promotion and Disease Prevention Rachel Nugent and Felicia Knaul Governments use fiscal policy to encourage healthy behavior. The instruments of government for this purpose are taxes and subsidies, and direct provision of certain health services for free or at subsidized rates. Examples of fiscal policies for health are taxes on tobacco and alcohol, subsidies on certain foods, and tax incentives for health care purchases. Government intervention through fiscal policy works best when public institutions and credibility are strong, the design and application of the fiscal instruments are appropriate, and consumers’ and producers’ responsiveness to a price signal is high. When these conditions are not present, direct provision, information and education campaigns, or legislation may be preferable in conjunction with fiscal policy. The purpose of this chapter is to review country experiences with promoting health through fiscal policies and to examine the usefulness and success of these policies. The chapter considers both the role of fiscal policies in the production of health and the effect of these policies on the well-being of the economy— fiscal policy for health and healthy fiscal policy.1 Little exists in the literature linking fiscal policy and health promotion except in relation to tobacco. This work contributes to filling that gap. The chapter deals specifically with experiences at the country level with tax policies affecting some goods related to health, such as food, tobacco, alcohol, and condoms; subsidized provision of workplace promotion of healthy behavior and caregiving; and direct subsidies affecting food provision and fortification, cooking fuels, water purification and soap, condoms, bednets, vaccines, and medical research. The chapter only touches on health care provision and does not discuss its financing, either directly by governments or through insurance, because other chapters deal with those topics. The chapter is divided into five sections: • The first section provides a general framework through which fiscal policy options can be considered in terms of their impact on health and the health sector. • The next section examines the experiences in developing countries of using subsidies to achieve health-related objectives (columns 1 and 2 in table 11.1). • The third section presents examples of how taxes are used in a number of countries to promote health (column 3 in table 11.1). • The next section discusses nonhealth goods where fiscal policies are often used and have important indirect health benefits (lower part of column 2 in table 11.1). • The final section presents conclusions and suggestions for further research and policy development. USE OF FISCAL POLICY FOR HEALTH IN DEVELOPING COUNTRIES Fiscal policies come in a wide range of designs, but the main effect is either to alter the price of health-related goods or to alter the quantity available. Table 11.2 summarizes the health interventions subject to fiscal policies. The behaviors that require these interventions are divided into the following categories: • unhealthy consumption (foods, tobacco, and alcohol), for which the most salient fiscal policies are taxes on consumers and producers, and fines 211 ©2006 The International Bank for Reconstruction and Development / The World Bank 15 Table 11.1 Fiscal Policies for Health Promotion Covered in Chapter 11 Subsidy for or tax imposed on Health-related products receiving direct subsidies Subsidized provision of health Health-related products directly taxed Government financing of health care Consumer Medicine Caregiving (partially covered) Tobacco Not covered Food Alcohol Cooking fuel Food Water purification Imported medicine and supplies Soap Condoms Bednets Producer Vaccinations Food additives Workplace promotion of healthy behavior (partially covered) Fuel usage (partially covered) Medical research Source: Authors. • health promotion and disease and accident prevention (hygiene, pollution, safety, public health, maternal and child and reproductive health, infectious disease, and healthy lifestyles), for which the most important fiscal instruments are subsidies, but which may also be affected by tax policy • health care goods and inputs, including insurance and human resources, that may be exempted from taxation, subsidized, or guaranteed as a constitutional right • other goods that indirectly promote health (education, housing, agriculture, energy, charitable giving, charities that provide targeted subsidies, and so on), which are often subject to their own particular tax regime or sets of subsidies that affect their production or consumption and, therefore, also affect health behavior • research and development initiatives that can be applied to health and health care goods and are sensitive to tax exemptions and subsidies. Fiscal interventions can have various rationales, such as macroeconomic benefits, equity, or efficiency—and promoting health may or may not be the primary goal. A fiscal policy may be designed to affect some other sphere of behavior or a good other than health—for instance, education—and the effects on health or the use of health care may be indirect. The shaded boxes in table 11.3 indicate the possible rationale behind each type of fiscal policy. A fiscal policy should be effective, efficient, and costeffective and should promote or maintain equity goals. An effective tax or subsidy reaches the intended target and alters health-related behavior in the desired manner. An efficient policy minimizes resource distortions and involves minimal administrative costs. A cost-effective policy has the lowest cost relative to the desired health goal. SUBSIDIES FOR HEALTH AND HEALTH-RELATED PRODUCTS Using examples primarily from developing countries, this section of the chapter analyzes the range of subsidies that are available to promote healthy behavior and the consumption of health-related goods. The first sections deal with consumer subsidies both to promote the consumption of healthproducing goods and of health care. The second section discusses producer subsidies. Consumer Subsidies Governments use consumer subsidies to encourage the use of a beneficial product by lowering the price consumers pay— usually in situations where the consumers are too poor, the market prices of the good are too high, or both situations apply—to otherwise achieve a socially optimal consumption level. Examples include subsidies for staple foods, condoms, soap, insecticide-treated bednets, cooking fuels, and medicines. Staple Foods. Ample evidence indicates that food subsidies are effective in improving nutrition; however, appropriate targeting is often a problem (Alderman 2002). Subsidies may be targeted to specific foods, specific delivery locales or geographic areas, or specific populations. Often targeting includes all three. Food-specific subsidies, whether in the form of general subsidies, ration cards, quotas, or food stamps, increase food consumption. They will have a positive effect on health if this consumption occurs in undernourished populations that require increased caloric or nutrient intake. In some cases, food subsidy programs have had unintended macroeconomic and 212 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul ©2006 The International Bank for Reconstruction and Development / The World Bank 16 Table 11.2 Use of Taxes and Subsidies to Promote Health by Type of Intervention Tax preferences Taxes Intervention Payroll Consumer (sales and value added tax) Excise Production Input Fines Credits and exemptions Subsidies Directorsubsidized provision Subcontract provision Targeted consumer subsidies Producers Rights Foods Alcohol Tobacco Health promotion and disease and accident prevention Hygiene (soap) Pollution (for example, fuels) Safety (for example, seat belts) Public health (vaccines, clean water, supplementation, education, and information) Fiscal Policies for Health Promotion and Disease Prevention | 213 ©2006 The International Bank for Reconstruction and Development / The World Bank 17 Unhealthy consumption Child, maternal, and reproductive health (information and education, supplementation, and medical attention) Infectious disease (condoms and healthy workplaces) Healthy lifestyles (food, exercise, and healthy workplaces) (Continues on the following page.) ©2006 The International Bank for Reconstruction and Development / The World Bank 18 214 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul Table 11.2 Continued Tax preferences Taxes Intervention Payroll Consumer (sales and value added tax) Health care goods Insurance Medical attention Medicines Human resources for health Other goods that indirectly promote health Personal leave Caregiving and family leave (maternity, paternity, and chronic illness) Education (for mothers or women, early childhood, or special needs) Housing (provision, flooring, and roofing) Agriculture (type of products produced or imported) Energy (types of fuels, heating, cars, and gasoline) Charitable giving Research and development Source: Authors. Note: Shaded boxes indicate possible fiscal policies for each type of intervention. Excise Production Input Fines Credits and exemptions Subsidies Directorsubsidized provision Subcontract provision Targeted consumer subsidies Producers Rights Table 11.3 Taxes and Subsidies by Policy Rationale Rationale Policy Healthy behavior Payroll Consumer (value added tax/sales) Levy Excise Producer Input Fines Subsidies Tax credits Direct provision Fiscal Policies for Health Promotion and Disease Prevention | 215 ©2006 The International Bank for Reconstruction and Development / The World Bank 19 Taxes Subsidized provision Subcontracted provision Targeted subsidies Subsidies to producers Rights Tax exemptions Source: Authors. Note: The shaded boxes indicate the possible rationale behind each type of fiscal policy. Macroeconomic or fiscal benefits Equity Efficiency Promotion of another good or type of behavior microeconomic consequences (Adams 2000; del Ninno and Dorosh 2002; Pinstrup-Anderson 1988; Siamwalla 1988). They become expensive if they are too widely available, can create incentives for black market activities, and can affect prices and volumes in agricultural and trade markets. Indonesia switched from a general rice support system to a limited subsidy during the 1997 macroeconomic crisis. The earlier system had successfully reduced food insecurity to low levels, but higher prices increased the cost of maintaining the subsidy and led to food being smuggled out of the country (Tabor and Sawit 2001). The government targeted the new rice subsidy to the poor and issued ration cards. Within roughly a year of implementation, the subsidy was reaching an estimated 85 percent of the poor. Only about 10 percent of the subsidy appeared to be reaching nontarget population groups. India has subsidized essential consumer goods for decades, including health-related goods such as food grains, edible oils, sugar, and fuels (S. Jha 1992). The government rationed certain goods in the belief that only the truly needy would endure waiting in lines and purchasing the poorer quality products that were involved in the subsidy schemes. This is called selftargeting. However, Jha shows that 40 percent of the population purchased subsidized rice in 1990, only half of whom were poor. The government recently modified the program to better target the subsidy to the poor and removed such barriers as bulk purchasing (Rao 2000). The Arab Republic of Egypt’s generalized program also illustrates the problems that beset broad food subsidy programs. The program reached its zenith in 1980, when it subsidized 20 food products and accounted for 15 percent of government expenditures (Adams 2000). The program has been scaled back to cover four staple foods and now accounts for 6 percent of government expenditures. Nevertheless, about 75 percent of the population holds ration cards entitling them to purchase the subsidized foods. The program is intended to achieve selftargeting, but the nonpoor purchase many of the subsidized foods. The program accounts for 44 percent of the total calorie supply of the poorest quintile group, but in rural areas, the rich obtain more calories from subsidized food than the poor do. Musgrove (1993) reviews 104 supplementary feeding programs in 19 countries in Latin America and the Caribbean. The review covers a range of program sizes, from those serving 1,000 individuals to those supplying 28 million people; of types of subsidies (namely, food distribution, direct feeding, and direct payments); of levels of coverage of the targeted population, ranging from 1.9 to 100.0 percent; and of extent of coverage of the poor, varying from 5.8 to 88.0 percent. The per capita costs of reaching beneficiaries differed widely. The most common reasons for program ineffectiveness were spreading resources too thinly across beneficiaries, targeting foods with minor health benefits, choosing inappropriate beneficiaries, and encountering excessive costs in distributing resources. These kinds of issues underscore the importance of design considerations and country conditions in creating effective and efficient food subsidy programs. In sum, many food subsidy programs avoid the political and administrative challenges of explicit targeting by allowing universal access to the subsidies on the assumption that the needy will self-select into the programs. However, Adams (2000) shows that countries with targeted food programs—for example, Chile, Jamaica, and Peru—provide much higher income transfers to the poor than do self-targeted programs of the kind used in Egypt, Morocco, and Tunisia. Condoms. Preliminary investigation indicates that subsidies on condoms can be effective in increasing their use in both general and high-risk populations, but whether price reduction, increased access, or education leads to greater use is not clear (Price 2001) because information campaigns about the health benefits of condoms usually accompany price subsidies. Recent surges in social-marketing schemes to distribute condoms as part of the fight against HIV/AIDS, especially in Africa, have increased condom use.2 Few researchers have compared HIV infection rates—or even condom use rates—before and after the introduction of a subsidy on condoms. Cohen and others (1999) conclude that in a particular jurisdiction in Louisiana, free distribution through public clinics and 1,000 small businesses in areas with high levels of HIV and other sexually transmitted diseases achieved significantly higher condom distribution than a fee-based system (77 percent use during the last sexual encounter compared with 64 percent) and that the revenues from cost recovery were insufficient to justify imposition of the fee. The dropoff in condom use during the cost-recovery period persuaded the jurisdiction to reinstate the free distribution program. Another example suggests that promotion and information are also effective. A social-marketing effort in Turkey in the early 1990s offered condoms at a commercial price but included intensive advertising and other promotional efforts. It achieved sales well beyond original expectations and gained 41 percent of the market share (Yaser 1993). Water Purification. The U.S. Centers for Disease Control and Prevention and the Pan American Health Organization designed the Safe Water System Initiative to improve the quality of drinking water for households that draw their water from sources outside the home. The principle underlying the initiative is to subsidize storage containers, disinfectant, and education on proper handling to avoid contamination (Quick and others 1999). Numerous countries have implemented similar initiatives, including Bangladesh, Bolivia, Burkina Faso, Kenya, and Zambia. The government provides containers and chemicals at subsidized prices, but the costs are still higher than the cost of boiling water (Quick and others 2002). 216 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul ©2006 The International Bank for Reconstruction and Development / The World Bank 20 Soap. Another proven method for reducing the incidence of diarrhea and other hygiene-related diseases is hand washing, with or without soap. Whether the key factor is education or the subsidized provision of soap is unclear. Some investigators claim that small-scale programs that subsidize soap and educate households about the benefits of hand washing are selffinancing because of the consequent reduction in disease (Borghi and others 2002). Luby and others’ (2001) results from Pakistan suggest that education alone may be just as effective as education accompanied by soap provision in reducing diarrheal disease. By contrast, Hoque (2003) and other researchers suggest that the cost of soap is a barrier to its widespread use among extremely poor populations and that behavioral change may be difficult to achieve without a subsidy. Insecticide-Treated Bednets. The degree of subsidization of bednets has become a controversial issue, with some arguing for full subsidization and others for partial subsidization. Most long-term studies indicate that consumers resist purchasing bednets even at subsidized prices after they have had access to free bednets (Snow and others 1999). A number of researchers have undertaken studies in various locations in Africa to assess the effect of selling bednets rather than providing them free to vulnerable populations (Armstrong-Schellenberg and others 2001; Kolaczinski and others 2004; Snow and others 1999). The key issue is consumers’ responsiveness to changes in the prices of bednets, through either subsidies or a reduction in taxes and tariffs. Many households do not own a bednet because they cannot afford it, while other reasons are lack of information, poor access to markets, and cultural preferences (Hanson and Worrall 2002; Simon and others 2002). The evidence suggests that responsiveness to price changes alone may be modest, but in combination with removing some of the other barriers, demand for bednets could increase substantially in malariaaffected regions (Simon and others 2002). Nigeria removed tariffs and taxes on bednet insecticide in 2001, and the 18 percent price drop resulted in an estimated 9 to 27 percent increase in purchases (Simon and others 2002). Another study that reviewed a public sector subsidy for bednets combined with private sector marketing and distribution by means of a social-marketing scheme concluded that the program was successful because 18 percent of children slept under bednets as a result; however, the low insecticide retreatment rate led the authors to conclude that subsidies were needed on both bednets and insecticide (Armstrong-Schellenberg and others 2001). Clean Cooking Fuels. High rates of respiratory illness occur as a result of exposure to smoke and particle emissions from biomass burning in many developing countries. Fuel subsidy pro- grams have been designed to promote the use of liquid petroleum gas, natural gas, or kerosene, which burn more cleanly and emit a low amount of smoke and particulates, but none has been efficacious or efficient (UNDP 2003). Liquid petroleum gas subsidies have been shown to benefit middle- and higher-income families in urban areas rather than the poor (UNDP 2003). In attempting to target the poor more accurately, Côte d’Ivoire and Senegal focused subsidies on smaller liquid petroleum gas cylinders but found that poor consumers still preferred charcoal (UNDP 2003). Electricity subsidies in low-income countries are also often skewed toward the well off, who are more likely than the poor to be connected to the electricity grid (Alderman 2002). Medicines and Medical Supplies. In relation to the direct provision of health-related goods, including drugs, supplies, and services of medical personnel, governments may subsidize and regulate drug prices, make bulk purchases from manufacturers for distribution at reduced prices, and distribute certain drugs with complete or partial subsidies to target populations. Specific interventions—for instance, antiretrovirals, vaccines, or reproductive health care—are often more heavily subsidized or may be targeted by population group or disease—for example, child and maternal health, tuberculosis, and malaria. With the exception of antiretroviral drugs, the health benefits and low costs of these medicinal interventions make them good targets for subsidization. General Health Care. In developing countries, where informal sectors tend to be large, providing subsidized health care is an important tool for health promotion. Some countries have chosen direct provision of health goods, whereas others combine the public provision of services with subsidized health insurance for families below a certain income cutoff. Both models require identifying the families that are unable to afford health care and the types of services that are considered public goods. One example of subsidizing the production and provision of health care is the Mexican program originally called PROGRESA and now known as Oportunidades. This program is also an example of how income transfers for other goods can affect health and how cross-subsidies can be used to strengthen the incentive effects of a fiscal policy to promote healthy behavior. The government launched the program in 1997 to provide subsidized health, nutrition, and education to poor families. By mid 2004, it was serving the majority of those living below the poverty line. Oportunidades combines a cash transfer equivalent to 20 to 30 percent of families’ incomes that includes incentives for positive behaviors in relation to health, nutrition, and schooling with subsidized basic health interventions. The program is largely financed from federal budgets. Fiscal Policies for Health Promotion and Disease Prevention | 217 ©2006 The International Bank for Reconstruction and Development / The World Bank 21 Oportunidades is successful both in terms of targeting the poorest households and in terms of achieving measurable gains in health, health care use, nutritional status and growth, school attendance, and school achievement. Gertler (2004), for example, finds significant and cumulative reduction in illness rates among children, lower prevalence of anemia, and an additional centimeter of growth in the first year of the program. The program’s success is attributable to many factors, including a rigorous longitudinal evaluation process; an integrated package of services; and the presence of financial stimuli tied to school attendance, visits to health clinics, and participation in health education initiatives. Furthermore, the program incorporates several targeting methods. Producer Subsidies Governments use producer subsidies to encourage production that improves health by lowering manufacturers’ costs in situations in which the private market supply is inadequate to meet social needs. Examples include medical supplies, vaccines, food additives, and medical research. Food Fortification. Governments sometimes subsidize the fortification of staple foods through the addition of selected micronutrients as a way of achieving broadly based nutrition improvements. Challenges involve maintaining a relationship between the public sector, which initiates and funds the program, and the private sector, which implements the fortification. Incentives for private providers are often needed in the form of tax exemptions, import preferences, subsidies for startup costs, quality control, and training. Illegal markets selling nonfortified products at a lower price often arise in response (Alderman 2002; Dorosh, del Ninno, and Sahn 1996; Rao 2000). Health Research. Government support for health research consists of the provision of direct subsidies for private sector investment, the granting of tax benefits for private research and development (R&D) investment, the establishment of property rights and a system to protect them, and the promotion of private goods by other means (OECD 2003). Despite the strong evidence from developed countries that the private sector will underinvest in R&D and that tax incentives increase R&D investment, developing countries should be cautious in applying those results to their own situations. Empirical investigations tend to conclude that producer subsidies for R&D in developing countries are not effective (Shah 1995; Zee, Stotsky, and Ley 2002). Many conditions need to be in place to realize high social returns and to minimize rent seeking and profiteering, including a strong private sector research effort that is stimulated by the public investment, the presence of appropriate targeting, a transparent and fair set of public laws and insti- tutions to grant and monitor the tax benefits, and the ability to forgo alternative public investments. TAXES AND TAX EXPENDITURES: DESIGN AND OUTCOMES The following section describes various examples of the use of taxation directed at both consumers and producers. This section of the chapter also analyzes the design issues that are important in order to guarantee that these instruments contribute to achieving healthy fiscal policy. Taxes on Consumers Sales taxes—including excise taxes and value added taxes— and exemptions from those taxes are the most common fiscal policy tools used to influence consumers’ health purchases. Examples are exempting medicines and foods from sales tax and imposing an excise tax on cigarettes and alcohol. Developed countries often use income tax incentives to provide deductions and credits for specific health care purchases. Box 11.1 discusses issues surrounding use of taxes for health. “Sin” Taxes on Tobacco and Alcohol. A wide range of countries and local jurisdictions have taxed tobacco, with acknowledged success in reducing consumption (P. Jha 1999). The health benefits of curbing the demand for cigarettes may go beyond eliminating the health consequences of smoking and secondhand smoke if consumer expenditures are diverted from cigarettes to healthier alternatives (for example, food). Taxes on alcohol are widespread and are used primarily to raise revenue. Governments typically impose taxes at the producer, wholesale, and retail levels that are levied as a percentage of the sale price or are based on a flat amount per unit. Harmful alcohol consumption is controlled through prohibition, government monopolization of sales, “dry” days, restrictions on hours when sales are legal, restrictions on age and locations for sales and consumption, laws against drinking and driving, limits to alcohol content, laws against the sale of certain types of alcohol, and licensing. Alcohol taxes do contribute revenue to government coffers in developing countries, generally in higher proportions than in developed countries (WHO 2002a), but smuggling and tax evasion are common. For example, Zimbabwe raised taxes on certain beers in 1995 but repealed the increase within months when tax revenues dropped significantly (WHO 2002a). Some developing countries have lowered alcohol taxes with consequent negative results. Mauritius experienced a dramatic increase in drunk-driving arrests, alcohol-related fatalities, and hospital admissions after it reduced taxes on alcohol (WHO 2002a). In sum, alcohol taxes do reduce drinking, but the evidence that such taxes are well targeted to those most at risk of problem drinking is not strong. 218 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul ©2006 The International Bank for Reconstruction and Development / The World Bank 22 Box 11.1 Using Taxes to Influence Consumption and Production Behavior Taxes as a tool for health policy face significant implementation obstacles. First, targeting can be difficult. A close link must exist between the consumption of the product or behavior to be taxed and a specific population with a health risk. For instance, all consumers would pay a tax on “junk” food, even though it would only present a health threat to a small percentage of them. The taxed good must also be appropriately defined in relation to close substitutes; for example, taxing only certain forms of tobacco such as cigarettes, but not chewing tobacco, may increase consumption of the latter. Governments may also distinguish between locally produced goods and imported goods, often because of lobby groups. If governments place a higher tax on the good that is less harmful, this action will encourage greater consumption of the more harmful good. Key weaknesses in using taxes for health policy include the feasibility of smuggling and the existence of large informal or illegal markets. Smuggled or contraband products that cannot be regulated or certified for quality and safety, such as alcohol or tobacco in particular, may be more harmful to health than goods that are legally produced and sold. Any tax should be efficient in terms of both its administration and its effect on resource allocation. Tax authorities need a well-functioning system for imposing, collecting, and monitoring taxes and taxed products, and the public should perceive the system as fair and credible in order to achieve a high degree of compliance. Finally, a tax should be cost-effective in achieving its stated goal of improving health outcomes. The net costs of imposing the tax should compare favorably with the net costs of using another policy instrument, such as regulation or direct government provision. Depending on the characteristics of the tax base, the health goal and the revenue goal may even be at odds. Source: Authors. Food Taxes. The issue of taxing unhealthy foods has received increasing attention in the wake of the Global Strategy on Diet, Physical Activity, and Health, which was approved by member countries of the World Health Organization (WHO 2004). The global strategy points to the rising prevalence of obesity and overweight in developing countries, along with that of nutrition-related noncommunicable diseases, and recommends that countries consider fiscal policies and other measures to reduce those problems. Governments can use excise taxes to reduce the consumption of unhealthy foods only if tax rates are sufficient to change consumption in a way that improves health outcomes, if they tax enough harmful foods or food ingredients, and if they levy the taxes in an effective manner. Guo and others’ (1999) study in China demonstrates significant potential for price changes to affect consumption. The researchers studied dietary intake in a sample of urban and rural Chinese households and show that a 10 percent increase in the price of pork potentially reduces fat consumption by 8 percent. Energy and protein intake would both drop by 2 percent. The overall effect may be different for the poor and the rich. The potentially harmful effects on the poor of increasing the price of pork would be buffered by substitutions from other food groups, such as oil, wheat flour, and coarse grains, but concerns remain that overall nutrition would worsen. These results suggest that using price changes to alter dietary intake in a setting where overnutrition and undernutrition coexist may have mixed outcomes. A natural experiment in Poland during the economic downturn of the 1990s suggests a beneficial role for price policy in a consumer switch from animal fats to vegetable fats with lower amounts of trans fatty acids (Zatonski, McMichael, and Powles 1998). A dramatic decline in ischemic heart disease and related circulatory system diseases during the first half of the 1990s is most easily explained by the removal of consumer subsidies from foods of animal origin, the aggressive marketing of margarines, and a general decline in food purchasing power. The major change in the food supply appeared to be a reduction in foods containing animal fats; however, no direct relationship can be conclusively attributed without further study of the Polish experience and the experiences of other countries undergoing similar transitions. Governments may choose to address food-related health problems by taxing imports of high-fat or high-sugar food; however, such efforts conflict with rules governing international trade. Fiji, for example, tried to ban the import of mutton flaps, an extremely fatty food that was contributing to the country’s obesity problem. To comply with its World Trade Organization obligations, Fiji had to ban the sale of all mutton flaps, not just imports (Evans and others 2001). One analysis suggests that the same kind of broad treatment would be necessary to grant subsidies to healthy foods, but taxing unhealthy Fiscal Policies for Health Promotion and Disease Prevention | 219 ©2006 The International Bank for Reconstruction and Development / The World Bank 23 domestic foods alone would probably not pose a problem under World Trade Organization rules (WHO 2003). Furthermore, avenues for using other regulatory and economic policies to improve the consumption of healthy foods may be acceptable under the World Trade Organization Agreement on Technical Barriers to Trade and the Agreement on Agriculture if countries can justify them as contributing to legitimate national health objectives. Agricultural policies affect food prices, food choices, and farm incomes in addition to the food security of both rural and urban populations. Each country must assess the potential for reorienting its agricultural policies so as to produce a healthier food supply. Developing countries are generally more likely to directly subsidize food consumption than food production; however, they frequently make indirect subsidies available through the provision of cheap fuel, chemical inputs, water, and loans to the agriculture sector. These policies may be environmentally and fiscally costly and rarely contribute to improved population health. Research is needed on individual countries’ agricultural policies and food supply needs to make them more compatible (Nugent 2004). At the same time, the dynamics of food choice and the effects of price manipulation need to be better understood before tax and subsidy systems can be designed to effectively promote healthy food choices. Sales Tax Exemptions on Healthy and Staple Foods and Medicines and Other Health Care Goods. Governments may set tax policies to ensure that certain expenditures on healthrelated behaviors and health goods are tax deductible or tax exempt for firms, employers, or individuals. Exemptions should apply to a limited number of goods that are easily differentiated from goods that are not exempted. Note that in countries with large informal sectors, income tax systems are weak, and fiscal policies for the deductibility of credits are unlikely to be effective. South Africa provided value added tax exemptions for a short list of essential foods and found a varied consumption pattern by commodity, with the poor receiving most of the benefits of the maize exemption, but few of the benefits of the milk exemption (Alderman and del Ninno 1999). Mexico imposes a 15 percent value added tax on almost all goods. Exemptions include medicines, physician’s services, and some foods. Recently, a government proposal to make drug and food purchases eligible for value added tax and to channel the resulting revenues into financing programs targeted to the poor has given rise to extensive debate. Those in favor have argued that the existing subsidy is regressive because most drug and food purchases are by the wealthy (Fundación Mexicana para la Salud 2001). Many developing countries concerned about the spread of HIV/AIDS have dropped import taxes on condoms, but others continue to impose tariffs on imports. For example, Malaysia is a major producer and imposes a 25 percent tax on imports. Brazil used to impose both an import tax and a distribution tax that amounted to a total of 45 percent of the original condom price, but it granted a permanent sales tax exemption when condom sales increased following a temporary tax holiday. Taxes on Producers Producer taxes are usually aimed at discouraging socially harmful products or processes. They can be imposed either on the use of certain inputs, such as more heavily polluting fuels, or on their outputs, such as emissions of air pollutants. Theoretical and simulation models have examined the use of taxes on fuels and emissions taxes to control air pollution (World Bank 1994a, 1994b), but empirical data are lacking. Models of taxes suggest potential to induce substitution by cleaner fuels and reductions in overall energy use, but actual results will depend on the availability of fuel substitutes within countries. Data on Chilean manufacturing support the possibility of clean fuels substitution but indicate the likelihood of uneven sectoral incidence of the emissions tax. For example, bakeries were responsive to changes in relative prices, whereas metal products plants were unresponsive, and meat packers were unable to adjust their electricity demand but could reduce energy from other sources (World Bank 1994b). If this potential were realized on a global or regional basis—for example, through agreements to the Kyoto Protocol—a double benefit of reducing harmful externalities and raising significant revenues might be achieved. FISCAL POLICY TO PROMOTE HEALTH The fiscal policies discussed in this chapter in relation to health and health care goods can be applied to other goods and markets, such as housing and education, some of which may have important effects on health. This chapter does not provide an exhaustive discussion of the goods that indirectly promote health, but it does briefly consider some of these policies in relation to workplaces, employment leave policies, and day care. Note that policies focused on formal labor markets will not be effective in reaching large segments of the population in many countries. Policies that provide health-related services, such as day care, that are not based on formal labor market participation may have a broader effect. Workplace Health Governments can use tax relief and financial support to producers to encourage firm-specific actions to promote health. Many countries mandate safeguards in the workplace and levy penalties against occupational health violations. Most 220 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul ©2006 The International Bank for Reconstruction and Development / The World Bank 24 government actions are mandates rather than fiscal policies, but a combination of approaches may also be used. A growing area for workplace health promotion is HIV/AIDS. Bloom and others (2004) suggest that the failure on the part of most firms to act—even if they correctly perceive the business, human, and social challenges HIV/AIDS poses— is attributable to a lack of incentives. Significant externalities (benefits to society and firms) are likely to result from promoting greater action by firms. Some private firms have begun providing HIV/AIDS prevention and treatment services to employees, families, and their communities (“Face Value: AIDS and Business” 2004). Sometimes government support is involved, but little information is available to evaluate the potential of fiscal policy. Maternity Leave, Sick Leave, and Family Care Leave Government policy can alter choices regarding different types of worker leave. Many countries have financial or legislative support for caregiving, although most focus on children. Rhum (1998) cites evidence that more than 100 countries— and almost all the industrial countries—have some legislation about parental leave, although in several countries it is unpaid. Caregiving policies that allow people to take time off work to care for aged and chronically ill family members are less common than policies for child care, particularly in developing countries, but tax benefits and allowances for these types of caregiving are becoming increasingly available in the industrial countries (Brodsky, Habib, and Mizrahi 2000; Pijl 2003; Wiener 2003). Although the provision of services in kind by the government is still an important mechanism, the trend is toward empowering consumers by offering subsidies or tax deductions that allow them to choose among caregiving options. Countries tend to use a combined approach to financing that relies on payroll taxes imposed on employees and employers, general taxation, and copayments. Important issues that developing countries need to address in this respect include targeting compared with universal provision, the mechanisms to pay for or to insure care, and the extent to which long-term care should be integrated into the health care and social service systems (Brodsky, Habib, and Mizrahi 2000; WHO 2003). Day Care and Early Childhood Education Some countries use targeted fiscal policies, such as income tax deductions or direct provision, to increase the use and quality of early childhood education and child care services. Important health, labor market efficiency, growth, and equity arguments support subsidizing these services, particularly for low-income families, because without subsidies women may be forced to limit their work or to leave the labor market, and families may have to use low-quality care or leave children unattended. Van der Gaag and Tan (1997) argue for public subsidies based on cost-benefit analysis of early childhood development programs. They conclude that the greatest payoff comes from targeting the most deprived families and that the private benefits are sufficient to expect better-off parents to pay. Two large-scale, home-based day care programs targeted to poor families are Community Well-Being Homes (Hogares Comunitarios de Bienestar) run by the Colombian Institute for Family Well-Being (Instituto Colombiano de Bienestar Familiar) in Colombia (Myers 1995) and the Integrated Child Development Program in Bolivia. The former is an interesting case of a targeted cross-subsidy because the financing comes from the wealthier formal sector by means of a payroll tax, whereas the services are targeted to the poorest families. The program was 85 percent subsidized in the early 1990s. Parents paid a proportion of the caregivers’ wages on a sliding-scale user fee (Young 1996). The Bolivian program includes nutrition, health, and cognitive development interventions and is one of the few early childhood programs in developing countries that has been formally evaluated (Behrman, Cheng, and Todd 2000). The evaluation shows that the program significantly increases cognitive achievement, although the results depend on age and the duration of exposure to the program. CONCLUSIONS A broad range of experiences cited in this chapter demonstrates that fiscal and health policies interact in a number of areas. Although substantial research has focused on tobacco and alcohol, other links—for example the promotion of health in the workplace—have been much less recognized or studied, particularly in developing countries. The research presented in this chapter suggests that fiscal policy can be a useful tool for influencing health in developing countries. Nevertheless, budgetary limitations to withstand pressure for program expansion, leakages to unintended beneficiaries, public compliance with the tax system, and corruption among both government officials and the public are important factors to take into account in design and implementation. Table 11.4 summarizes some lessons learned on the use of fiscal policy to promote health. Governments may find it worthwhile to examine their use of fiscal policies to identify the entire range of effects and have health ministries participate in this exercise. More generally, the chapter indicates an area for increased interaction between ministries of health and finance. Healthy fiscal policy and fiscal policy for health should be topics that are debated, agreed on, and formalized between the two areas of policy making to guarantee that those developing fiscal policy take both its economic and its health implications into account. Fiscal Policies for Health Promotion and Disease Prevention | 221 ©2006 The International Bank for Reconstruction and Development / The World Bank 25 Table 11.4 Lessons in Using Fiscal Policy for Health Promotion Intervention choice Program design Instrument design Policy regime • Select interventions that directly • Ensure that the health benefits of • Choose the appropriate recipients • Ensure that policy is consistent address the health objective. • Ensure that interventions are sufficient to effect the health change, but not excessive. • Choose interventions with high health returns and low costs relative to alternatives. the desired change are apparent and significant. • Make sure that the tax base is adequate and stable and that no untaxed close substitutes are available. • Be aware that a large informal labor sector will limit the effectiveness and equity of benefit delivery. • Avoid programs whose expenses may become unsustainable because of uncontrollable factors. for a subsidy or tax preference. • Do not spread the benefits across too large a group. • Note that targeting by demographic, geographic, or need categories is more efficient than no targeting or self-targeting. • Be aware of the price elasticities of a taxed good so that its incidence is clear. and predictable. • Ensure that institutions carrying out a policy are open, accountable, and uncorrupted. • Consider tradeoffs between efficiency and distributional goals. • Seek non–health sector opportunities to effect health goals. Source: Authors. Rigorous evaluation studies are needed of most of the fiscal policy interventions discussed in this chapter. Such studies should address the health, fiscal, macroeconomic, and distributional effects of using fiscal policy to achieve health goals and should be performed in a range of countries with mixed public and private sector capacity to deliver health services. The studies should also examine the differing effects of policies in urban and rural settings and across income quintiles. Of particularly high priority are further studies of the results of subsidizing drugs, medical supplies, and hygiene interventions with or without education campaigns. Those areas may reveal new fiscal approaches for addressing the disease burden in developing countries. DISCLAIMER AND ACKNOWLEDGMENT The results and conclusions in this chapter are those of the authors and do not necessarily reflect the opinions of the institutions for which they work. The authors are grateful to Ana Mylena Aguilar, Hector Arreola, Ania Burczysnka, Esperanza Calleja, Marissa Courey, Monica Hurtado, Vanesa Leyva, Eugenia Rocha, Swathi Sista, and Sinaia Urrusti for excellent research assistance and to the Mexican Health Foundation’s Council on Health and Competitiveness and the National Council for Science and Technology of Mexico for support for this research. NOTES 1. The idea of healthy fiscal policy is discussed in Cook and Vlaisavljevich (1994), Joffe and Mindell (2004), and Secretaría de Salud (2001). 2. Social marketing is defined as the use of marketing principles to influence behavior for a socially desirable outcome. It provides a desirable product at an affordable price with adequate promotion and placement (that is, access). REFERENCES Adams, R. 2000. “Self-Targeted Subsidies: The Political and Distributional Impact of the Egyptian Food Subsidy System.” Economic Development and Cultural Change 49 (1): 115–36. Alderman, H. C. 2002. Price and Tax Subsidization of Consumer Goods. Social Safety Net Primer Series. Washington, DC: World Bank Institute. Alderman, H. C., and C. del Ninno. 1999. “Poverty Issues for Zero Rating VAT in South Africa.” Journal of African Economies 8 (2): 182–208. Armstrong-Schellenberg, J., S. Abdulla, R. Nathan, O. Mukasa, T. J. Marchant, N. Kikumbih, and others. 2001. “Effect of Large-Scale Social Marketing of Insecticide-Treated Nets on Child Survival in Rural Tanzania.” Lancet 357 (9264): 1241–47. Behrman, J. R., Y. Cheng, and P. Todd. 2000. The Impact of the Bolivian Integrated “PIDI” Preschool Program. Philadelphia: University of Pennsylvania. Bloom, D., L. R. Bloom, D. Steven, and M. Weston. 2004. “Business and HIV/AIDS: Who Me? A Global Review of the Business Response to HIV/AIDS.” Paper prepared for the World Economic Forum’s Global Health Initiative in partnership with the Harvard School of Public Health and Joint United Nations Programme on HIV/AIDS, World Economic Forum, Geneva. Borghi, J., L. Guinness, J. Ouedraogo, and V. Curtis. 2002. “Is Hygiene Promotion Cost-Effective? A Case Study in Burkina Faso.” Tropical Medicine and International Health 7 (11): 960–69. Brodsky, J., J. Habib, and I. Mizrahi. 2000. Long-Term Care in Five Developed Countries. Geneva: World Health Organization. Cohen, D., R. Scribner, R. Bedimo, and T. Farley. 1999. “Cost as a Barrier to Condom Use: The Evidence for Condom Subsidies in the United States.” American Journal of Public Health 89 (4): 567–68. Cook, E., and M. Vlaisavljevich. 1994. “Implications of Health Reform for State and Local Fiscal Policy.” National Tax Journal 47 (3): 639–54. del Ninno, C., and P. Dorosh. 2002. “In-Kind Transfers and Household Food Consumption: Implications for Targeted Food Programs in Bangladesh.” Discussion Paper 134, International Food Policy Research Institute, Washington, DC. Dorosh, P., C. del Ninno, and D. Sahn. 1996. “Market Liberalization and the Role of Food Aid in Mozambique.” In Economic Reform and the Poor in Africa, ed. D. Sahn. Oxford, U.K.: Clarendon Press. Evans, M., R. Sinclair, C. Fusimalohi, and V. Liav’a. 2001. “Globalization, Diet, and Health: An Example from Tonga.” Bulletin of the World Health Organization 79 (9): 856–62. 222 | Disease Control Priorities in Developing Countries | Rachel Nugent and Felicia Knaul ©2006 The International Bank for Reconstruction and Development / The World Bank 26 “Face Value: AIDS and Business.” 2004. Economist 373 (8404): 68. Fundación Mexicana para la Salud. 2001. “Política Fiscal Saludable: Propuestas a Considerar sobre la Reforma Fiscal.” Unpublished paper, Fundación Mexicana para la Salud, México, D.F. Gertler, P. 2004. “Do Conditional Cash Transfers Improve Child Health? Evidence from PROGRESA’s Control Randomized Experiment.” American Economic Review 94 (2): 336–41. Water Treatment and Safe Storage: A Promising New Strategy.” Epidemiology Infection 122 (1): 83–90. Rao, V. 2000. “Price Heterogeneity and Real Inequality: A Case Study of Poverty and Prices in Rural South India.” Review of Income and Wealth 46 (2): 201–11. Rhum, C. 1998. “Parental Leave and Child Health.” NBER Working Paper 6554, National Bureau of Economic Research, Cambridge, MA. Guo, X., B. Popkin, T. Mroz, and F. Zhai. 1999. “Food Price Policy Can Favorably Alter Macronutrient Intake in China.” Journal of Nutrition 129 (5): 994–1001. Secretaría de Salud. 2001. “Programa Nacional de Salud: 2001–2006— La democratización de la salud en México: Hacia un sistema universal de salud.” Secretaría de Salud, México, D.F. Hanson, K., and E. Worrall. 2002. Report on the Analysis of SMITN2 Endof-Project Survey. London: London School of Hygiene and Tropical Medicine. Shah, A., ed. 1995. Fiscal Incentives for Investment and Innovation. New York: Oxford University Press. Hoque, B. 2003. “Handwashing Practices and Challenges in Bangladesh.” International Journal of Environmental Health Research 13 (Suppl. 1): S81–87. Jha, P. 1999. Curbing the Epidemic: Governments and the Economics of Tobacco Control. Washington, DC: World Bank. Jha, S. 1992. “Consumer Subsidies in India: Is Targeting Effective?” Development and Change 23 (4): 101–28. Joffe, M., and Mindell, J. 2004. “A Tentative Step towards Healthy Public Policy.” Journal of Epidemiology and Community Health 58: 966–68. Kolaczinski, J., N. Muhammad, Q. Khan, Z. Jan, N. Rehman, T. Leslie, and others. 2004. “Subsidized Sales of Insecticide-Treated Nets in Afghan Refugee Camps Demonstrate the Feasibility of a Transition from Humanitarian Aid Towards Sustainability.” Malaria Journal 3 (15): 1–11. Luby, S. P., M. Agboatwalla, A. Raza, J. Sobel, E. Mintz, K. Baier, and others. 2001. “Microbiologic Effectiveness of Hand Washing with Soap in an Urban Squatter Settlement, Karachi, Pakistan.” Epidemiological Infection 127 (2): 237–44. Musgrove, P. 1993. “Feeding Latin America’s Children.” World Bank Research Observer 8 (1): 23–45. Siamwalla, A. 1988. “Some Macroeconomic Policy Implications of Consumer-Oriented Food Subsidies.” In Food Subsidies in Developing Countries: Costs, Benefits and Policy Options, ed. P. Anderson. Baltimore: Johns Hopkins University Press, 323–30. Simon, J., B. Larson, A. Zusman, and S. Rosen. 2002. “How Will the Reduction of Tariffs and Taxes on Insecticide-Treated Bednets Affect Household Purchases?” Bulletin of the World Health Organization 80 (11): 892–99. Snow, R. W., E. McCabe, D. Mbogo, C. Molyneux, V. Mung’ala, and C. Nevill. 1999. “The Effect of Delivery Mechanisms on the Uptake of Bed Net Re-impregnation in Kilifi District, Kenya.” Health Policy and Planning 14 (1): 18–25. Tabor, R., and M. Sawit. 2001. “Social Protection via Rice: The OPK Rice Subsidy Program in Indonesia.” Developing Economies 39 (3): 267–94. UNDP (United Nations Development Programme). 2003. World Bank Energy Sector Management Assistance Program—India: Access of the Poor to Clean Household Fuels. New York: UNDP. http://www. worldbank.org/esmap/. Van der Gaag, J., and J. P. Tan. 1997. The Benefits of Early Child Development Programs: An Economic Analysis. Washington, DC: World Bank. Myers, R. 1995. Early Childhood Care and Development Programs in Latin America and the Caribbean: A Review of Experience. Washington, DC: Inter-American Development Bank. http://www.ecdgroup.com/ download/gw1eccdl.pdf. WHO (World Health Organization). 2002. Alcohol in Developing Societies: A Public Health Approach. Geneva: WHO. Nugent, R. 2004. “Food and Agricultural Policy in the Prevention of Non-communicable Diseases.” Food and Nutrition Bulletin 25 (2): 200–8. ———. 2004. Strategic Plan for Diet, Nutrition, and Physical Activity. Geneva: World Health Assembly. OECD (Organisation for Economic Co-operation and Development). 2003. Tax Incentives for Research and Development: Trends and Issues. Paris: Science Technology Industry. Pijl, M. 2003. “The Support of Carers and Their Organizations in Some Northern and Western European Countries.” In Key Policy Issues in Long-Term Care: World Health Organization Collection on Long-Term Care, ed. J. Brodsky, J. Habib, M. J. Hirschfeld. Geneva: World Health Organization. Pinstrup-Anderson, P. 1988. Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options. Washington, DC: International Food Policy Research Institute; Baltimore, MD: Johns Hopkins University Press. Price, N. 2001. “The Performance of Social Marketing in Reaching the Poor and Vulnerable in AIDS Control Programs.” Health Policy and Planning 16 (3): 231–39. Quick, R., A. Kimura, A. Thevos, M. Tembo, I. Shamputa, L. Hutwanger, and E. Mintz. 2002. “Diarrhea Prevention through Household-Level Water Disinfection and Safe Storage in Zambia.” American Journal of Tropical Medicine 66 (5): 584–89. Quick, R., L. Venczel, E. Mintz, L. Soleto, J. Aparicio, M. Gironaz, and others. 1999. “Diarrhea Prevention in Bolivia through Point-of-Use ———. 2003. Using Domestic Law in the Fight against Obesity: An Introductory Guide for the Pacific. Geneva: WHO. Wiener, J. 2003. “The Role of Informal Support in Long-Term Care.” In Key Policy Issues in Long-Term Care. World Health Organization Collection on Long-Term Care, ed. J. Brodsky, J. Habib, M. J. Hirschfeld. Geneva: World Health Organization. World Bank. 1994a. Energy Pricing and Air Pollution: Econometric Evidence from Manufacturing in Chile and Indonesia. Washington, DC: World Bank. ———. 1994b. How Relative Prices Affect Fuel Use Patterns in Manufacturing: Plant-Level Experience from Chile. Washington, DC: World Bank. Yaser, Y. 1993. “Extensive Advertising.” Integration 32: 32–33. Young, M. E. 1996. Desarrollo del Niño en la Primera Infancia: Una Inversión en el Futuro. Washington, DC: World Bank, Human Resources Department. http://web.worldbank.org/WBSITE/EXTERNAL/ TOPICS/EXTEDUCATION/EXTECD/. Zatonski, W. A., A. J. McMichael, and J. W. Powles. 1998. “Ecological Study of Reasons for Sharp Decline in Mortality from Ischaemic Heart Disease in Poland since 1991.” British Medical Journal 316 (7137): 1047–51. Zee, H., J. Stotsky, and E. Ley. 2002. “Tax Incentives for Business Investment: A Primer for Policymakers in Developing Countries.” World Development 30 (9): 1497–516. Fiscal Policies for Health Promotion and Disease Prevention | 223 ©2006 The International Bank for Reconstruction and Development / The World Bank 27 ©2006 The International Bank for Reconstruction and Development / The World Bank 28 Chapter 42 Indoor Air Pollution Nigel Bruce, Eva Rehfuess, Sumi Mehta, Guy Hutton, and Kirk Smith Access to modern energy sources has been described as a “necessary, although not sufficient, requirement for economic and social development” (IEA 2002). It is, therefore, of great concern that almost half the world’s population still relies for its everyday household energy needs on inefficient and highly polluting solid fuels, mostly biomass (wood, animal dung, and crop wastes) and coal. The majority of households using solid fuels burn them in open fires or simple stoves that release most of the smoke into the home. The resulting indoor air pollution (IAP) is a major threat to health, particularly for women and young children, who may spend many hours close to the fire. Furthermore, the reliance on solid fuels and inefficient stoves has other, far-reaching consequences for health, the environment, and economic development. NATURE, CAUSES, AND BURDEN OF CONDITION About 3 billion people still rely on solid fuels, 2.4 billion on biomass, and the rest on coal, mostly in China (IEA 2002; Smith, Mehta, and Feuz 2004). There is marked regional variation in solid fuel use, from less than 20 percent in Europe and Central Asia to 80 percent and more in Sub-Saharan Africa and South Asia. This issue is inextricably linked to poverty. It is the poor who have to make do with solid fuels and inefficient stoves, and many are trapped in this situation: the health and economic consequences contribute to keeping them in poverty, and their poverty stands as a barrier to change. Where socioeconomic circumstances improve, households generally move up the energy ladder, carrying out more activities with fuels and appliances that are increasingly efficient, clean, convenient, and more expensive. The pace of progress, however, is extremely slow, and for the poorest people in Sub-Saharan Africa and South Asia, there is little prospect of change. Illustrated in figures 42.1 and 42.2 are findings for Malawi and Peru, respectively, from Demographic and Health Surveys (ORC Macro 2004). The examples are selected from available national studies with data on main cooking fuel use to represent the situation in poor African and South American countries. The main rural and urban cooking fuels are illustrated in figures 42.1a and 42.2a; the findings are then broken down nationally by level of education of the principal respondent (woman of childbearing age) in figures 42.1b and 42.2b, and in urban areas by her level of education in figures 42.1c and 42.2c. Biomass is predominantly, though not exclusively, a rural fuel: indeed, in many poor African countries, biomass is the main fuel for close to 100 percent of rural homes. Marked socioeconomic differences (indicated by women’s education) exist in both urban and rural areas. During the 1990s, use of traditional fuels (biomass) in Sub-Saharan Africa increased as a percentage of total energy use, although in most other parts of the world the trend has generally been the reverse (World Bank 2002). In many poorer countries, the increase in total energy use accompanying economic development has occurred mainly through increased consumption of modern fuels by better-off minorities. In Sub-Saharan Africa, however, the relative increase in biomass use probably reflects population growth in rural and poor urban areas against a background of weak (or negative) national economic growth. Reliable data on trends in 793 ©2006 The International Bank for Reconstruction and Development / The World Bank 29 a. Primary household fuel use in urban and rural areas a. Primary household fuel use in urban and rural areas Wood, straw Wood, straw, dung Charcoal Charcoal Kerosene Kerosene Electricity Electricity, gas 0 20 40 60 Percentage Urban 80 100 0 20 Rural 40 60 Percentage Urban 80 Rural b. Primary household fuel use, by level of education of respondent b. Primary household fuel use, by level of education of respondent Wood, straw Wood, straw, dung Charcoal Charcoal Kerosene Kerosene Electricity Electricity, gas 0 20 40 60 Percentage Primary or less 80 100 0 Secondary or higher 20 40 60 Percentage Primary or less 80 c. Primary household fuel use in urban areas, by level of education of respondent Wood, straw Wood, straw, dung Charcoal Charcoal Kerosene Kerosene Electricity Electricity, gas 20 40 60 Percentage Primary or less 80 100 0 Secondary or higher 20 40 60 Percentage Primary or less 100 Secondary or higher c. Primary household fuel use in urban areas, by level of education of respondent 0 100 80 100 Secondary or higher Source: Unpublished data derived from Demographic and Health Survey. Source: Unpublished data derived from Demographic and Health Survey. Figure 42.1 Patterns of Household Fuel Use in Malawi, 2000 Figure 42.2 Patterns of Household Fuel Use in Peru, 2000 household energy use are not available for most countries. Information is available from India, where the percentage of rural homes using firewood fell from 80 percent in 1993–94 to 75 percent in 1999–2000 (D’Sa and Narasimha Murthy 2004). Nationally, liquid petroleum gas (LPG) use increased from 9 to 16 percent over the same period, with a change from 2 percent to 5 percent in rural areas, and it is expected to reach 36 percent nationally and 12 percent for rural homes by 2016. International Energy Agency projections to 2030 show that, although a reduction in residential biomass use is expected in most developing countries, in Africa and South Asia the decline will be small, and the population relying on biomass will 794 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 30 increase from 2.4 billion to 2.6 billion, with more than 50 percent of residential energy consumption still derived from this source(OECD and IEA 2004). The number of people without access to electricity is expected to fall from 1.6 billion to 1.4 billion. Because electricity is used by poor households for lighting and not as a cleaner substitute for cooking, electrification will not, at least in the short to medium term, bring about substantial reductions in IAP. Levels of Pollution and Exposure Biomass and coal smoke emit many health-damaging pollutants, including particulate matter (PM),1 carbon monoxide (CO), sulfur oxides, nitrogen oxides, aldehydes, benzene, and polyaromatic compounds (Smith 1987). These pollutants mainly affect the lungs by causing inflammation, reduced ciliary clearance, and impaired immune response (Bruce, PerezPadilla, and Albalak 2000). Systemic effects also result, for example, in reduced oxygen-carrying capacity of the blood because of carbon monoxide, which may be a cause of intrauterine growth retardation (Boy, Bruce, and Delgado 2002). Evidence is emerging, thus far only from developed countries, of the effects of particulates on cardiovascular disease (Pope and others 2002, 2004). Saksena, Thompson, and Smith (2004) have recently compiled data on several of the main pollutants associated with various household fuels from studies of homes in a wide range of developing countries. Concentrations of PM10, averaged over 24-hour periods, were in the range 300 to 3,000 (or more) micrograms per cubic meter (␮g/m3). Annual averages have not been measured, but because these levels are experienced almost every day of the year, the 24-hour concentrations can be taken as a reasonable estimate. By comparison, the U.S. Environmental Protection Agency’s annual air pollution standard for PM10 is 50 ␮g/m3, one to two orders of magnitude lower than levels seen in many homes in developing countries. During cooking, when women and very young children spend most time in the kitchen and near the fire, much higher levels of PM10 have been recorded—up to 30,000 ␮g/m3 or more. With use of biomass, CO levels are generally not as high in comparison, typically with 24-hour averages of up to 10 parts per million (ppm), somewhat below the World Health Organization (WHO) guideline level of 10 ppm for an eighthour period of exposure. Much higher levels of CO have been recorded, however. For example, a 24-hour average of around 50 ppm was found in Kenyan Masai homes (Bruce and others 2002), and one Indian study reported carboxyhemoglobin levels similar to those for active cigarette smokers (Behera, Dash, and Malik 1988). The health effects of chronic exposure of young children and pregnant women to levels of CO just below current WHO guidelines have yet to be studied. For additional information on levels of other pollutants in biomass and coal smoke, see Saksena, Thompson, and Smith (2004). Fewer studies of personal exposure have been done than of area pollution, mainly because measurement of personal PM typically requires wearing a pump, a cumbersome procedure. CO can be measured more easily and has been used as a proxy: time-weighted (for example, 24-hour average) CO correlates well with PM if a single main biomass stove is used (Naeher and others 2001). Time-activity and area pollution information can also be combined to estimate personal exposure (Ezzati and Kammen 2001). These various methods indicate that personal 24-hour PM10 exposures for cooks range from several hundred ␮g/m3 to more than 1,000 ␮g/m3 (Ezzati and Kammen 2001), with even higher exposures during cooking (Smith 1989). Few studies have measured personal PM exposures of very young children: one study in Guatemala found levels a little lower than those of their mothers (Naeher, Leaderer, and Smith 2000). Health Impacts of IAP A systematic review of the evidence for the impact of IAP on a wide range of health outcomes has recently been carried out (Smith, Mehta, and Feuz 2004; see table 42.1). This review identified three main outcomes with sufficient evidence to include in the burden-of-disease calculations and a range of other outcomes with as yet insufficient evidence. Studies for the key outcomes used in the burden-of-disease calculations—acute lower respiratory infection (ALRI), chronic obstructive pulmonary disease (COPD), and lung cancer—had to be primary studies (not reviews or reanalyses), written or abstracted in English (and for lung cancer, Chinese), that reported an odds ratio and variance (or sufficient data to estimate them) and provided some proxy for exposure to indoor smoke from the use of solid fuels for cooking and heating purposes. A limitation of almost all studies has been the lack of measurement of pollution or exposure: instead, proxy measures have been used, including the type of fuel or stove used, time spent near the fire, and whether the child is carried on the mother’s back during cooking. The studies do not, therefore, provide data on the exposure-response relationship, although a recent study from Kenya has gone some way to addressing this omission (Ezzati and Kammen 2001). In some countries, household fuels carry locally specific risks. It has been estimated that more than 2 million people in China suffer from skeletal fluorosis, in part resulting from use of fluoride-rich coal (Ando and others 1998). Arsenic, another contaminant of coal, is associated with an increased risk of lung cancer in China (Finkelman, Belkin, and Zheng 1999). There has been concern, however, that reducing smoke could increase risk of vectorborne disease, including malaria. Some Indoor Air Pollution | 795 ©2006 The International Bank for Reconstruction and Development / The World Bank 31 Table 42.1 Status of Evidence Linking Biomass Fuels and Coal with Child and Adult Health Outcomes Health outcome Age Status of evidence Sufficient evidence for burden-of-disease calculation Acute lower respiratory infections Children ⬍ 5 years Chronic obstructive pulmonary disease Adult women Lung cancer (coal exposure) Adult women Chronic obstructive pulmonary disease Adult men Lung cancer (coal exposure) Adult men Strong. Some 15–20 observational studies for each condition, from developing countries. Evidence is consistent (significantly elevated risk in most though not all studies); the effects are sizable, plausible, and supported by evidence from outdoor air pollution and smoking. Moderate-I. Smaller number of studies, but consistent and plausible. Not yet sufficient evidence for burden-of-disease calculation Lung cancer (biomass exposure) Adult women Tuberculosis Adult Asthma Child and adult Cataracts Adult Adverse pregnancy outcomes Perinatal Cancer of upper aerodigestive tract Adult Interstitial lung disease Adult Ischemic heart disease Adult Moderate-II. Small number of studies, not all consistent (especially for asthma, which may reflect variations in definitions and condition by age), but supported by studies of outdoor air pollution, smoking, and laboratory animals. Tentative. Adverse pregnancy outcomes include low birthweight and increased perinatal mortality. One or a few studies at most for each of these conditions, not all consistent, but some support from outdoor air pollution and passive-smoking studies. Several studies from developed countries have shown increased risk for exposure to outdoor air pollution at much lower levels than IAP levels seen in developing countries. As yet, no studies from developing countries. Source: Smith, Mehta, and Feuz 2004. studies have shown that biomass smoke can repel mosquitoes and reduce biting rates (Palsson and Jaenson 1999; Paru and others 1995; Vernede, van Meer, and Alpers 1994). Few studies have examined the impact of smoke on malaria transmission: one from southern Mexico found no protective effect of smoke (adjusted odds ratio 1.06 [0.72–1.58]; Danis-Lozano and others 1999), and another from The Gambia found that wood smoke did not protect children in areas of moderate transmission (Snow and others 1987). Method Used for Determining Attributable Disease Burden Smith, Mehta, and Feuz (2004) have provided a full explanation of the calculation of the disease burden associated with IAP. Summarized here are the methods they used to estimate the two most critical components of these calculations: the number of people exposed and the relative risks. Exposure. The absence of pollution or exposure measurement in health studies required use of a binary classification: the use or nonuse of solid fuels. The authors obtained estimates of solid fuel use for 52 countries from a range of sources, mostly household surveys, and statistical modeling was used for countries with no data (the majority) (Smith, Mehta, and Feuz 2004). They assumed, conservatively, that all countries with a 1999 per capita gross national product (GNP) greater than US$5,000 had made a complete transition either to electricity or cleaner liquid and gaseous fuels or to fully ventilated solid fuel devices. To account for differences in exposure caused by variation in the quality of stoves, they applied a ventilation factor (VF), set from 1 for no ventilation to 0 for complete ventilation. In China, a VF of 0.25 was used for child health outcomes and 0.5 for adult outcomes, reflecting a period of higher exposure (to open fires) before the widespread introduction of chimney stoves. Countries with a 1999 GNP per capita greater than US$5,000 were assigned a VF of 0, and all other countries a value of 1, reflecting the very low rates of use of clean fuels or effective ventilation technologies. The authors obtained the final point estimate for exposure by multiplying the percentage of solid fuel use by the VF. They arbitrarily assigned an uncertainty range of ⫾5 percent to the estimates. Risk. Smith, Mehta, and Feuz (2004) carried out metaanalyses for the three health outcomes with sufficient evidence (table 42.2). They used fixed-effects models and sensitivity analysis that took account of potential sources of heterogeneity, including the way in which exposure was defined and whether adjustment had been made for confounders (Smith, Mehta, and Feuz 2004). 796 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 32 Table 42.2 Summary of Relative Risk Estimates for Health Outcomes Used in Burden-of-Disease Estimates Age and sex group ALRI Children ⬍ 5 years 8 2.3 1.9–2.7 COPD Women ⬎ 30 years 8 3.2 2.3–4.8 Lung cancer (coal) Number of studies Relative risk 95 percent confidence interval Health outcome Men ⬎ 30 yearsa 2 1.8 1.0–3.2 Women ⬎ 30 years 9 1.9 1.1–3.5 Men ⬎ 30 years 3 1.5 1.0–2.5 Sources: Smith, Mehta, and Feuz 2004. a. Because of the limited quantity and quality of available evidence, the male COPD relative risk and range have been fixed to include 1.0 (no effect) as the lower estimate. The Burden of Disease from Solid Fuel Use Information on the proportions exposed and risk of key disease outcomes was combined with total burden-of-disease data to obtain the population attributable fractions associated with IAP (WHO 2002b). Globally, solid fuels were estimated to account for 1.6 million excess deaths annually and 2.7 percent of disability-adjusted life years (DALYs) lost, making them the second most important environmental cause of disease, after contaminated water, lack of sanitation, and poor hygiene (table 42.3). Approximately 32 percent of this burden (DALYs) occurs in Sub-Saharan Africa, 37 percent in South Asia, and 18 percent in East Asia and the Pacific. In developing countries with high child and adult mortality, solid fuel use is the fourth most important risk factor behind malnutrition, unsafe sex, and lack of water and sanitation, and it is estimated to account for 3.7 percent of DALYs lost (WHO 2002b). Overall, there are more female deaths but similar numbers of male and female DALYs (table 42.3b). The reason can be found by looking further at the health outcomes. Deaths and DALYs from ALRI in children under five years of age are slightly greater for males (table 42.3c). Women experience twice the DALYs and three times the deaths from COPD (male smoking-attributable COPD deaths excluded). Far fewer cases of lung cancer are attributable to IAP, but women experience about three times the burden of men. Table 42.3 also shows how the poorest regions of the world carry by far the greatest burden, particularly for ALRI. More than half of all the deaths and 83 percent of DALYs lost attributable to solid fuel use occur as a result of ALRI in children under five years of age. In high-mortality areas, such as SubSaharan Africa, these estimates indicate that approximately 30 percent of mortality and 40 percent of morbidity caused by ALRI can be attributed to solid fuel use, as can well over half of the deaths from COPD among women. Because they derive from WHO risk assessments, these estimates include age weights, such that years of life lost at very young or advanced ages count less than years lost in the prime of adult life. Age weighting makes little difference to the DALYs lost per death up to age five; how much it affects the DALY cost of adult deaths depends on the age distribution of deaths from COPD. Because these are likely to occur at age 45 or beyond, the DALY losses are underestimated compared with estimates without age weighting that follow the usual practice in this volume. Other Effects of Household Energy Use in Developing Countries A number of other health impacts—for example, burns from open fires—were not assessed because the burden-of-disease assessment process allowed inclusion of only those health effects resulting directly from pollution. Children are at risk of burns and scalds, resulting from falling into open fires and knocking over pots of hot liquid (Courtright, Haile, and Kohls 1993; Onuba and Udoidiok 1987). Modern fuels are not always safe either, because children are also at risk of drinking kerosene, which is often stored in soft drink bottles (Gupta and others 1998; Reed and Conradie 1997; Yach 1994). Families—mainly the women and children—can spend many hours each week collecting biomass fuels, particularly where environmental damage and overpopulation have made them scarce. This time could be spent more productively on child care and household or income-generating tasks. There are also risks to health from carrying heavy loads and dangers from mines, snake bites, and violence (Wickramasinghe 2001). Inefficient stoves waste fuel, draining disposable income if fuel is bought. Although women carry out most of the household activities requiring fuels, they often have limited control over how resources can be spent to change the situation (Clancy, Skutsch, and Batchelor 2003). These conditions can combine to restrict income generation from home-based activities that require fuel energy (for example, processing and preparing food for sale). Homes that are heavily polluted and dark can hinder productivity of householders, including children doing homework and others engaged in home-based income-generating activities such as handicrafts. In many poor homes, lighting is obtained from the open fire and simple kerosene wick lamps, which provide poor light and add to pollution. Solid fuel use has important environmental consequences. Domestic use of solid fuels in high-density rural and urban environments contributes to outdoor air pollution. Many lowincome urban populations rely on charcoal, the production of which can place severe stress on forests. The use of wood as fuel can contribute to deforestation, particularly where it is combined with population pressure, poor forest management, and clearance of land for agriculture and building timber. Damage to forest cover can increase the distance traveled to obtain wood and can result in the use of freshly cut (green) wood, dung, and Indoor Air Pollution | 797 ©2006 The International Bank for Reconstruction and Development / The World Bank 33 Table 42.3 Deaths and DALYs Lost Because of Solid Fuel Use a. Overall Deaths (thousands) World Bank region DALYs (thousands) Total burden (percent) East Asia and the Pacific 540 7,087 18.4 Europe and Central Asia 21 544 1.4 26 774 2.0 118 3,572 9.3 Latin America and the Caribbean Middle East and North Africa South Asia 522 14,237 36.9 Sub-Saharan Africa 392 12,318 32.0 1,619 38,532 100.0 World b. All causes, by sex Deaths (thousands) World Bank region Male Female 152 388 East Asia and the Pacific Europe and Central Asia Latin America and the Caribbean Middle East and North Africa DALYs (thousands) All Male Female All 540 3,028 4,060 7,087 9 13 21 251 293 544 12 14 26 368 405 774 57 61 118 1,849 1,724 3,572 South Asia 218 304 522 6,641 7,596 14,237 Sub-Saharan Africa 211 181 392 6,901 5,417 12,318 World 658 961 1,619 19,037 19,495 38,532 c. From ALRI (children under age five) Deaths (thousands) World Bank region Male Female DALYs (thousands) All Male Female All East Asia and the Pacific 40 41 81 1,502 1,535 3,036 Europe and Central Asia 7 6 13 235 204 439 8 7 15 324 281 605 51 44 95 1,794 1,571 3,365 Latin America and the Caribbean Middle East and North Africa South Asia 177 178 355 6,228 6,278 12,506 Sub-Saharan Africa 198 153 351 6,777 5,191 11,967 World 481 429 910 16,860 15,058 31,918 d. From COPD (men and women 30 years and over) Deaths (thousands) World Bank region East Asia and the Pacific DALYs (thousands) Male Female All Male Female All 105 338 443 1,461 2,430 3,891 Europe and Central Asia 2 7 9 16 89 104 Latin America and the Caribbean 4 7 11 44 125 168 Middle East and North Africa South Asia Sub-Saharan Africa World 6 17 23 55 153 208 41 126 167 410 1,314 1,724 13 28 41 124 227 351 171 522 693 2,110 4,336 6,446 Source: Modified by authors to World Bank regions, from Smith, Mehta, and Feuz 2004. 798 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 34 twigs, which are more polluting and less efficient. In some urban communities, poverty and supply problems are resulting in the use of plastic and other wastes for household fuel (IEA 2002). Stoves with inefficient combustion produce relatively more products of incomplete combustion, such as methane, which have a markedly higher global-warming potential than carbon dioxide (Smith, Uma, and others 2000). It has, therefore, been argued that, although the energy use and greenhouse gas emissions from homes in developing countries are small relative to the emissions generated in industrial countries, cleaner and more efficient energy systems could provide the double benefit of reduced greenhouse gas emissions (with opportunities for carbon trading) and improved health through reduced IAP (Wang and Smith 1999). The evidence available for assessing these effects, which together could have a substantial influence on health and economic development, is patchy at best. This area is important for research (Larson and Rosen 2002). INTERVENTIONS AND POLICY The uses of energy in the home—for example, for cooking and keeping warm and as a focus of social activities—have important attributes that are specific to the locality, culture, and individual households and are often associated with established traditions and deeply held beliefs. Encouraging the use of cleaner and more efficient energy technologies by populations that are among the poorest in the world has not been easy, but recent years have seen progress being made with respect to suitable technology that meets the needs of households and with respect to the development of supportive policy. Poverty Reduction and the Millennium Development Goals Given the close relationship between socioeconomic conditions and solid fuel use, poverty reduction must be a key element of policy to alleviate IAP. The United Nations Millennium Development Goals set targets for poverty eradication, improvements in health and education, and environmental protection; they represent the accepted framework for the world community to achieve measurable progress (United Nations Statistics Division 2003). Although reducing IAP can contribute to achieving a number of these goals, it is particularly relevant to reducing child mortality (Goal 4) from ALRI. Goal 7, Target 9, aims at integrating sustainable development into country policies and programs. The proportion of population using solid fuels has been adopted as an indicator for Target 9. Alleviating drudgery resulting from collecting fuel and using inefficient stoves, together with the involvement of women in implementing changes, can promote gender equality and empower women (Goal 3). Household energy interventions can also contribute to eradicating extreme poverty (Goal 1) through health improvements, time saving, and better environments for education and facilitating income generation (WHO 2004a). Interventions Although the main focus of this chapter is IAP, the many other ways in which household energy can affect health and development emphasize why interventions should aim to achieve a range of benefits, including the following: • reduced levels of IAP and human exposure • increased fuel efficiency • reduced time spent collecting fuel and using inefficient stoves • reduced stress on the local environment • increased opportunities for income generation • contribution to an overall improvement in the quality of the home environment—in particular, the working environment and conditions for women. Interventions for reducing IAP can be grouped under three headings: those acting on the source of pollution, those improving the living environment (aspects of the home), and changes to user behaviors (table 42.4). It should not be assumed that an intervention that reduces IAP will necessarily achieve other aims listed previously. For example, in colder areas, an enclosed stove with a flue that reduces IAP may reduce radiant heat and light, forcing households to use other fuels for those purposes. If not addressed with households, such problems may well result in disappointing reductions in IAP exposure, poor acceptance of interventions, and lack of motivation to maintain them. Policy Instruments Although a range of interventions is available, poor households face many barriers to their adoption, and enabling policy is needed (table 42.5). This area of practice is complex and evolving, often requiring solutions that are highly setting specific. INTERVENTION COSTS AND EFFECTIVENESS The cost-effectiveness analysis discussed in this chapter is based on recent work by Mehta and Shahpar (2004). The key components of this analysis are described here, with particular emphasis on the underlying assumptions. Indoor Air Pollution | 799 ©2006 The International Bank for Reconstruction and Development / The World Bank 35 Table 42.4 Interventions for Reducing Exposure to IAP Source of pollution Living environment User behaviors Improved cooking devices Improved ventilation Reduced exposure through operation of source • Improved biomass stoves without flues • Hoods, fireplaces, and chimneys built into the • Fuel drying structure of the house • Improved stoves with flues attached • Windows and ventilation holes (such as in Alternative fuel-cooker combinations roof), which may have cowls to assist extraction • Briquettes and pellets • Charcoal Kitchen design and placement of the stove • Kerosene • Kitchen separate from house to reduce • Liquid petroleum gas exposure of family (less so for cook) • Biogas, producer gas • Using pot lids to conserve heat • Properly maintaining stoves and chimneys and other appliances Reductions by avoiding smoke • Keeping children away from smoke—for example, in another room (if available and safe to do so) • Stove at waist height to reduce direct exposure of cook leaning over fire • Solar cookers (thermal) • Other low-smoke fuels • Electricity Reduced need for the fire • Insulated fireless cooker (haybox) • Efficient housing design and construction • Solar water heating Source: Modified from Ballard-Tremeer and Mathee 2000. Table 42.5 Policy Instruments for Promoting Implementation of Effective Household Energy Interventions Policy instruments Examples Applications Information, education, and communication Schools Learning about household energy, health, and development should be integrated in school curricula, particularly in countries where these topics are a priority for health and economic development. This goal can be achieved through programs such as the WHO Global School Health Initiative, which promotes environmental health education, including education about IAP. Media Local and national radio, television, and newspapers can be used to raise awareness and disseminate information on technologies and opportunities to support implementation, such as promotions and microcredit. These media can be directed at a range of audiences, including decision makers, professionals, and the public where radio is widely used. Community education Opportunities such as adult literacy programs can be used to raise awareness and share experience of interventions, and innovative methods can be used (for example theater). Tax on fuels and appliances Reduced tax on fuels and appliances may promote development of distribution networks and uptake, and it may be seen as efficient if there is evidence of health, education, and economic benefits. Subsidy on fuels and appliances General (for example, national) subsidies on fuels such as kerosene have been applied to promote use by poor households. Subsidies have been found to be inefficient instruments, however, often benefiting the better off rather than the poor. Time-limited subsidy on specific products (for example, clean fuel appliances, connection to grid) may be a useful method for promoting initial uptake, generating demand, and thereby providing market conditions for lower prices and more consistent quality. Air quality standards Although some developing countries have air quality standards for urban air, none have them for indoor air in settings where solid fuels are widely used. Routine monitoring and enforcement is not practical, but it may be useful to set standards and targets linked to specific assessments. For more routine use, information from censuses and surveys, such as fuel type, stove type, and venting for smoke, offers a practical alternative for setting air quality standards for IAP in developing countries. Design standards for appliances Design standards can be applied to safety (prevention of burns, gas leaks, and explosions); venting of emissions; and efficiency. Although such standards may be difficult to enforce in an informal economy, they could become valuable with wider-scale production. Public program provision of appliances Large-scale public provision of appliances, such as improved stoves or clean-fuel appliances, has generally been found unsuitable. Some form of targeted provision or partial subsidy where households have made informed choices and commit to cost sharing may be useful to stimulate demand and act in favor of equity. Taxes and subsidies Regulation and legislation Direct expenditures 800 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 36 Table 42.5 Continued Policy instruments Research and development Examples Applications Funding of finance schemes Experience has shown that credit is most likely to be made available and adopted for energy applications that contribute directly to productive, income-generating activities (such as food processing for sale). Meeting everyday cooking and space-heating needs is seen as a lower priority. Good opportunities may exist where biomass fuel is purchased and where cost saving combines with other valued benefits, such as increased prestige and cleaner kitchens. Support for such schemes, mainly in the form of raising awareness, skills training in managing funds, and seed funding (the main source of funds being from users) may be cost-effective. Surveys Surveys of fuel and appliance use, knowledge of risks to health, willingness to pay for interventions, knowledge of and confidence in credit schemes, and the like are important for planning interventions. Development and evaluation of interventions Evaluation of interventions should be conducted in a range of settings, using harmonized methods, if possible, that allow local flexibility but permit comparison with other types of interventions and other locations. Studies of health effects Stronger and better-quantified evidence of the effects on health of reducing IAP, which includes exposure measurement, is required not only for key outcomes such as ALRI, but also for other health outcomes for which evidence is currently tentative. Research capacity development Capacity for carrying out a wide range of research—from national and local surveys, to monitoring and evaluation of interventions, to more complex health studies—requires strengthening in those countries where the problems associated with household energy and IAP are most pressing. Source: Authors. Costs Intervention costs have a number of components, the relative importance of which will vary with the type of fuel and device (box 42.1). The level of costs incurred by consumers and others, including government, depends not only on the type of intervention but also on how it is delivered, supplied, and adopted. Experience indicates that successful interventions are sustainable in local markets, implying that the consumer pays the majority of initial and recurrent costs. The contributions of the government, utilities, nongovernmental organizations (NGOs), and the commercial sector will depend on many factors, including the type of intervention and fuel, location (urban or rural), existing level of supply and distribution Box 42.1 Cost Components for Household Energy Interventions • Fuels, which vary from zero (in direct cash terms, though not in opportunity cost) for collected biomass to a U.S. dollar or so per week for kerosene and several U.S. dollars per week for electricity (where used for cooking). • Stove appliances, which vary from zero for a simple three-stone fire (stones arranged on the floor to support cooking pots, with the fire lit between the stones), to US$50 (and in some cases more than US$100) for a good-quality woodstove with a chimney and up to several hundred U.S. dollars for a biogas installation. Source: Authors. • Additional appliances—for example, an LPG storage bottle has a moderately high initial cost but should last for many years. • Maintenance costs, which vary from zero for a three-stone fire up to modest, but not negligible, costs of repairing (and periodically replacing) woodstoves and chimneys. Appliances for using kerosene, LPG, and electricity also require maintenance and periodic replacement. • Program costs, which apply to various aspects of provision of energy services, particularly LPG and electricity, but may also include costs of, for example, establishing more sustainable biomass reserves and administrative costs. Indoor Air Pollution | 801 ©2006 The International Bank for Reconstruction and Development / The World Bank 37 Box 42.2 Cost Issues in Switching to Cleaner Fuels for a “Typical” Poor Kenyan Family Ruth1 and her family live 3 kilometers from a small town on the main road about one hour by bus from Kisumu. They are subsistence farmers, with a small income from selling vegetables, from irregular laboring work obtained by her husband, and from making and selling handicrafts. Ruth, a mother of five, cooks over a three-stone fire using mostly wood, which she collects every other day from plots up to two hours walking distance from home. She spends 8 to 12 hours each week collecting wood. Ruth and her family use about 2 liters of kerosene each week for wick lamps and for cooking. They use dry cell batteries for the radio; grid electricity runs nearby, but connection is far too expensive. In all, the family spends an equivalent of US$1 to US$2 per week on fuel and batteries. Through her women’s group, Ruth hears that a few families are using LPG, now available at a nearby petrol station. The women say it is very quick and easy to use, and it keeps pots, clothes, and walls clean. The women and children seem to feel better, with less cough, runny eyes, and headaches. But those families run small shops and have been able to find the money to buy the gas bottle and cooker. She talks with her husband about LPG, and although quite supportive, her husband thinks they cannot afford it. They could spend a little more on fuel, but income is irregular. Why abandon free fuel when they are so poor? Ruth thinks she could earn more money from her handicrafts in the time she saves collecting wood. On balance, they reckon they could probably afford the cost of the gas if they could be sure of more regular income, but they do not know where they could find the money to pay for the cooker and bottle. Ruth then learns about a revolving fund set up by her women’s group with the help of an NGO. If she can make small regular payments, she and her husband could get a loan to buy the stove and gas bottle next year. But they have never saved before, and what if they need money for medicines or for the children at school? Will they be able to keep saving each week to make sure they have enough to refill the gas bottle when needed? 1. Not her real name. Source: Authors. networks, and support for credit (for example, seed funds and fund capital) and targeted subsidies. Some degree of market support may be required to stimulate demand and to encourage adoption by poor households, particularly those using three-stone fires (and other simple stoves) and collected biomass, because those methods do not incur direct monetary costs. Some countries have applied subsidies on fuels such as kerosene to assist poor families, but general subsidies are now considered to be an inefficient instrument for this purpose (von Schirnding and others 2002). Targeted subsidy and small-scale credit may be more appropriate ways of helping poor families acquire new household energy technologies and can have low default rates. Experience shows, however, that households are more likely to access credit for directly productive (with regard to income) uses of energy, rather than for everyday cooking and space-heating needs. Because the latter are the most important sources of IAP, more promotion of other benefits is needed, such as improved family health; fuel cost savings; time saved by faster cooking and reduced need for biomass; greater prestige; and cleaner homes, clothes, and utensils. A number of these benefits may result in reduced expenditure or increased income generation. Box 42.2 illustrates how these various issues can influence the decisions of a “typical” poor rural African household considering transition from gathered biomass to predominant use of a commercial fuel (LPG). Effectiveness Most evidence available for assessing intervention effectiveness deals with the effect on IAP levels and in some cases personal exposure. No experimentally derived evidence is available, however, on the effect of reducing IAP exposure on incidence of ALRI or the course of COPD in adults. A randomized trial of an improved chimney stove is currently under way in Guatemala, focusing on ALRI in children up to 18 months of age (Dooley 2003). A cohort study in Kenya by Ezzati and Kammen (2001) describes significant exposure-response relationships for all acute respiratory infections—and for ALRI specifically—associated with the use of traditional and improved woodstoves and charcoal. However, those effect estimates require confirmation because the study has small numbers of children (93 children under age five, living in 55 homes). For the other major health outcome, lung cancer, Lan and 802 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 38 others (2002) reported adjusted hazard ratios of 0.59 (95 percent confidence interval: 0.49 to 0.71) for men and 0.54 (0.44 to 0.65) for women using improved coal stoves compared with traditional open coal fires in a 16-year retrospective cohort study in rural China. Measuring evidence on reductions in pollution and exposure is nonetheless an important step in assessing effectiveness. Summarized here are the main findings of studies that have measured pollution levels in homes using traditional open fires, various improved stoves, kerosene, and LPG (see also Saksena, Thompson, and Smith 2004) and one that examined the effect of rural electrification in South Africa (Rollin and others 2004). Effect of Improved Stoves. In East Africa, cheap improved stoves without flues, burning either wood or charcoal, are popular. These wood-burning stoves can reduce kitchen pollution by up to 50 percent, but levels still remain high (Ezzati, Mbinda, and Kammen 2000). Charcoal emits much less PM (but with a higher CO-to-PM ratio than wood), and stoves such as the Kenyan jiko yield particulate levels in the region of 10 percent of those from wood fires. In a number of Asian and Latin American countries, improved stoves with flues have been promoted quite extensively, although many such stoves are found to be in poor condition after a few years. Some studies from India have shown minimal or small reductions in PM (Ramakrishna 1988; Smith, Aggarwal, and Dave 1983). Other studies, from Nepal, have shown reductions of about two-thirds, although the very high baseline levels mean that homes with stoves still recorded total suspended particulate values of 1,000 to 3,000 ␮g/m3 during cooking (Pandey and others 1990; Reid, Smith, and Sherchand 1986). Results from Latin American countries are similar, although the IAP levels are generally lower. Studies have shown that plancha-type stoves (made of cement blocks, with a metal plate and flue) reduce PM by 60 to 70 percent and by as much as 90 percent when they are in good condition. Typical 24-hour PM levels (PM10, PM3.5 [respirable], and PM2.5 have variously been reported) with open fires of 1,000 to 2,000 ␮g/m3 have been reduced to 300 to 500 ␮g/m3, and in some cases to less than 100 ␮g/m3 (Albalak and others 2001; Brauer and others 1996; Naeher, Leaderer, and Smith 2000). One study from Mexico found little difference between homes with open fires and with improved stoves (Riojas-Rodriguez and others 2001), but the 16-hour levels of PM10 at about 300 ␮g/m3 with open fires were relatively low. Improved stoves with flues have so far had little success in Sub-Saharan Africa, although recent work developing hoods with flues for highly polluted Kenyan Masai homes reported reductions in 24-hour mean respirable PM of 75 percent from more than 4,300 ␮g/m3 to about 1,000 ␮g/m3 (Bruce and others 2002). Personal exposures were usually found to have been reduced proportionately less than area pollution levels. For example, in Kenya, where hoods with flues achieved a 75 percent reduction in 24-hour mean kitchen PM3.5 and CO, the woman’s mean 24-hour CO exposure was reduced by only 35 percent (Bruce and others 2002). Similar results were found for child exposures in a study of improved wood stoves in Guatemala (Bruce and others 2004). We are aware of only one study that has used direct measurement of personal particulate exposure in very young children (Naeher, Leaderer, and Smith 2000). This study, also in Guatemala, reported mean 10- to 12-hour (daytime) PM2.5 levels for children under 15 months of age of 279 ␮g/m3 (⫹SD of 19.5) for the open fire and 170 ␮g/m3 (⫹154) for the plancha stoves, a 40 percent reduction. Impact of Cleaner Fuels. Good evidence shows that kerosene and LPG can deliver much lower levels of pollution, although it is important to determine the extent to which the cleaner fuel is substituting for biomass. For example, a study in rural Guatemala comparing LPG with open fires and plancha chimney stoves found that LPG-using households typically also used an open fire for space heating and cooking with large pots. As a result, the plancha stoves achieved the lowest pollution levels in that setting (Albalak and others 2001). Still, a number of studies, mainly from India, show that introducing kerosene and LPG dramatically reduces kitchen pollution, which perhaps reflects different cooking requirements and less need for space heating. In rural Tamil Nadu, two-hour (mealtime) kitchen respirable PM levels of 76 ␮g/m3 using kerosene and of 101 ␮g/m3 using gas contrasted with levels of 1,500 to 2,000 ␮g/m3 using wood and animal dung (Parikh and others 2001). Personal (cook) 24-hour exposure to respirable PM was 132 ␮g/m3 with the use of kerosene as opposed to 1,300 and 1,500 ␮g/m3, respectively, with the use of wood and dung (Balakrishnan and others 2002). Other studies confirm those findings, for example, with the use of gas in Mexico (Saatkamp,Masera,and Kammen 2000). Delivering electricity to rural homes requires extensive infrastructure, and most poor people with access to electricity can afford to use it only for lighting and running low-demand electrical appliances. Without marked improvements in socioeconomic conditions, electrification has little potential to bring about substantial reductions in IAP. South Africa is one of the few countries with a large rural population traditionally dependent on biomass that has the resources for rural electrification. An investigation of three rural villages with similar socioeconomic characteristics, two not electrified and one electrified, in the North West province found that 3.6 years (average) after connection to the grid, 44 percent of the electrified homes had never used an electric cooker (Rollin and others 2004). Only 27 percent of electrified homes cooked primarily with electricity; the remainder used a mix of electricity, kerosene, and solid fuels. Despite the mixed fuel use, households Indoor Air Pollution | 803 ©2006 The International Bank for Reconstruction and Development / The World Bank 39 cooking with electricity had the lowest pollution levels. Overall, homes in the electrified village had significantly lower 24-hour mean respirable PM and CO levels and significantly lower mean 24-hour CO exposure for children under 18 months of age than homes in the nonelectrified villages. Effect of Other Interventions. Little systematic evaluation has been made of other interventions listed in table 42.4. Investigation of the potential of improving ventilation has, overall, shown that although enlarging eaves can be quite effective (Bruce and others 2002), removing smoke generally requires a well-functioning flue or chimney. Behavioral changes are currently the subject of an intervention study in South Africa (Barnes and others 2004a, 2004b). Cost-Effectiveness Analysis Although clean fuels can be expected to have a greater health effect than improved stoves (even those with flues), clean fuels may be too expensive and inaccessible for many poor communities over the short to medium term. Furthermore, even though clean fuels may be the best longer-term goal, an intermediate stage of improved biomass stoves may promote change by raising awareness of benefits and thus creating demand by improving health, saving time, and mitigating poverty. For those reasons, this cost-effectiveness analysis (CEA) examines both improved biomass stove and clean fuel options in the following scenarios: • access to improved stoves (stoves with flues that vent smoke to the exterior), with coverage of 95 percent • access to cleaner fuels (LPG or kerosene), with coverage of 95 percent • part of the population with access to cleaner fuels (50 percent) and part with improved stoves (45 percent). In each case, the intervention is compared with the current level of coverage of the respective technology or fuel. Cost Assumptions. The assumptions for costs include program costs, fixed costs (including stoves), and recurrent fuel costs. Household costs for each region were drawn from the most comprehensive estimates available in the literature (von Schirnding and others 2002; Westoff and Germann 1995). For LPG, costs include the initial price of a cooker and cylinder and the recurrent refill costs. Assumed household annual costs, discounted at 3 percent, range from US$1 to US$10 for improved stoves and from US$3 to US$4 for kerosene or up to US$30 for LPG. Recurrent costs of fuel were found to be the most significant cost for the cleaner fuel interventions. Wood fuel costs are estimated at US$0.25 per week and assumed to be the same for traditional and improved stoves. Costs were estimated separately for cleaner fuel and improved stove programs, using an “ingredients” approach (Johns, Baltussen, and Hutubessy 2003) and a costing template developed by WHO (2003). In summary, all the ingredients— including administrative, training, and operational costs— necessary to set up and maintain a given program must be added up. For regional estimates, costs of all traded goods were in U.S. dollars, whereas nontraded (local) costs were estimated in local currency and converted to U.S. dollars using relevant exchange rates. All costs were annualized using a 3 percent discount rate. Costs for tradable goods are scaled, using region-specific standardized price multipliers to reflect the increasing costs of expanding coverage caused by higher transportation costs to more remote areas (Johns, Baltussen, and Hutubessy 2003). Price multipliers were not applied to improved stoves because they tend to be manufactured locally with mainly local materials. Program costs were found to make up a small proportion of the overall intervention costs. Savings from averted health care costs are not included; because many of these cases currently go untreated, it can be argued that including treatment costs could result in inflated cost-effectiveness ratios (CERs). Effectiveness and Health Outcome Assumptions. For this analysis, cleaner fuels are assumed to remove exposure completely, whereas improved stoves are assumed to reduce exposure by 75 percent (ventilation factor of 0.25). The effect on health of the exposure reduction will vary from region to region, because it depends on current levels of exposure as well as region-specific rates of morbidity and mortality. A number of assumptions have been made about households in carrying out analyses at the regional level. First, regional estimates of household composition (numbers of people, by age group and sex) and, hence, the effect of interventions on exposure and health apply at the level of individual households. Second, the age distribution of household members is similar in exposed and nonexposed groups; for example, the number of children per household is the same irrespective of household fuel use and ventilation characteristics. That assumption is likely to be conservative, since poorer, more polluted homes will typically have higher fertility and more children under five; all other factors being equal, such households would therefore experience a higher burden of disease from IAP exposure. The health outcomes included are ALRI and COPD, because they were responsible for nearly all of the 1.6 million deaths attributable to IAP. The risk estimates used are those derived from the meta-analyses, as summarized in table 42.2. Smoking is an important confounding variable for COPD, particularly with men, because they generally smoke more than women do in developing countries. At present, information is sparse on the independent effect of solid fuel use on COPD in the presence of smoking. To avoid possible overestimation of the impact of IAP on COPD, attributable fractions for COPD from solid fuel use 804 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 40 were applied to disease burdens remaining after removal of smoking-attributable burdens (Ezzati and Lopez 2004). Current estimates of exposure are used in combination with estimates of disease burden to obtain region-specific disease burdens for exposed and unexposed populations. Regional patterns of disease for 2000 have been used, including incidence, mortality, remission, duration, and case-fatality rate, obtained from WHO (2004b). In contrast to the estimates of burden in table 42.3, no age weighting has been used in the cost-effectiveness analysis. Health impacts are discounted at 3 percent. Implementation Period. The implementation period is 10 years, although effects have been evaluated over 100 years in order to approximate the benefits for an entire population cohort. Thus, health effects are calculated for a cohort with a typical age structure for the population concerned that experiences the intervention for 10 years. It is assumed that after 100 years, all of the cohort (including children born during the 10-year implementation period) will have died. The implementation period has critical implications, particularly in situations in which it takes several years to establish an intervention (for example, developing local markets for cleaner fuel), in which there are high start-up costs, and in which disease prevention is experienced in the distant future. This is especially true for chronic health effects (for example, COPD) that result from exposure over many years. If the intervention is implemented and exposure reduced for only 10 years, the disease burden is effectively deferred by 10 years, whereas longer-term implementation would result in many more cases being averted. For this analysis, using the 10-year intervention scenario specified for the Disease Control Priorities Project-2, incident cases are deferred by 10 years. For COPD, it is assumed that reduced exposure results in a milder form of COPD, accounted for by using a lower severity weighting. Findings for Cost-Effectiveness Analysis. Findings from the CEA are expressed, for the four intervention scenarios with differing coverage (50 percent, 80 percent, 95 percent), by region, as (a) total healthy years gained in each region, (b) CERs in U.S. dollars per healthy year gained, and (c) healthy years gained per US$1 million (table 42.6). For all regions, the cleaner fuels yield the greatest gain in healthy years, but improved stoves also have a significant effect. The largest total population gains in healthy years are in Sub-Saharan Africa and South Asia for all types of interventions and in East Asia and the Pacific (mainly China) for cleaner fuels. In the two regions with the largest burden of disease attributable to solid fuel use (Sub-Saharan Africa and South Asia), CERs are lowest (most favorable) for improved stoves, although in both regions kerosene has CERs just over twice those of improved stoves. In East Asia and the Pacific, kerosene is most cost-effective, followed by improved stove and clean fuel combinations and then by LPG (for coverage over 50 percent). In Latin America and the Caribbean, kerosene has the most favorable CER, followed by kerosene in combination with improved stoves. When the 50 percent and 80 percent coverage scenarios are compared, large differences in the ratio are seen in regions where coverage for that intervention is already substantial, and there is much less health gain at lower levels of coverage. Where no result is given, the specified coverage of the intervention has already been reached. Multivariate sensitivity analysis was conducted to assess the effect of uncertainty in cost and effectiveness estimates. Costs were assumed to vary with a standard deviation of 5 percent, and effectiveness by the range of the confidence interval around the relative risk for each health endpoint. Results for Southeast Asia are shown in figure 42.3: the “clouds,” or uncertainty regions, illustrate the range of possible point estimates emerging from the sensitivity analysis. This example is representative because other regions show essentially similar results. Despite the uncertainty, the ranking of the interventions remains the same (Mehta and Shahpar 2004). Discussion. Results of this cost-effectiveness analysis indicate that an improved biomass stove is the most cost-effective intervention for South Asia and Sub-Saharan Africa, the two regions with the highest solid fuel–related disease burden. This finding is important given International Energy Agency projections to 2030, which indicate that biomass will remain the principal household fuel for the poor in South Asia and Sub-Saharan Africa and that actual numbers of users will increase over that period (IEA 2002). Cleaner fuels (particularly kerosene) are the most cost-effective options for East Asia and the Pacific, the other region with a high burden of solid fuel–related disease. Cleaner fuels, in particular LPG, appear relatively costly for South Asia and Sub-Saharan Africa, but circumstances in individual countries may vary considerably and in ways that make this fuel much more cost-effective. Sudan, for example, has abundant cheap supplies of LPG and favorable excise arrangements for imported appliances, which would result in a lower CER for LPG than in other countries in the region. Furthermore, as will be discussed later, costs and benefits from the user’s perspective will differ markedly, depending on whether the starting point is free fuel collection or purchased biomass fuel. In interpreting the results, one should bear in mind the assumptions underlying the CEA. Much of the evidence indicates that, although improved biomass stoves may reduce kitchen pollution by up to 75 percent, the reduction in exposure of women and children is typically no more than 30 to 40 percent (equivalent to VF of 0.6 to 0.7). Achievement of the 75 percent reduction in exposure (VF ⫽ 0.25) assumed for this analysis is consistent only with well-designed and -maintained chimney stoves that meet most of the cooking and heating energy needs of the household and high population coverage (to avoid exposure from neighbors and others). Those conditions may be Indoor Air Pollution | 805 ©2006 The International Bank for Reconstruction and Development / The World Bank 41 Table 42.6 Intervention Scenarios for World Bank Regions a. Healthy years gained Intervention LPG Kerosene Improved stove Combined (with stove) Coverage (percent) SubSaharan Africa Latin America and the Caribbean Middle East and North Africa 50 22,160,000 160,000 n.a. n.a. 44,810,000 2,560,000 80 60,370,000 4,670,000 15,570,000 1,330,000 149,300,000 228,710,000 Europe and Central Asia South Asia East Asia and the Pacific 95 75,630,000 11,260,000 22,510,000 4,810,000 184,940,000 568,640,000 50 22,160,000 160,000 n.a. n.a. 44,810,000 2,560,000 80 60,370,000 4,670,000 15,570,000 1,330,000 149,300,000 228,710,000 95 75,630,000 11,260,000 22,520,000 4,810,000 184,940,000 568,640,000 50 18,010,000 n.a. n.a. n.a. 48,880,000 1,120,000 80 40,270,000 1,380,000 6,630,000 n.a. 101,670,000 6,980,000 95 51,540,000 2,600,000 11,640,000 n.a. 128,380,000 32,760,000 LPG 69,250,000 8,650,000 19,540,000 3,230,000 170,340,000 427,350,000 Kerosene 69,250,000 8,650,000 19,540,000 3,230,000 170,340,000 427,350,000 Middle East and North Africa Europe and Central Asia South Asia East Asia and the Pacific b. Cost-effectiveness ratios (US$ per healthy year gained) Intervention LPG Kerosene Improved stove Combined (with stove) Coverage (percent) SubSaharan Africa Latin America and the Caribbean 50 715 1,405 n.a. n.a. 542 1,695 80 518 783 756 1,221 312 115 95 518 814 762 1,321 314 100 50 84 631 n.a. n.a. 63 225 80 60 115 95 183 36 14 95 60 106 95 167 36 12 50 25 n.a. n.a. n.a. 15 297 80 21 947 457 n.a. 13 587 95 20 1,101 368 n.a. 13 327 LPG 295 761 606 1,375 177 83 Kerosene 45 296 220 507 26 25 Latin America and the Caribbean Middle East and North Africa c. Healthy years gained per US$1 million Intervention LPG Kerosene Improved stove Combined (with stove) Coverage (percent) SubSaharan Africa Europe and Central Asia South Asia East Asia and the Pacific 50 1,400 710 n.a. n.a. 1,840 590 80 1,930 1,280 1,320 820 3,210 8,680 10,040 95 1,930 1,230 1,310 760 3,190 50 11,970 1,580 n.a. n.a. 16,000 4,440 80 16,600 8,690 10,500 5,470 27,850 72,840 95 16,620 9,470 10,560 6,000 27,680 85,840 50 39,640 n.a. n.a. n.a. 67,330 3,360 80 47,940 1,060 2,190 n.a. 74,750 1,700 95 49,510 910 2,720 n.a. 76,300 3,060 LPG 3,390 1,310 1,650 5,660 5,660 12,020 Kerosene 22,250 3,380 4,550 38,590 35,590 40,730 Source: Authors. n.a. ⫽ not applicable because the specified coverage of the intervention has already been reached. 806 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 42 Average annual cost in millions of international dollars 25,000 20,000 15,000 10,000 5,000 0 0 50 100 150 200 250 Average annual gain in healthy years in millions 300 Improved stoves Kerosene Propane and LPG Combination: LPG and improved stoves Combination: kerosene and improved stoves Source: Mehta and Shahpar 2004. Figure 42.3 Multivariate Sensitivity Analysis for Three Types of Interventions and Combined Intervention Scenarios, Southeast Asia Region achievable and should be the goal, but they are not currently widespread. The relative cost-effectiveness advantage for improved stoves over cleaner fuel reported here should therefore be viewed as relating more to what might be achievable with good biomass stoves rather than to what is currently being achieved. The assumption that kerosene and LPG are equally clean and achieve zero exposure (VF = 0) presumes, at the very least, the use of high-quality kerosene fuel and pressurized burners. In many places, kerosene is of low quality, and the types of kerosene stoves and lamps used result in poor combustion. Cost comparisons for the various fuels also need careful consideration. For example, the cost of solid fuel has been assumed to be constant for traditional open fires and improved stoves. As a general assumption this is reasonable, because the efficiency of “improved” stoves varies, and some may even be less fuel efficient than are open fires. However, new stove technology is markedly improving efficiency, and some designs reduce daily fuel consumption by 40 percent or more, resulting in savings of time (where fuel is collected) and money (where fuel is bought) (Boy and others 2000). Transition from biomass (collected free) or charcoal (typically paid for daily in small amounts) to LPG would almost certainly require changes in saving and budgeting habits for a poor household (see also box 42.2). Those changes may entail arranging a loan to purchase the gas bottle and stove and saving money for the relatively large, periodic outlay to refill the cylinder. Such changes are very likely to have other consequences for the family that should not be overlooked. However, those consequences are complex and difficult to allow for within the current CEA framework. Empirical data are required on how household budgets change with various interventions and approaches to implementation. The calculations have been undertaken for whole regions and provide no indication of how CERs differ among countries and specific communities. As local data on exposure, risk factors, health outcomes, and intervention effectiveness become available, similar analyses should be conducted at national and subnational levels. Averted treatment costs have not been included on the grounds that most users of solid fuel are poor and have limited access to health services; many do not seek medical care for ALRI and even fewer do so for COPD. Inclusion of averted costs would increase cost-effectiveness. However, efforts to raise awareness about health risks and the importance of seeking care for ALRI (and COPD), which should accompany an intervention program, may increase care seeking and costs to the consumer. As more complete information becomes available, future CEAs should include treatment costs, with the option of allowing for an increase in care seeking associated with the intervention. Interpretation of the results of this CEA, particularly with respect to comparisons with other types of intervention, needs to acknowledge that, although public organizations and other agencies will (or may) have some involvement in funding intervention programs, most of the cost of market-based interventions will be borne by households and those involved in production and marketing. Furthermore, it is hoped that, in addition to reducing IAP, interventions (and the means of accessing them) will have other positive effects, including on household budgets, in creating opportunities for income generation and empowering women in decisions about how energy is used. The promotion of market-based solutions implies new opportunities for artisans and entrepreneurs, but also the loss of traditional employment. The balance sheet for interventions is therefore complex, is specific to the setting, and will evolve as markets and enterprise develop. Cost-Benefit Analysis The CER gives cost per unit of health gained (healthy year) based on reduced risk of specified disease outcomes (ALRI, COPD). As discussed earlier, however, household energy interventions can affect a wide range of social, economic, and environmental issues, with important implications for health and development. In an economic analysis of water and sanitation interventions, Hutton and Haller (2004) found that time saved was the most important benefit. Those other effects cannot easily be expressed in units of health gain. Cost-benefit analysis (CBA) offers an alternative approach that may be better suited to environmental health interventions, given that health arguments alone will not motivate the multiple sectors involved in financing and implementing household energy interventions. All main benefits in CBA are expressed in a common unit of monetary value and compared with costs in the cost-benefit Indoor Air Pollution | 807 ©2006 The International Bank for Reconstruction and Development / The World Bank 43 ratio (CBR). The assessment of costs in CBA would have many assumptions and methods in common with CEA. The key differences lie in the selection of effects for inclusion as benefits and the methods for valuing them. In principle, there is no reason all the full range of effects discussed earlier could not be included (table 42.7), although in practice some, such as global climate effects, may be too uncertain. Where disadvantages of interventions are identified, they should also be included. Benefit valuation presents particular challenges: effects are highly setting specific; evidence for some is limited, and their effects poorly quantified; and valuation in monetary units of benefits, such as lost working time averted for women, is difficult because women frequently are unpaid or work in informal markets. As a result, methods of valuation based on human capital may not be suitable, and alternative approaches such as contingent valuation, in which communities are involved in agreeing on market values for nontradable commodities, may be preferable. A related issue is valuing benefits that relate to sustainability and health, which would be experienced after many years and by subsequent generations. Larson and Rosen (2002) used a mix of valuation of statistical life and contingent valuation methods to examine the CBRs for improved stoves with respect to mortality (Guatemala, East Africa) and morbidity (Pakistan), concluding that ratios appeared favorable. Although they discuss other benefits, those benefits were not included in their valuations. Their observation that the favorable CBRs are not reflected in the generally low adoption of improved stoves led them to conclude that the information required for assessing household demand correctly is not currently available. IMPLEMENTATION OF CONTROL STRATEGIES: LESSONS FROM EXPERIENCE The past 30 to 40 years have seen many diverse programs on household energy, from small-scale NGO- and community-led Table 42.7 Possible Data Requirements for Quantifying Benefits Impact category Variables or elements to identify Direct benefits related to specific health outcomes Expenditure and time for health care–seeking Health service use of those with diseases caused by IAP (number of cases, visits or days per case) Health service use of those having accidents or injuries due to reasons related to fuel use: Direct: burns, poisoning Indirect: injuries in collecting fuel Access features to get to health services (distance, mode of transport, time; average visits per case) Other consumption related to health care–seeking Time loss of seeking health care, both of the patient and of those accompanying patient Other direct benefits in and around the home Time gained owing to less illness and death Activities of those with diseases caused by IAP Impact of disease on activities (time input, productivity) Value of time of various occupations Time saving of changed technology Reduced time spent collecting fuel Reduced time spent cooking and on other tasks requiring fire or stove Value of time of various occupations Income-generating activities achieved through increased time Impact on household cleanliness and hygiene and need for cleaning Change in household environment and production Effect of improved lighting on evening activities (education, production) Effect of availability of electricity and other fuels on household production activities Impact on ergonomics related to cooking Consequences of process of acquiring new technology and related changes Increased confidence in capacity of the household to save for immediate or future needs More involvement of women in decision making with respect to changes in household energy use and related issues Indirect benefits related to the environment Local environment Impact of fuel scarcity on local environment, average fuel collection time Global environment Contribution of local area to greenhouse gases Increased risk of environmental effects (such as soil fertility) or disasters (such as flooding, landslides) Source: Authors. 808 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 44 initiatives to ambitious national programs, the largest of which has been the installation of some 200 million improved stoves in rural China. Although few have been subjected to rigorous evaluation, an assessment has been made of the Indian national stove program (box 42.3; ESMAP and World Bank 2001); the Chinese national stove program (box 42.4; Sinton and others 2004; Smith and others 1993); and LPG promotion (box 42.5; UNDP and ESMAP 2002). Experience with a number of smaller initiatives has also been reported—for example, the ceramic and metal stoves in East Africa, which have proved popular and provided local employment (Njenga 2001), and improved stove interventions in Guatemala (UNDP and ESMAP 2003). Implementation of the Chinese national program differed substantially from that in India. Although the Chinese rural populations concerned are poor, they do have greater effective purchasing power than the poor in many developing countries, allowing development of a program with the majority of con- sumers purchasing stoves at close to full cost (Smith and others 1993). Among the key features of the Chinese program reported to have contributed to its success are decentralization of administration; a commercialization strategy that provided subsidies to rural energy enterprise development and quality control through the central production of critical components, such as parts of the combustion chamber; and engagement of local technical institutions in modifying national stove designs to local needs. National-level stove competitions were held among counties for contracts, ensuring local interest and allowing the best-placed counties to proceed first; financial payments were provided to counties only after completion of an independent review of their achievements. No large flows of funds came from the central government (in contrast, for example, to India); local governments provided the major financial contributions. As a result, delays and other problems associated with transferring large amounts of money have been Box 42.3 Key Features and Lessons from India’s National Stove Program The Indian National Programme of Improved Cookstoves was established in 1983 with goals common to many such initiatives: • conserving fuel • reducing smoke emissions in the cooking area and improving health conditions • reducing deforestation • limiting the drudgery of women and children and reducing cooking time • improving employment opportunities for the rural poor. Although the Ministry of Non-Conventional Energy Sources was responsible for planning, setting targets, and approving stove designs, state-level agencies relayed this information to local government agencies or NGOs. A technical backup unit in each state trained rural women or unemployed youths to become self-employed workers to construct and install the stoves. Between 1983 and 2000, the program distributed more than 33 million improved stoves. Despite extensive government promotion efforts, improved stoves now account for less than 7 percent of all stoves. Among those that have been adopted, poor quality and lack of maintenance have resulted in a life span of two years at most and typically much less. Evaluation of the program identified four main problems: • Most states placed inadequate emphasis on commercialization, now seen as crucial for effective and sustainable uptake. • Overall, there was insufficient interaction with users, self-employed workers, and NGOs, so the designs did not meet needs of households, and there was very poor acceptance of user training. • Quality control for installation and maintenance of the stove and its appropriate use was lacking. • High levels of subsidy (about 50 percent of the stove cost) were found to reduce household motivation to use and maintain the stove. Some more successfully managed areas of the program focused resources on technical assistance, research and development, marketing, and information dissemination. Recently, the government of India decentralized the program and transferred all implementation responsibility to state level. Since 2000, the program promotes only durable cement stoves with chimneys that have a minimum life span of five years. The introduction of these stoves will make adhesion to technical specifications and quality control much easier. Source: Authors, based on ESMAP and World Bank 2001. Indoor Air Pollution | 809 ©2006 The International Bank for Reconstruction and Development / The World Bank 45 Box 42.4 Household Effects of China’s National Improved Stove Program In 2002, an independent multidisciplinary evaluation was undertaken by a team of U.S. and Chinese researchers to evaluate (a) implementation methods used to promote improved stoves; (b) commercial stove production and marketing organizations that were created; and (c) effects of the program on households, including health, stove performance, socioeconomic factors, and monitoring of indoor air quality. The first two objectives were assessed through a facility survey of 108 institutions at all levels. The third objective was assessed through a household survey of nearly 4,000 households in three provinces: Zheijang, Hubei, and Shaanxi. Key findings were as follows: not among households using combinations of fuels that included coal or LPG), improved stoves showed significantly lower PM4 and CO concentrations than traditional stoves. • In both children and adults, coal use was associated with higher levels of exposure (as measured by CO in exhaled breath) and improved biomass stoves with lower levels. Reported childhood asthma and adult respiratory disease were negatively associated with use of improved stoves and good stove maintenance. These results should, however, be treated as indicative because of limited sample size. • The household survey revealed highly diverse fuel usage patterns: 28 and 34 different fuel combinations were used in kitchens in winter and summer, respectively. Most households owned at least one or more coal and one or more biomass stoves. Of the biomass stoves 77 percent, but only 38 percent of the coal stoves, were classified as improved. On average, improved stoves had a mean efficiency of 14 percent, which is well below the program target of between 20 and 30 percent, but above the mean efficiency of 9 percent for traditional stoves. • With respect to air quality (measured with PM4, the “thoracic fraction” of particulate matter, and CO), coal stoves showed significantly higher concentrations than biomass stoves during the summer, but not during the winter. Among households using biomass fuels (but Overall, several important conclusions emerge with relevance to future improved stove programs: • A wide range of combinations of different fuel and stove types may limit the effect of an improved stove program. • Given the importance of space heating, making available an improved biomass stove for cooking may not be a sufficient strategy to reduce IAP. Improved coal stoves need to be promoted among rural Chinese households. • Even among households using improved stoves, PM4 and CO levels were higher than Chinese national indoor air standards, implying that a large fraction of China’s rural population is still chronically exposed to pollution levels substantially above those determined by the Chinese government to harm human health. Source: Authors, based on Sinton and others 2004. avoided. The Chinese program succeeded in shifting norms: most biomass stoves now available on the market have flues and other technical features that classify them as improved. Experience in the promotion of LPG has also been reported, for example, from the Indian Deepam Scheme (ESMAP and World Bank 2004; UNDP and ESMAP 2002) and from the LPG Rural Energy Challenge (UNDP 2005). The latter initiative, developed by UNDP and the World LPG Association in 2002, is promoting the development of new, viable markets for LPG in developing countries. Key elements include developing partnerships in countries; enabling regulatory environments that facilitate LPG business development and product delivery; reducing barriers, for example, by introducing smaller (more affordable) gas bottles; and raising government and consumer awareness of costs and benefits. McDade (2004) has recently identified a number of key lessons emerging from experience with the promotion of LPG markets (box 42.5). Electrification has an important role in development (IEA 2002). Evidence from South Africa suggests that communities with grid access experience lower IAP exposure (Rollin and others 2004). Electricity is not expected to bring about large reductions in IAP exposure in most low-income countries, however, because most poor households can afford it only for uses such as lighting and running entertainment appliances and not for cooking and space heating. The International Energy Agency has recently carried out a detailed review of electrification, including the issues involved in supply and cost recovery among poor (and especially rural) communities (IEA 2002). 810 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 46 Box 42.5 Key Lessons Learned in the Promotion of New Markets for LPG in Developing Countries • LPG can be affordable outside of urban areas, where wood fuel is currently purchased. On the other hand, for many consumers who do not participate in the monetized economy, it will be premature to promote LPG markets. • One-time subsidies on appliances could be a good use of government (or other) resources. • Microcredit initiatives should emphasize the costsaving and productive potential and should seek to package both the gas (and appliances) and the financing. • Concerns about safe handling, cylinder refilling, and transportation can be serious barriers to market expansion. These issues need to be addressed by raising awareness among consumers and strengthening regulatory environments. • Appliances for a range of end uses required by consumers must be available. • Government leadership is essential, backed up by policy that sets the basic parameters for successful market expansion and avoids conflict between, for example, subsidies on competing fuels that undermine efforts to promote LPG markets. • Specific initiatives, such as integrated energy centers (as in Morocco and South Africa) offer an effective means of developing markets in rural areas. Source: Authors, based on McDade 2004. The key lessons from experience with interventions to date may be summarized as follows: • Too often, intervention technologies have been developed without adequate reference to users’ needs and, as a result, have been poorly used and maintained or abandoned. Consequently, it is important to involve users—particularly women—in assessing needs and developing suitable interventions. • Sustainable adoption should also be promoted through greater availability of a choice of appropriately priced interventions through local commercial outlets (artisans, shops, markets). This situation will come about only if demand is sufficient and if producers and distributors recognize this demand. • All too commonly, communities most at risk exhibit low awareness, low demand, and poverty (often extreme poverty). A combination of user involvement and market approaches is needed, supported by the promotion and availability of targeted subsidies or microcredit facilities or both. The nature and extent of such financial support should depend on the purchasing power of the community. • Local initiatives such as those outlined above must be led by national (and subnational) policy that acknowledges the contributions of a range of actors (government, business, NGOs, and so on) and sectors (energy, health, environment, finance, and so on) and that results in coordinated action. The instruments listed in table 42.5 should be considered when developing national policy. In a recent review of the situation in Guatemala, the United Nations Development Programme and Energy Sector Management Assistance Programme (UNDP and ESMAP 2003) found that, despite the almost total reliance of the rural population on biomass, a marked lack of national policy, leadership, and coordinated action existed in relation to household energy. Countries need to develop mechanisms for action and coordination in light of local needs, available institutional capacity, and leadership potential. THE RESEARCH AND DEVELOPMENT AGENDA WHO has, through a process involving multistakeholder meetings and reviews, developed some consensus on research and development priorities for household energy, IAP, and health (see for example WHO 2002a). Effective coordination is a prerequisite because of the need for input from, and collaboration between, many different organizations and “actors” that have generally not previously worked in partnership on this issue. One recent response to this need has been the establishment of the Partnership for Clean Indoor Air, following the Johannesburg World Summit on Sustainable Development in 2002 (EPA 2004; http://www.pciaonline.org/). The evidence base on health effects requires further strengthening, particularly to quantify the effect of a measured reduction in IAP exposure on the risk of key outcomes (for example, ALRI). A randomized controlled trial is currently under way in Guatemala, focusing primarily on ALRI in children up to 18 months of age (Dooley 2003); however, at least one other such trial on another continent would be desirable. Also required are observational studies for outcomes for which few studies currently exist, including tuberculosis, low Indoor Air Pollution | 811 ©2006 The International Bank for Reconstruction and Development / The World Bank 47 birthweight and perinatal mortality, cataracts, asthma, and cardiovascular disease. A small number of such studies are in progress, but further effort is required, with perinatal outcomes being a particular priority. Despite limitations in the evidence on health effects, what is known about the health, social, and economic consequences of current patterns of household energy use in poor countries is of sufficient concern to press ahead with an active program of research and development regarding interventions. This activity should address both the technology (and associated knowledge and behavior) and the approaches taken for implementation. Although some development and innovation in technology and fuels (for example, clean fuels derived from biomass) are likely to be valuable, the single greatest challenge is to promote wider access to—and adoption of—existing knowledge and interventions. Projects and programs currently in progress or being developed should be carefully evaluated using quantitative and qualitative methods to assess a range of effects. Work is currently under way to develop suitable methods and tools for this purpose (WHO 2005). Experience and lessons learned need to be disseminated widely to ensure that they reach governments, donors, researchers, NGOs, and communities. As part of this effort, WHO is developing a resource for countries that offers information on the effectiveness of interventions as well as the enabling factors that facilitate longterm, sustained adoption and use of suitable improved technologies in different settings (WHO 2004c). Economic assessment, including cost-effectiveness analysis, has a valuable part to play. Critical issues resulting from limited evidence have been identified about estimations and assumptions for costs, exposure reductions, health effects, and averted treatment costs, as well as the current inability to assess national and subnational cost-effectiveness. CBA may be more suitable for interventions in this and similar areas but will require better description of environmental, social, and economic effects and further development of valuation methods. New health studies and broadly based evaluations of interventions should help fill some of these gaps. Determination of the macroeconomic costs to countries of current household energy use and the potential gains resulting from change to more efficient and cleaner options could substantially add to the case for action. Monitoring progress requires the development and testing of standard indicators for use in such policy documents as the World Development Report and for routine application at national and subnational levels. The Millennium Development Goal Indicator on the proportion of the population using solid fuels is a key starting point, and WHO, the reporting agency, is working to broaden the monitoring of this indicator through international surveys, such as demographic and health surveys (ORC Macro 2004), the Multiple Indicator Cluster Survey (UNICEF 2004), and the World Health Survey (WHO 2004d), as well as through work on regional and national indicators conducted under the Global Initiative on Children’s Environmental Health Indicators (WHO 2004e). Future reporting will need to be further refined by taking into account differences in cooking practices (for example, type of stove and cooking location), as well as in fuel use for lighting and heating. Advocacy for stronger action, internationally and in countries, is required. Products and guidance for a range of audiences should be prepared, with clear messages on the extent of the problem, the population groups most affected, what works, and what should be avoided. Tools such as the recently published guidelines on estimating the national burden of disease from solid fuels will help provide local evidence to argue for greater attention and action (Desai, Mehta, and Smith 2004). CONCLUSIONS IAP from solid fuel use is responsible for a large burden of disease among the world’s poorest and most vulnerable populations. Inefficient and polluting household energy systems hold back development through resulting ill health, constraints on women’s time and income generation, environmental impacts, and other factors. Although there is a trend toward cleaner and more efficient energy with increasing prosperity, little improvement is in prospect for more than 2 billion of the world’s poorest people, particularly in South Asia and Sub-Saharan Africa. The number of people relying on traditional biomass is actually expected to increase until 2030. Although the development of new energy technologies has a part to play in addressing this problem, many effective interventions are already available. The single greatest challenge is to dramatically increase the access of poor households to cleaner and more efficient household energy systems. Much valuable experience has been gained from successful—and unsuccessful—programs in household energy over the past three to four decades. Despite this experience, coherent, evidence-based policy is lacking in most of the countries concerned, where the lessons from experience now need to be implemented. Implementation will require greater awareness of the problem at international and national levels, provision of support for national collaborative action, and a focus on supporting appropriate, mainly market-based interventions. Better information is crucial to this effort, including stronger evidence of the health effects of IAP exposure; assessment of the social, economic, and environmental benefits of interventions; and indicators to monitor progress. Economic analysis can help bring the case for action into policy, but it needs to be applied at country level and to include a wider range of benefits. Results from analysis at the regional level show that interventions can be cost-effective, particularly improved stoves, as long as these interventions can deliver substantial exposure reductions in practice. This conclusion, as 812 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 48 well as its qualification, is important given the expectation that biomass will remain the principal household fuel in many developing countries for more than 20 years. The balance of effort and resources put into promoting cleaner biomass interventions rather than cleaner fuels, or vice-versa, will be an important policy issue for many countries and for the international community (Smith 2002). With a range of innovative projects and programs under way in a number of countries and regions of the world, now is an important time to focus attention and effort on achieving the health, social, and economic gains that should result from improvements in household energy systems in developing countries. NOTE 1. Particles are typically described according to the aerodynamic diameter, and although the devices used to separate particles of a given size do not yield a very sharp cutoff, this classification is functionally useful because smaller particles are able to penetrate farther into the lungs. Total suspended particles (TSP) include suspended particles of all sizes. Commonly defined smaller particles include PM10 (up to 10 microns diameter); respirable PM (includes all very small particles, about 50 percent of those 4 microns in diameter, and none above 10 microns in diameter); and PM2.5 (up to 2.5 microns in diameter). Boy, E., N. G. Bruce, and H. Delgado. 2002. “Birthweight and Exposure to Kitchen Wood Smoke during Pregnancy.” Environmental Health Perspectives 110 (1): 109–14. Boy, E., N. G. Bruce, K. R. Smith, and R. Hernandez. 2000. “Fuel Efficiency of an Improved Wood Burning Stove in Rural Guatemala: Implications for Health, Environment, and Development.” Energy for Sustainable Development 4 (2): 21–29. Brauer, B., K. Bartlett, J. Regaldo-Pineda, R. Perez-Padilla. 1996. “Assessment of Particulate Concentrations from Domestic Biomass Combustion in Rural Mexico.” Environmental Science and Technology 30: 104–9. Bruce, N. G., E. Bates, R. Nguti, S. Gitonga, J. Kithinji, and A. Doig. 2002. “Reducing Indoor Air Pollution through Participatory Development in Rural Kenya.” In Proceedings of 9th International Conference on Indoor Air Quality and Climate, Monterey, CA, 590–95. Bruce, N. G., J. P. McCracken, R. Albalak, M. Schei, K. R. Smith, V. Lopez, and C. West. 2004. “The Impact of Improved Stoves, House Construction, and Child Location on Levels of Indoor Air Pollution and Exposure in Young Guatemalan Children.” Journal of Exposure Analysis and Environmental Epidemiology 14 (Suppl. 1): S110–17. Bruce, N. G., R. Perez-Padilla, and R. Albalak. 2000. “Indoor Air Pollution in Developing Countries: A Major Environmental and Public Health Challenge.” Bulletin of the World Health Organization 78 (9): 1078–92. Clancy, J. S., M. Skutsch, and S. Batchelor. 2003. The Gender-EnergyPoverty Nexus: Finding the Energy to Address Gender Concerns in Development. Project report CNTR998521. London: Department for International Development. Courtright, P., D. Haile, and E. Kohls. 1993. “The Epidemiology of Burns in Rural Ethiopia.” Journal of Epidemiology and Community Health 47 (1): 19–22. Danis-Lozano, R., M. H. Rodriguez, L. Gonzalez-Ceron, and M. Hernandez-Avila. 1999. “Risk Factors for Plasmodium Vivax Infection in the Lacandon Forest, Southern Mexico.” Epidemiology of Infection 122 (3): 461–69. REFERENCES Albalak, R., N. G. Bruce, J. P. McCracken, and K. R. Smith. 2001. “Indoor Respirable Particulate Matter Concentrations from an Open Fire, Improved Cookstove, and LPG/Open Fire Combination in a Rural Guatemalan Community.” Environmental Science and Technology 35 (13): 2650–55. Ando, M., M. Tadano, S. Asanuma, K. Tamura, S. Matsushima, T. Watanabe, and others. 1998. “Health Effects of Indoor Fluoride Pollution from Coal Burning in China.” Environmental Health Perspectives 106 (5): 239–44. Desai, M. A., S. Mehta, and K. R. Smith. 2004. Indoor Smoke from Solid Fuels: Assessing the Environmental Burden of Disease at National and Local Levels. Environmental Burden of Disease Series 4. Geneva: World Health Organization. Dooley, E. E. 2003. “New Stoves for Better Children’s Health?” Environmental Health Perspectives 111 (1): A33. D’Sa, A., and K. V. Narasimha Murthy. 2004. “LPG as a Cooking Fuel Option for India.” Energy for Sustainable Development 8 (3): 91–106. Balakrishnan, K., J. Parikh, S. Sankar, R. Padmavathi, K. Srividya, V. Venugopal, and others. 2002. “Daily Average Exposures to Respirable Particulate Matter from Combustion of Biomass Fuels in Rural Households of Southern India.” Environmental Health Perspectives 110 (11): 1069–75. EPA (U.S. Environmental Protection Agency). 2004. Partnership for Clean Indoor Air. Washington, DC: EPA. http://www.epa.gov/iaq/pcia.html. Ballard-Tremeer, G., and A. Mathee. 2000. “Review of Interventions to Reduce the Exposure of Women and Young Children to Indoor Air Pollution in Developing Countries.” Paper prepared for USAID/WHO International Consultation on Household Energy, Indoor Air Pollution and Health, Washington, DC, May 4–6. ———. 2004. Clean Household Energy for India: Reducing the Risks to Health. Delhi: World Bank. Barnes, B., A. Mathee, L. Shafritz, L. Krieger, L. Sherburne, and M. Favin. 2004a. “Testing Selected Behaviours to Reduce Indoor Air Pollution Exposure in Young Children.”Health Education Research 19 (5): 543–50. Barnes, B., A. Mathee, L. Shafritz, L. Krieger, and S. A. Zimicki. 2004b. “A Behavioural Intervention to Reduce Child Exposure to Indoor Air Pollution: Identifying Possible Target Behaviours.” Health Education and Behaviour 13 (3): 306–17. Behera, D., S. Dash, and S. Malik. 1988. “Blood Carboxyhaemoglobin Levels Following Acute Exposure to Smoke of Biomass Fuel.” Indian Journal of Medical Research 88 (December): 522–24. ESMAP (Energy Sector Management Assistance Programme) and World Bank. 2001. Indoor Air Pollution: Energy and Health for the Poor, Issue 5 (September). Delhi: World Bank. Ezzati, M., and D. M. Kammen. 2001. “Quantifying the Effects of Exposure to Indoor Air Pollution from Biomass Combustion on Acute Respiratory Infections in Developing Countries.” Environmental Health Perspectives 109 (5): 481–88. Ezzati, M., and A. Lopez. 2004. “Mortality and Morbidity Due to Smoking and Oral Tobacco Use: Global and Regional Estimates for 2000.” In Comparative Quantification of Health Risks: The Global Burden of Disease Due to Selected Risk Factors, ed. M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Murray. Geneva: World Health Organization. Ezzati, M., M. B. Mbinda, and D. M. Kammen. 2000. “Comparison of Emissions and Residential Exposure from Traditional and Improved Cookstoves in Kenya.” Environmental Science and Technology 34 (4): 578–83. Indoor Air Pollution | 813 ©2006 The International Bank for Reconstruction and Development / The World Bank 49 Finkelman, R. B., H. E. Belkin, and B. Zheng. 1999. “Health Impacts of Domestic Coal Use in China.” Proceedings of the National Academy of Science 96 (7): 3427–31. Gupta, S.,Y. C. Govil, P. K. Misra, R. Nath, and K. L. Srivastava. 1998.“Trends in Poisoning in Children: Experience at a Large Referral Teaching Hospital.” National Medical Journal of India 11 (4): 166–68. Pope, C. A. III, M. L. Hansen, R. W. Long, K. R. Nielsen, N. L. Eatough, W. E. Wilson, and D. J. Eatough. 2004. “Ambient Particulate Air Pollution, Heart Rate Variability, and Blood Markers of Inflammation in a Panel of Elderly Subjects.” Environmental Health Perspectives 112 (3): 339–45. Ramakrishna, J. 1988. “Patterns of Domestic Air Pollution in Rural India.” Ph.D. dissertation. University of Hawaii, Honolulu. Hutton, G., and L. Haller. 2004. Evaluation of the Costs and Benefits of Water and Sanitation Improvements at the Global Level. Geneva: World Health Organization. Reed, R. P., and F. M. Conradie. 1997. “The Epidemiology and Clinical Features of Paraffin (Kerosene) Poisoning in Rural African Children.” Annals of Tropical Paediatrics 17 (1): 49–55. IEA (International Energy Agency). 2002. “Energy and Poverty.” In World Energy Outlook 2002. Paris: International Energy Agency. Reid, H., K. R. Smith, and B. Sherchand. 1986. “Indoor Smoke Exposures from Traditional and Improved Cookstoves: Comparisons among Rural Nepali Women.” Mountain Research and Development 6 (4): 293–304. Johns, B., R. Baltussen, and R. Hutubessy. 2003. “Programme Costs in the Economic Evaluation of Health Interventions. Cost Effectiveness and Resource Allocation 1 (1). http://www.resource-allocation.com/ content/1/1/1. Lan, Q., R. S. Chapman, D. M. Schreinemachers, L. Tian, and X. He. 2002. “Household Stove Improvement and Risk of Lung Cancer in Xuanwei, China.” Journal of the National Cancer Institute 94 (11): 826–35. Larson, B. A., and S. Rosen. 2002. “Understanding Household Demand for Indoor Air Pollution Control in Developing Countries.” Social Science and Medicine 55 (4): 571–84. Riojas-Rodriguez, H., P. Romano-Riquer, C. Santos-Burgoa, and K. R. Smith. 2001. “Household Firewood Use and the Health of Children and Women of Indian Communities in Chiapas, Mexico.” International Journal of Occupational and Environmental Health 7 (1): 44–53. Rollin, H., A. Mathee, N. G. Bruce, J. Levin, and Y. E. R. von Schirnding. 2004. “Comparison of Indoor Air Quality in Electrified and UnElectrified Dwellings in Rural South African Villages.” Indoor Air 14 (3): 208–16. McDade, S. 2004. “Fueling Development: The Role of LPG in Poverty Reduction and Growth.” Energy for Sustainable Development 8 (3): 74–81. Saatkamp, B. D., O. R. Masera, and D. M. Kammen. 2000. “Energy and Health Transitions in Development: Fuel Use, Stove Technology, and Morbidity in Jarácuaro, Mexico.” Energy for Sustainable Development 4 (2): 5–14. Mehta, S., and C. Shahpar. 2004. “The Health Benefits of Interventions to Reduce Indoor Air Pollution from Solid Fuel Use: A Cost-Effectiveness Analysis.” Energy for Sustainable Development 8 (3): 53–59. Saksena, S., L. Thompson, and K. R. Smith. 2004. “Indoor Air Pollution and Exposure Database: Household Measurements in Developing Countries.” http://ehs.sph.berkeley.edu/hem/page.asp?id=33. Naeher, L., B. Leaderer, and K. R. Smith. 2000. “Particulate Matter and Carbon Monoxide in Highland Guatemala: Indoor and Outdoor Levels from Traditional and Improved Wood Stoves and Gas Stoves.” Indoor Air 10 (3): 200–205. Sinton, J. E., K. R. Smith, J. Peabody, L. Yaping, Z. Ziliang, R. Edwards, and G. Quan. 2004. “An Assessment of Programs to Promote Improved Household Stoves in China.” Energy for Sustainable Development 8 (3): 33–52. Naeher, L., K. R. Smith, B. Leaderer, L. Neufeld, and D. Mage. 2001. “Carbon Monoxide as a Tracer for Assessing Exposures to Particulate Matter in Wood and Gas Cookstove Households in Highland Guatemala.” Environmental Science and Technology 35 (3): 575–81. Smith, K. R. 1987. Biofuels, Air Pollution, and Health: A Global Review. New York: Plenum Press. Njenga, B. K. 2001. “Upesi Rural Stoves Project.” In Generating Opportunities: Case Studies on Energy and Women, ed. G. V. Karlsson and S. Misana, 45–51. Washington, DC: United Nations Development Programme. ———. 2002. “In Praise of Petroleum?” Science 298: 1847. OECD and IEA (Organisation for Economic Co-operation and Development and International Energy Agency). 2004. “Energy and Development.” In World Energy Outlook 2004. Paris: OECD and IEA. Onuba, O., and E. Udoidiok. 1987. “The Problems of Burns and Prevention of Burns in Developing Countries.” Burns 13 (5): 382–85. ORC Macro. 2004. “Demographic and Health Survey.” ORC Macro. http://www.measuredhs.com/. Palsson, K., and T. G. Jaenson. 1999. “Plant Products Used as Mosquito Repellants in Guinea Bissau, West Africa.” Acta Tropica 72 (1): 39–52. Pandey, M., R. Neupane, A. Gautam, and I. Shrestha. 1990. “The Effectiveness of Smokeless Stoves in Reducing Indoor Air Pollution in a Rural Hill Region of Nepal.” Mountain Research and Development 10 (4): 313–20. Parikh, J., K. Balakrishnan, V. Laxmi, and B. Haimanti. 2001. “Exposures from Cooking with Biofuels: Pollution Monitoring and Analysis for Rural Tamil Nadu, India.” Energy 26: 949–62. Paru, R., J. Hii, D. Lewis, and M. P. Alpers. 1995. “Relative Repellancy of Woodsmoke and Topical Applications of Plant Products against Mosquitoes.” Papua New Guinea Medical Journal 38 (3): 215–21. Pope, C. A. III, R. T. Burnett, M. J. Thun, E. E. Calle, D. Krewski, K. Ito, and G. D. Thurston. 2002. “Lung Cancer, Cardiopulmonary Mortality, and Long-Term Exposure to Fine Particulate Air Pollution.” Journal of the American Medical Association 287 (9): 1132–41. ———. 1989. “Dialectics of Improved Stoves.” Economic and Political Weekly 24 (10): 517–22. Smith, K. R., A. L. Aggarwal, and R. M. Dave. 1983. “Air Pollution and Rural Biomass Fuels in Developing Countries: A Pilot Village Study in India and Implications for Research and Policy.” Atmospheric Environment 17 (11): 2343–62. Smith, K. R., S. Mehta, and M. Feuz. 2004. “Chapter 18: Indoor Smoke from Household Use of Solid Fuels.” In Comparative Quantification of Health Risks: The Global Burden of Disease Due to Selected Risk Factors, ed. M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Murray, vol 2, 1435–93. Geneva: World Health Organization. Smith, K. R., G. Shuhua, H. Kun, and Q. Daxiong. 1993. “One Hundred Million Improved Cookstoves in China: How Was It Done?” World Development 21 (6): 941–61. Smith, K. R., R. Uma, V. V. N. Kishore, J. Zhang, V. Joshi, and M. A. K. Khalil. 2000. “Greenhouse Implications of Household Stoves: An Analysis for India.” Annual Review Energy Environment 25: 741–63. Snow, R. W., A. K. Bradley, R. Hayes, P. Byass, and B. M. Greenwood. 1987. “Does Woodsmoke Protect against Malaria?” Annals of Tropical Medical Parasitology 81 (4): 449–51. UNDP (United Nations Development Programme). 2005. “LP Gas Rural Challenge.” http://www.undp.org/energy/lpg.htm. UNDP and ESMAP (United Nations Development Programme and Energy Sector Management Assistance Programme). 2002. India: Household Energy, Indoor Air Pollution, and Health. Delhi: UNDP and World Bank. ———. 2003. Health Impacts of Traditional Fuel Use in Guatemala. Washington, DC: UNDP and World Bank. 814 | Disease Control Priorities in Developing Countries | Nigel Bruce, Eva Rehfuess, Sumi Mehta, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 50 UNICEF (United Nations Children’s Fund). 2004. “Monitoring the Situation of Women and Children: The Multiple Indicator Cluster Survey.” http://www.childinfo.org. United Nations Statistics Division. 2003. “Millennium Development Goals.” http://millenniumindicators.un.org/unsd/mi/mi_goals.asp. Vernede, R., M. M. van Meer, and M. P. Alpers. 1994. “Smoke as a Form of Personal Protection against Mosquitoes: A Field Study in Papua New Guinea.” Southeast Asian Journal of Tropical Medicine and Public Health 25 (4): 771–75. von Schirnding, Y. E. R., N. G. Bruce, K. R. Smith, G. Ballard-Tremeer, M. Ezzati, and K. Lvovsky. 2002. “Addressing the Impact of Household Energy and Indoor Air Pollution on the Health of the Poor: Implications for Policy Action and Intervention Measures.” Paper prepared for the Commission on Macroeconomics and Health, WHO/HDE/HID/02.9. World Health Organization, Geneva. Wang, X., and K. R. Smith. 1999. “Secondary Benefits of Greenhouse Gas Control: Health Impacts in China.” Environmental Science and Technology 33 (18): 3056–61. Westoff, B., and D. Germann. 1995. Stove Images: A Documentation of Improved and Traditional Stoves in Africa, Asia and Latin America. Brussels: Commission of the European Communities, DirectorateGeneral for Development. WHO (World Health Organization). 2002a. “Addressing the Links between Indoor Air Pollution, Household Energy, and Human Health: Based on the WHO-USAID Global Consultation on the Health Impact of Indoor Air Pollution and Household Energy in Developing Countries.” Meeting Report WHO/HDE/HID/02.10. WHO, Geneva. ———. 2002b. Reducing Risks, Promoting Healthy Life: World Health Report 2002. Geneva: WHO. ———. 2003. Making Choices in Health: WHO Guide to Cost-Effectiveness Analysis. Geneva: WHO. ———. 2004a. “Indoor Air Thematic Briefing 1: Indoor Air Pollution, Household Energy and the Millennium Development Goals.” Geneva: WHO. http://www.who.int/indoorair/info/en/iabriefing1rev.pdf. ———. 2004b. “Burden of Disease Statistics. Burden of Disease Unit.” Geneva: WHO. http://www.who.int/research/en/. ———. 2004c. “Evidence for Policy Makers: Indoor Air Pollution.” Geneva: WHO. http://www.who.int/indoorair/policy/en/. ———. 2004d. “The World Health Survey.” Geneva: WHO. http://www3. who.int/whs/. ———. 2004e. “From Theory to Action: Implementing the WSSD Global Initiative on Children’s Environmental Health Indicators.” Geneva: WHO. http://www.who.int/ceh/publications/924159188_9/en/. ———. 2005. “Development of a Catalogue of Methods: Indoor Air Pollution.” Geneva: WHO. http://www.who.int/indoorair/interventions/ methodology/en/. Wickramasinghe, A. 2001. “Gendered Sights and Health Issues in the Paradigm of Biofuel in Sri Lanka.” Energia News 4 (4): 14–16. World Bank. 2002. “Energy Efficiency and Emissions.” In World Development Indicators. Washington, DC: World Bank. Yach, D. 1994. “Paraffin Poisoning: Partnership Is the Key to Prevention.” South African Medical Journal 84 (11): 717. Indoor Air Pollution | 815 ©2006 The International Bank for Reconstruction and Development / The World Bank 51 ©2006 The International Bank for Reconstruction and Development / The World Bank 52 Chapter 44 Prevention of Chronic Disease by Means of Diet and Lifestyle Changes Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, Courtenay Dusenbury, Pekka Puska, and Thomas A. Gaziano Coronary artery disease (CAD), ischemic stroke, diabetes, and some specific cancers, which until recently were common only in high-income countries, are now becoming the dominant sources of morbidity and mortality worldwide (WHO 2002). In addition, rates of cancers and cardiovascular disease (CVD) among migrants from low-risk to high-risk countries almost always increase dramatically. In traditional African societies, for example, CAD is virtually nonexistent, but rates among African Americans are similar to those among Caucasian Americans. These striking changes in rates within countries over time and among migrating populations indicate that the primary determinants of these diseases are not genetic but environmental factors, including diet and lifestyle. Thus, considerable research has been aimed at identifying modifiable determinants of chronic diseases. Prospective epidemiological studies, some randomized prevention trials, and many short-term studies of intermediate endpoints such as blood pressure and lipids have revealed a good deal about the specific dietary and lifestyle determinants of major chronic diseases. Most of these studies have been conducted in Western countries, in part because of the historical importance of these diseases in the West, but also because they have the most developed research infrastructure. A general conclusion is that reducing identified, modifiable dietary and lifestyle risk factors could prevent most cases of CAD, stroke, diabetes, and many cancers among high-income populations (Willett 2002). These findings are profoundly important, because they indicate that these diseases are not inevitable consequences of a modern society. Furthermore, low rates of these diseases can be attained without drugs or expensive medical facilities, an outcome that is not surprising, because their rates have historically been extremely low in developing countries with few medical facilities. However, preventing these diseases will require changes in behaviors related to smoking, physical activity, and diet; investments in education, food policies, and urban physical infrastructure are needed to support and encourage these changes (see box 44.1). CHRONIC DISEASE PREVENTION In this section, we briefly review dietary and lifestyle changes that reduce the incidence of chronic disease. The potential magnitude of benefit is also discussed. Recommended Lifestyle Changes Specific changes in diet and lifestyle and likely benefits are summarized in table 44.1. These relationships and supporting evidence are summarized here. Avoid Tobacco Use. Avoidance of smoking by preventing initiation or by cessation for those who already smoke is the single most important way to prevent CVD and cancer (chapter 46). Avoiding the use of smokeless tobacco will also prevent a good deal of oral cancer. Maintain a Healthy Weight. Obesity is increasing rapidly worldwide (chapter 45). Even though obesity—a body mass index (BMI) of 30 or greater—has received more attention 833 ©2006 The International Bank for Reconstruction and Development / The World Bank 53 Box 44.1 The Insulin Resistance Syndrome In recent years, researchers have recognized the insulin resistance syndrome (also known as the metabolic syndrome) as a common contributing factor to the development of diabetes, CAD, and some cancers. The syndrome is characterized by increased waist circumference, low HDL (high-density lipoprotein) cholesterol, high levels of triglycerides, hypertension, and glucose intolerance. The most direct causes are overweight and inactivity, but dietary factors contribute. Genetic factors, which are probably beneficial during periods of food shortages, also play a role. Recent evidence indicates that the populations of Asia, Latin America, and probably Africa are particularly susceptible (Dickinson and others 2002; Harris and others 1998). Table 44.1 Convincing and Probable Relationships between Dietary and Lifestyle Factors and Chronic Diseases Dietary and lifestyle factors CVD Type 2 diabetes Cancer Dental disease Fracture Cataract Birth defects Obesity Metabolic syndrome Depression Sexual dysfunction Avoid smoking Pursue physical activity Avoid overweight Diet Consume healthy types of fatsa Eat plenty of fruits and vegetables Replace refined grains with whole grains Limit sugar intakeb Limit excessive calories Limit sodium intake Source: Authors’ summary of a review by the WHO and FAO 2003; Bacon and others 2003; Fox 1999; IARC 2002. Note: Bold ⫽ convincing; Standard ⫽ probable relation; ↑ ⫽ increase in risk; ↓ ⫽ decrease in risk. a. Replace trans and saturated fats with mono- and polyunsaturated fats, including a regular source of N-3 fatty acids. b. Includes limiting sugar-based beverages. than overweight, overweight (BMI of 25 to 30) is typically even more prevalent and also confers elevated risks of many diseases. For example, overweight people experience a two- to threefold elevation in the risks of CAD and hypertension and a more than tenfold increase in the risk of type 2 diabetes compared with lean individuals (BMI less than 23) (Willett, Dietz, and Colditz 1999). Both overweight and obese people also experience elevated mortality from cancers of the colon, breast (postmenopausal), kidney, endometrium, and other sites (Calle and others 2003). Many people with a BMI of less than 25 have gained substantial weight since they were young adults and are also at increased risk of these diseases, even though they are not technically overweight (Willett, Dietz, and Colditz 1999). For 834 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 54 example, in rural China, where the average BMI was less than 21 for both men and women, F. B. Hu and others (2000) found that the prevalence of hypertension was nearly five times greater for those with a BMI of approximately 25 than for the leanest people. Because many Asians are experiencing adverse consequences of excess body fat with a BMI of less than 25, the definition of overweight for Asia has recently been expanded to include a BMI of 23 to 25 (WHO 2000). For most people, unless obviously malnourished as an adolescent or young adult, bodyweight should ideally not increase by more than 2 or 3 kilograms after age 20 to maintain optimal health (Willett, Dietz, and Colditz 1999). Thus, a desirable weight for most people should be within the BMI range of 18.5 to 25.0, and preferably less than 23. Additional valuable information can be obtained by measuring waist circumference, which reflects abdominal fat accumulation. In many studies, waist circumference is a strong predictor of CAD, stroke, and type 2 diabetes, even after controlling for BMI (Willett, Dietz, and Colditz 1999). A waist circumference of approximately 100 centimeters for men and 88 centimeters for women has been used as the criterion for the upper limit of the healthy range in the United States, but for many people this extent of abdominal fat would be far above optimal. Because abdominal circumference is easily assessed, even where scales may not be available, further work to develop locally appropriate criteria could be worthwhile. In the meantime, increases of more than 5 centimeters can be used as a basis for recommending changes in activity patterns and diet. Views about the causes of obesity and ways to prevent or reduce it have been controversial. Diets low in fat and high in carbohydrates were believed to limit caloric intake spontaneously and thus to control adiposity, but such diets have not reduced bodyweight in trials that have lasted for a year or more (Willett and Leibel 2002). Some researchers have suggested that diets with a high energy density, referring to the amount of energy per volume, offer an alternative explanation for the observed increases in obesity (Swinburn and others 2004), but long-term studies have not examined this theory. Sugarsweetened beverages contribute significantly to the overconsumption of calories, in part because calories in fluid form appear to be poorly regulated by the body (E. A. Bell, Roe, and Rolls 2003). In children, an increase in soda consumption of one serving per day was associated with an odds ratio of 1.6 for incidence of obesity (Ludwig, Peterson, and Gortmaker 2001), and in a randomized trial, replacement of a standard soda with a zero-calorie diet soda was associated with significant weight loss (Raben and others 2002). Reductions in dietary fiber and increases in the dietary glycemic load (large amounts of rapidly absorbed carbohydrates from refined starches and sugar) may also contribute to obesity (Ebbeling and others 2003; Swinburn and others 2004). Aspects of the food supply unrelated to its macronutrient composition are also likely to be contributing to the global rise in obesity. Inexpensive food energy from refined grains, sugar, and vegetable oils has become extremely plentiful in most countries. Food manufacturers and suppliers use carefully researched methods to make products based on these cheap ingredients maximally convenient and attractive. Maintain Daily Physical Activity and Limit Television Watching. Contemporary life in developed nations has markedly reduced people’s opportunities to expend energy, whether in moving from place to place, in the work environment, or at home (Koplan and Dietz 1999). Dramatic reductions in physical activity are also occurring in developing countries because of urbanization, increased availability of motorized transportation to replace walking and bicycle riding, and mechanization of labor. However, regular physical activity is a key element in weight control and prevention of obesity (IARC 2002; Swinburn and others 2004). For example, among middle-aged West African women, more walking was associated with a three-unit lower BMI (Sobngwi, Gautier, and Mbanya 2003), and in China, car owners are 80 percent more likely to be obese (Hu 2002). In addition to its key role in maintaining a healthy weight, regular physical activity reduces the risk of CAD, stroke, type 2 diabetes, colon and breast cancer, osteoporotic fractures, osteoarthritis, depression, and erectile dysfunction (table 44.1). Important health benefits have even been associated with walking for half an hour per day, but greater reductions in risk are seen with longer durations of physical activity and more intense activity. The number of hours of television watched per day is associated with increased obesity rates among both children and adults (Hernandez and others 1999; Ruangdaraganon and others 2002) and with a higher risk of type 2 diabetes and gallstones (F. B. Hu and others 2001; Leitzmann and others 1999). This association is likely attributable both to reduced physical activity and to increased consumption of foods and beverages high in calories, which are typically those promoted on television. Decreases in television watching reduce weight (Robinson 1999), and the American Academy of Pediatrics recommends a maximum of two hours of television watching per day. Eat a Healthy Diet. Medical experts have long recognized the effects of diet on the risk of CVD, but the relationship between diet and many other conditions, including specific cancers, diabetes, cataracts, macular degeneration, cholelithiasis, renal stones, dental disease, and birth defects, have been documented more recently. The following list discusses six aspects of diet for which strong evidence indicates important health implications (table 44.1). These goals are consistent with a detailed Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 835 ©2006 The International Bank for Reconstruction and Development / The World Bank 55 2003 World Health Organization (WHO) report (WHO and FAO 2003). • Replace saturated and trans fats with unsaturated fats, including sources of omega-3 fatty acids. Replacing saturated fats with unsaturated fats will reduce the risk of CAD (F. B. Hu and Willett 2002; Institute of Medicine 2002; WHO and FAO 2003) by reducing serum low-density lipoprotein (LDL) cholesterol. Also, polyunsaturated fats (including the long-chain omega-3 fish oils and probably alpha-linoleic acid, the primary plant omega-3 fatty acid) can prevent ventricular arrhythmias and thereby reduce fatal CAD. In a case-control study in Costa Rica, where fish intake was extremely low, the risk of myocardial infarction was 80 percent lower in those with the highest alpha-linoleic acid intake (Baylin and others 2003). Intakes of omega-3 fatty acids are suboptimal in many populations, particularly if fish intake is low and the primary oils consumed are low in omega-3 fatty acids (for example, partially hydrogenated soybean, corn, sunflower, or palm oil). These findings have major implications, because changes in the type of oil used for food preparation are often quite feasible and not expensive. Trans fatty acids produced by the partial hydrogenation of vegetable oils have uniquely adverse effects on blood lipids (F. B. Hu and Willett 2002; Institute of Medicine 2002) and increase risks of CAD (F. B. Hu and Willett 2002); on a gram-for-gram basis, both the effects on blood lipids and the relationship with CAD risk are considerably more adverse than for saturated fat. In many developing countries, trans fat consumption is high because partially hydrogenated soybean oil is among the cheapest fats available. In South Asia, vegetable ghee, which has largely replaced traditional ghee, contains approximately 50 percent trans fatty acids (Ascherio and others 1996). Independent of other risk factors, higher intakes of trans fat and lower intakes of polyunsaturated fat increase risk of type 2 diabetes (F. B. Hu, van Dam, and Liu 2001). • Ensure generous consumption of fruits and vegetables and adequate folic acid intake. Strong evidence indicates that high intakes of fruits and vegetables will reduce the risk of CAD and stroke (Conlin 1999). Some of this benefit is mediated by higher intakes of potassium, but folic acid probably also plays a role (F. B. Hu and Willett 2002). Supplementation with folic acid reduces the risk of neural tube defect pregnancies. Substantial evidence also suggests that low folic acid intake is associated with greater risk of colon—and possibly breast—cancer and that use of multiple vitamins containing folic acid reduces the risk of these cancers (Giovannucci 2002). Findings relating folic acid intake to CVD and some cancers have major implications for many parts of the developing world. In many areas, consumption • • • • of fruits and vegetables is low. For example, in northern China, approximately half the adult population is deficient in folic acid (Hao and others 2003). Consume cereal products in their whole-grain, high-fiber form. Consuming grains in a whole-grain, high-fiber form has double benefits. First, consumption of fiber from cereal products has consistently been associated with lower risks of CAD and type 2 diabetes (F. B. Hu, van Dam, and Liu 2001; F. B. Hu and Willett 2002), which may be because of both the fiber itself and the vitamins and minerals naturally present in whole grains. High consumption of refined starches exacerbates the metabolic syndrome and is associated with higher risks of CAD (F. B. Hu and Willett 2002) and type 2 diabetes (F. B. Hu, van Dam, and Liu 2001). Second, higher consumption of dietary fiber also appears to facilitate weight control (Swinburn and others 2004) and helps prevent constipation. Limit consumption of sugar and sugar-based beverages. Sugar (free sugars refined from sugarcane or sugar beets and highfructose corn sweeteners) has no nutritional value except for calories and, thus, has negative health implications for those at risk of overweight. Furthermore, sugar contributes to the dietary glycemic load, which exacerbates the metabolic syndrome and is related to the risk of diabetes and CAD (F. B. Hu, van Dam, and Liu 2001; F. B. Hu and Willett 2002; Schulze and others 2004). WHO has suggested an upper limit of 10 percent of energy from sugar, but lower intakes are usually desirable because of the adverse metabolic effects and empty calories. Limit excessive caloric intake from any source. Given the importance of obesity and overweight in the causation of many chronic diseases, avoiding excessive consumption of energy from any source is fundamentally important. Because calories consumed as beverages are less wellregulated than calories from solid food, limiting the consumption of sugar-sweetened beverages is particularly important. Limit sodium intake. The principle justification for limiting sodium is its effect on blood pressure, a major risk factor for stroke and coronary disease (chapter 33). WHO has suggested an upper limit of 1.7 grams of sodium per day (5 grams of salt per day) (WHO and FAO 2003). Potential of Dietary and Lifestyle Factors to Prevent Chronic Diseases Several lines of evidence indicate that realistic modifications of diet and lifestyle can prevent most CAD, stroke, diabetes, colon cancer, and smoking-related cancers. Less progress has been made in identifying practically modifiable causes of breast and prostate cancers. 836 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 56 Box 44.2 Success in Finland Finland provides one of the best-documented examples of a community intervention. In 1972, Finland had the world’s highest CVD mortality rate. Planners examined the policy and environmental factors contributing to CVD and sought appropriate changes, such as increased availability of low-fat dairy products, antismoking legislation, and improved school meals. They used the media; schools; worksites; and spokespersons from sports, education, and agriculture to educate residents. After five years, signifi- One line of evidence is based on declines in CAD in countries that have implemented preventive programs. Rates of CAD mortality have been cut in half in several high-income countries, including Australia, the United Kingdom, and the United States. The most dramatic example is that of Finland (box 44.2). Other evidence derives from randomized intervention studies. These often have serious limitations for estimating the potential magnitude of benefits, because typically only one or a few factors are modified, durations are usually only a few years, and noncompliance with lifestyle change is often substantial. Nevertheless, some examples are illustrative of the potential benefit. In two randomized studies among adults at high risk of type 2 diabetes, those assigned to a program emphasizing dietary changes, weight loss, and physical activity experienced only half the risk of incident diabetes (Knowler and others 2002; Tuomilehto and others 2001). The Lyon Heart Study, conducted among those with existing heart disease, found a Mediterranean-type diet high in omega-3 fatty acids reduced recurrent infarction by 70 percent compared with an American Heart Association diet (de Lorgeril and others 1994). A third approach is to estimate the percentage of disease that is potentially preventable by reducing multiple behavioral risk factors using prospective cohort studies. Among U.S. adults, more than 90 percent of type 2 diabetes, 80 percent of CAD, 70 percent of stroke, and 70 percent of colon cancer are potentially preventable by a combination of nonsmoking, avoidance of overweight, moderate physical activity, healthy diet, and moderate alcohol consumption (Willett 2002). Collectively, these findings indicate that the low rates of these diseases suggested by international comparisons and time trends are attainable by realistic, moderate changes that are compatible with 21st-century lifestyles. cant improvements were documented in smoking, cholesterol, and blood pressure. By 1992, CVD mortality rates for men age 35 to 64 had dropped by 57 percent. The program was so successful that it was expanded to include other lifestyle-related diseases. Twenty years later, major reductions in CVD risk-factor levels, morbidity, and mortality were attributed to the project. Recent data show a 75 percent decrease in CVD mortality (Puska and others 1998). INTERVENTIONS Interventions aimed at changing diet and lifestyle factors include educating individuals, changing the environment, modifying the food supply, undertaking community interventions, and implementing economic policies. In most cases, quantifying the effects of the intervention is difficult, because behavioral changes may take many years and synergies are potentially important but hard to estimate in formal studies. Substantial nihilism often exists regarding the ability to change populations’ diets or behaviors, but major changes are possible over extended periods of time. For example, per capita egg consumption in the United States decreased from approximately 420 to 270 per year between 1940 and 1990 following recommendations for preventing CAD (though in reality, the evidence for benefits was meager). Similarly, the prevalence of smoking, despite its being a physically addictive behavior, halved among men in the United States between 1965 and 2000. Because changing behaviors related to diet and lifestyle require sustained efforts, long-term persistence is needed. However, opportunities exist that do not require individual behavior changes, and these can lead to more rapid benefits. Educational Interventions Efforts to change diets, physical activity patterns, and other aspects of lifestyle have traditionally attempted to educate individuals through schools, health care providers, worksites, and general media. These efforts will continue to play an important role, but they can be strongly reinforced by policy and environmental changes. School-Based Programs. School-based programs include the roles of nutrition and physical activity in maintaining physical Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 837 ©2006 The International Bank for Reconstruction and Development / The World Bank 57 Box 44.3 The Planet Health Program Planet Health, developed for middle school students, in the United States, has an immediate goal of reducing television viewing time with the long-range goal of preventing unhealthy weight gain (Gortmaker and others 1999). Teachers incorporate messages about reducing television watching, nutrition, and increasing fitness into mathematics, social studies, science, and language arts lessons. Fitness units and periodic “FitChecks” during physical education complement the classroom lessons. Teacher training, student self-assessment using graphs, and student reflection about enjoyable activities that could replace at least a portion of the time they spend watching television are key elements. This program has reduced television watching and weight in girls (Gortmaker and others 1999). Because the program is integrated into existing classes, its cost is minimal. Box 44.4 Live for Life® Johnson & Johnson introduced Live for Life in 1979 with the goal of making its employees the healthiest in the world (Bly, Jones, and Richardson 1986). In 1993, the company integrated its health and wellness program with its disability management, employee assistance, and occupational medicine programs. Instead of using physicians and nurses to treat symptoms, the combined program sought to use a variety of health professionals to change individual behavior and improve health status. Employees and mental health (box 44.3). School food services should provide healthy meals, both because they directly affect health and because they provide a special opportunity to teach by example. In many countries, school-based physical education remains a significant source of physical activity for young people. In China, 72 percent of children age 6 to 18 engage in moderate to vigorous physical activity for a median of 90 to 100 minutes per week (Tudor-Locke and others 2003). Maintaining these programs should be a high priority because they have likely contributed to the historically low rates of obesity in such countries. Worksite Interventions. Worksite interventions can efficiently include a wide variety of health promotion activities because workers spend a large portion of their waking hours and eat a large percentage of their food there. Interventions can include educating employees; screening them for behavioral risk factors; offering incentive programs to walk, ride a bicycle, or take public transportation to work; offering exercise were offered US$500 in benefit credits for participation. The program included routine health risk assessment, health promotion after recovery from a medical event, and support when returning to work after a major illness. Even though the intervention program had little effect on bodyweight, physical fitness did increase. By the end of the third year, savings to the company were more than US$400 per year per employee. programs during breaks or after work; improving the physical environment to promote activity; and providing healthier foods in cafeterias (box 44.4). Worksite health promotion can result in a positive return on investment through lower health costs and fewer sick days. Interventions by Health Care Providers. Controlled intervention trials for smoking cessation and physical activity have shown that physician counseling, especially when accompanied by supporting written material, can be efficacious in modifying behavior. Studies of dietary counseling by physicians indicate that even brief messages about nutrition can influence behavior and that the magnitude of the effect is related to the intensity of the intervention (Pignone and others 2003). Identifying patients who are overweight or obese, or who are gaining weight but are not yet overweight, is an initial step in preventing and treating overweight. However, many physicians are not well trained to measure and calculate BMI and identify weight problems. 838 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 58 Box 44.5 Reducing Automobile Use in Brazil Curitiba, Brazil, provides an example of the benefits of a strategy that reduces automobile use and increases use of public transportation. In 1965, city planners adopted a master plan that promoted development along designated corridors along with a bus system so efficient that it has virtually eliminated the need for automobiles. Minibuses Transportation Policy and Environmental Design Transportation policies and the design of urban environments are fundamental determinants of physical activity and therefore influence the risks of obesity and other chronic diseases. Countries can take a number of steps to make positive changes. Limit the Role of Automobiles. In wealthy countries, the automobile has strongly influenced the trend toward lowdensity, automobile-based suburban developments, many built without sidewalks. These sprawling settlements tend to have few services within walking distance and are usually not linked to public transporationt. Dependence on automobiles affects physical activity, because those who use public transportation tend to walk more. In a prospective study in eight provinces in China, 14 percent of households acquired a car between 1889 and 1997, and the likelihood of men becoming obese during the same period was twice as great in households that acquired a car than in those that did not (A. C. Bell, Ge, and Popkin 2002). National policies strongly influence automobile use and dependency. In the United States, low taxes on gasoline, free parking, and wide streets encourage car ownership: almost 92 percent of U.S. households own at least one car, and 59 percent own two or more cars (Pucher and Dijkstra 2003). In contrast, in most of Western Europe, narrow streets, limited parking, and high gasoline prices make the costs of automobile use almost double those in the United States (Pucher and Dijkstra 2003). As a result, Europeans walk or bike more and use their cars approximately 50 percent less than their American counterparts. Investment in roads rather than in public transportation creates a vicious cycle: poor public transportation systems lead to more dependency on the automobile. As car use grows, injuries and deaths associated with automobile accidents also grow. In China, the number of fourwheeled vehicles increased from about 60,000 to more than 50 million between 1951 and 1999, and traffic fatalities increased from about 6,000 to more than 413,000 (S. Y. Wang and others 2003). Many innovative strategies have been are used to quickly and efficiently transport individuals from residential neighborhoods to express bus lines. These bus lines run almost every 90 seconds and can carry up to 270 passengers each. Compared with other Brazilian cities of its size, Curitiba uses 30 percent less gasoline per capita, and its air pollution is among the lowest in the nation. developed to discourage private automobile use and to promote public transportation, walking, and bicycling (see box 44.5). Singapore has long been in the lead in relation to such efforts: a combination of limiting the number of licenses issued, implementing a vehicle quota system, and introducing a road pricing system has limited personal car ownership and congestion throughout the country. Other nations and regions are now enacting similar road pricing systems or congestion taxes. For example, London’s congestion charging system levies a fee of approximately US$8 per day for cars entering central London. Since its inception in 2003, the charge has reduced congestion in the city and is expected to channel funds back into the city’s transportation facilities. Unfortunately some countries, particularly China, have taken a different approach to their future transportation needs. Government initiatives that encourage families to buy automobiles include lowering taxes, simplifying registration procedures, and allowing foreign financing. In Beijing alone, residents purchased 400,000 cars in 2003. Promote Walking and Bicycle Riding. Walking or cycling for transportation and leisure are effective and practical means of engaging in physical activity and are still the most common ways to travel in many developing countries. In Bangkok and Manila, only 25 percent of travel is by car, motorcycle, or taxi, compared with 75 percent by public transportation or walking (Pendakur 2000). In Madras, India, only 8 percent of the population travels by private, motorized transportation; 22 percent of people walk; 20 percent bike; and the rest use public transportation (Pendakur 2000). In China, approximately 90 percent of the urban population walks or rides a bicycle to work, shopping, or school each day (G. Hu and others 2002). Walking or biking is more likely to be prevalent in smaller cities—that is, those with 1 million to 5 million people—than in larger ones. Bicycle riding and walking are also important for children’s health. Most American children do not walk or bike to school, even when distances are short (box 44.6). In contrast, almost Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 839 ©2006 The International Bank for Reconstruction and Development / The World Bank 59 Box 44.6 Walking and Cycling to School One of the most effective ways to promote walking and cycling is through local schools. The Safe Routes to School program (http://www.saferoutestoschools.org/), established in Marin County, California, is a private-public partnership that created a citywide map of safe biking and 90 percent of Chinese children under 12 walk or ride a bicycle to school (Hu 2002). In many areas, the shift toward private car use has not yet begun and can perhaps be forestalled by policies that benefit walkers and cyclists rather than drivers. Such policies include implementing road designs that promote a safe and well-lit environment for walking and cycling, including traffic-calming measures to reduce automobile speeds. Many Western European countries have taken steps to increase safety for cyclists and walkers. In Germany and the Netherlands, bike paths serve as travel routes, not just weekend recreational destinations as they do in the United States. The former countries have invested heavily in bike paths and have also created extensive car-free areas in cities, with well-lit sidewalks, clearly marked crosswalks, and pedestrian islands that have improved safety. Both countries have increased the number of bicycle-friendly streets (on which cars are permitted but bicycles have the right of way) and have created systems to separate streams of traffic, including cars, pedestrians, and bicycles. A meta-analysis of selected traffic-calming studies in many countries reported reductions in traffic speed, accidents, injuries, and fatalities and an increase in bicycle use and walking (Bunn and others 2003). Design Cities and Towns to Promote Health. Handy and others’ (2002) comprehensive assessment of recent research on urban planning concludes that a combination of urban design, land-use patterns, and transportation systems that promotes walking and bicycling will help create active, healthier, and more livable communities. In densely developed cities that have been built around public transportation rather than away from it, individuals are much more likely to take public transit, walk, or bicycle than in other areas and to weigh less and be less likely to suffer from hypertension (Ewing, Schieber, and Zegeer 2003; Lopez 2004; Saelens, Sallis, and Frank 2003). Those living in walker-friendly neighborhoods also appear to be more mentally healthy and are more likely to know their neighbors, to be socially active, and to participate in the political process (Leyden 2003). In contrast, urban sprawl has been walking routes and proposed solutions for problem areas. The program also sponsors walk- and bike-to-school days, frequent-rider contests, and other promotional events (Staunton, Hubsmith, and Kallins 2003). linked to decreases in mental health and social capital (Frumkin 2002) as well as anger and frustration over long commutes (Surface Transportation Policy Project 1999). Sprawl adversely affects the elderly in particular because they are unable to walk to places of interest and many cannot drive. Such isolation does not promote good physical or mental health. The so-called smart growth movement has resulted from concerns about urban sprawl and unsustainable development and is encouraging governments worldwide to rethink how they develop new areas and redevelop older suburbs and cities. Smart growth principles include mixing land uses, using compact building designs, including a range of transportation and housing choices, building walker-friendly neighborhoods in attractive communities with a distinctive sense of place, and implementing a philosophy of directing development toward existing communities and the preservation of open space (Office of the Administrator 2001) (box 44.7). The involvement of public health practitioners in transportation planning and building design is becoming more common. In Edinburgh, a health impact assessment conducted on proposed options for transportation policy showed the effects of specific choices on both affluent members of the community and the poor. Its recommendations, now adopted, included new spending on pedestrian safety, a citywide bicycle network, more greenways and park-and-ride programs, and more rail transportation or bus services. Priorities are to benefit pedestrians first, cyclists second, public transportation users third, freight and delivery people fourth, and car users last. Establishing criteria for building design can also lead to increases in physical activity. For example, increasing signage promoting stair use, as well as the attractiveness of the facilities themselves, encourages people to use the stairs (Boutelle and others 2001) (box 44.8). Improved Food Supply People’s diets can be enhanced by improving the food supply. The usual position of the food industry is that it simply 840 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 60 Box 44.7 Enhancing Urban Life in the Republic of Korea In Seoul, the government is managing growth by creating six satellite communities with high-rise residential buildings outside the city center. These communities are intended to become new job-creation centers and to shift the balance of employment away from one centralized location to provide a more regional balance. Major expressways are being removed to create parks, sidewalks, and bikeways (http://www.itdp.org/STe/ste6/#seoul). Box 44.8 Promoting Physical Activity in Brazil One successful example of increasing activity is Agita São Paulo, a multilevel physical activity initiative designed for the 34 million citizens of Brazil’s São Paulo state (Matsudo and others 2002). The program was launched in 1996 to increase the public’s knowledge of the benefits of exercise and expand participation in physical fitness activities by encouraging people to do 30 minutes of moderate activity at least five times a week. As elsewhere, program designers perceived a lack of time as the major factor preventing daily exercise. They chose three settings as places to promote activity: home (gardening, chores, avoidance of television watching); transportation (walking, taking the stairs); and leisure time (dancing). Agitol, a prescription for exercise, was developed for physicians to dispense. Its provides whatever consumers demand, but this argument is misleading, because the industry spends more than US$12 billion annually to influence consumer choices just within the United States and many times this amount globally. Much of this sum goes to promote foods with adverse health effects, and children are primary targets. Improving Processing and Manufacturing. Altering the manufacturing process can rapidly and effectively improve diets because such action does not require the slow process of behavioral change. One example is eliminating the partial hydrogenation of vegetable oils, which destroys essential omega-3 fatty acids and creates trans fatty acids. European manufacturers have largely eliminated trans fatty acids from their food supply by altering production methods. Regulations can facilitate changes in manufacturing directly or indirectly by providing an incentive for manufacturers to change their processes. For example, in 2003, the U.S. Food and message is displayed on electricity bills and stickers, and it is touted by radio stations and other media outlets. After four years, 55.7 percent of those surveyed had heard about Agita, 37 percent knew its purpose, and those who knew of the program’s purpose were more likely to be active. Agita appears to have played a role in increasing activity in the region (Matsudo and others 2002). It is closely linked to a national program to promote healthy diets and active lifestyles by nutritional content labeling, promotion of healthy diets in schools, communication of guidelines for healthy eating, and encouragement of innovative community-based initiatives (Coitinho, Monteiro, and Popkin 2002). Drug Administration announced that food manufacturers had to include trans fatty acid content on the standard food label. Following imposition of this requirement, several large food companies said that they would reduce or eliminate trans fats, and many more are planning to do so (U.S. Food and Drug Administration 2003). In Mauritius, the government required a change in the commonly used cooking oil from mostly palm oil to soybean oil, which changed people’s fatty acid intake and reduced their serum cholesterol levels (Uusitalo and others 1996). Changes in types of fat can often be almost invisible and inexpensive. Omega-3 fatty acid intakes can be increased by incorporating oils from rapeseed, mustard, or soybean into manufactured foods, cooking oils sold for use at home, or both. Selective breeding and genetic engineering provide alternative ways to improve the healthfulness of oils by modifying their fatty acid composition. When the consumption of processed food is high, a reduction in salt consumption will usually require changes at the Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 841 ©2006 The International Bank for Reconstruction and Development / The World Bank 61 manufacturing level, because processed food is a major salt source. If the salt content of foods is reduced gradually, the change is imperceptible to consumers. Coordination among manufacturers or government regulation is needed; otherwise producers whose foods are lower in salt may be placed at a disadvantage. Unfortunately, good examples are not available. Another example of improved processing would be to reduce the refining of grain products, which can be done in small, almost invisible decrements. Fortifying Food. Food fortification has eliminated iodine deficiency, pellagra, and beriberi in much of the world. In regions where iodine deficiency remains a serious problem, fortification should be a high priority. Folic acid intake is suboptimal in many regions of both developing and developed countries. Fortifying foods with folic acid is extremely inexpensive and could substantially reduce the rates of several chronic diseases. Grain products—such as flour, rice, and pasta—are usually the best foods to fortify, and in many countries, they are already being fortified with other B vitamins. Since 1998, grain products in the United States have been fortified with folic acid, which has almost eliminated folate deficiency, and rates of neural tube defect pregnancies have declined by about 19 percent (Honein and others 2001). Where intakes of vitamins B12 and B6 are also low and contribute to elevations of homocysteine, as among vegetarian populations in India, simultaneous fortification of food with these vitamins should be considered. The effects of fortification on reducing CVD are not considered proven, but the potential benefits are huge; therefore, intervention trials to evaluate the effects of fortification should be a high priority. Increasing the Availability and Reducing the Cost of Healthy Foods. Policies regarding the production, importation, distribution, and sale of specific foods can influence their cost and availability. Policies may be directed at the focus of agricultural research and the types of production promoted by extension services. Policies often promote grains, dairy products, sugar, and beef, whereas those that encourage the production and consumption of fruits, vegetables, nuts, legumes, whole grains, and healthy oils would tend to enhance rather than reduce health. Promoting Healthy Food Choices and Limiting Aggressive Marketing to Children. Almost every national effort to improve nutrition incorporates the promotion of healthy food choices, such as fruits, vegetables, and legumes. Ideally, such efforts are coordinated among government groups, retailers, professional groups, and nonprofit organizations, and investment in such efforts should include the careful testing and refining of social-marketing strategies. Another strategy is to protect consumers from aggressive marketing of unhealthy foods. Producers spend billions of dollars a year encouraging children to consume foods that are detrimental to their health. Manufacturers and fast-food chains personify food products with cartoon characters; display food brands on toys; and issue “educational” card games that subvert children’s natural gift for play, story telling, and make believe. The willingness to limit advertising depends on a country’s political culture, but the public clearly distinguishes between advertising aimed at adults and that targeted at children. For example, in the United States, 46 percent of adults surveyed supported restrictions on advertising to children (Blendon 2002). Restrictions can range from banning advertising to children to limiting the types of products that advertisers may promote to this audience. Initiatives at the Community Level Nations and regions can promote a variety of initiatives to encourage greater physical activity and better nutrition. These initiatives are likely to be most effective when they are multifaceted and coordinated and when they are developed with the active involvement of individuals and organizations within communities (Puska and others 1998). Many countries are undertaking efforts to educate their populations about healthy lifestyles. In the Islamic Republic of Iran, the Isfahan Healthy Heart Program, a WHO collaborating center for research and training for CVD control, prevention, and rehabilitation for cardiac patients, has developed a comprehensive, integrated community intervention that involves schools, worksites, health care facilities, food services, urban planners, and the media. Physical activity is promoted by creating safe routes for walking and bicycle riding and by organizing recreational walking that involves entire families (http://ihhp.mui.ac.ir). South Africa’s Community Health Intervention Programme, a partnership between an insurance company and an academic institution, has created programs targeted to specific age groups, including children and older adults. The program’s twice-weekly classes have reduced blood pressure and increased strength and balance (Lambert, Bohlmann, and KolbeAlexander 2001) (box 44.9). Singapore’s Fit and Trim Program uses a multidisciplinary approach to increase physical activity and healthy diets among schoolchildren. Between 1992 and 2000, the rate of obesity declined by 13.1 to 16.6 percent for children age 11 to 12 and 15 to 16 (Toh, Cutter, and Chew 2002) (box 44.10 outlines the national program for adults). Economic Policies Economic policies can have important effects on behavior and choices, and these policies have been particularly useful in 842 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 62 Box 44.9 A Comprehensive Intervention Approach in South Africa The Coronary Risk Factor Study in South Africa (Rossouw and others 1993) tested community interventions at different levels of intensity in two communities with a third control community. The target population was Caucasian South Africans. Interventions included direct media campaigns, public health messages delivered in a variety of ways, and home mailings. Also included were community activities, such as fun walks, public meetings, involvement of community-based organizations, free screening for blood pressure, small-group personal interventions, and encouragement of food substitution in stores and restaurants. The results showed an improvement in the community risk factor profile for CAD in the intervention communities, especially in relation to blood pressure, smoking, and overall risk. The results indicate no additional benefit of the personal intervention for high-risk individuals beyond that already offered by the mass media program. Estimated per capita costs of the heavy intervention program were roughly four times as much as for the mild intervention program (US$22 per capita compared with US$5 per capita), and the low-intervention community received almost the same level of benefits as the high-intervention community. Box 44.10 The Singapore National Healthy Lifestyle Program Because CVD and cancer had become the major causes of death in Singapore, the government adopted the National Healthy Lifestyle Program in 1992 (Cutter, Tan, and Chew 2001). This coordinated, multisectoral approach involved government ministries, health professionals, employers, unions, and community organizations. The program aimed at improving the social and physical environment so as to promote healthy living. Healthy diets, regular physical exercise, and nonsmoking were emphasized. The program used the mass media; legislative measures to discourage smoking; and widespread school, workplace, and community health promotion packages. reducing the prevalence of smoking (see chapter 46). Policies that could influence diet and physical activity deserve careful consideration because they are rarely neutral and often support unhealthy behaviors. Consider the following examples: • Subsidies can favor the consumption of less healthy foods, such as sugar, refined grains, beef, and high-fat dairy products as opposed to fruits, vegetables, whole grains, nuts, legumes, and fish. Poland provides a striking example of how changes in subsidies can affect health (box 44.11). Governments often subsidize foods indirectly by sheltering them from sales taxes in the recognition that they are essential; however, this logic should not extend to foods with adverse health effects, such as sugar-sweetened beverages In a follow-up survey after six years, cigarette smoking had decreased from 34 to 27 percent among men, the proportion of adults who exercised regularly had increased from 14 to 17 percent, and the prevalence of obesity was stable. However, hypertension and high LDL cholesterol levels had increased modestly. From 1991 to 1999, the agestandardized incidence of myocardial infarction declined from 98.2 to 83.0 per 100,000 residents (Mak and others 2003) and age-standardized mortality from CAD decreased from 60.8 to 47.2 per 100,000 residents. and those high in trans fats. Legislation can make this distinction, providing a modest economic incentive for healthier choices and at the same time conveying important nutritional messages (see chapter 11). • Use of individual automobiles is often subsidized by building and maintaining highways, providing inexpensive parking, and imposing low taxes on petroleum products that do not fully reflect their societal and environmental costs. Increasing taxes on petroleum products and subsidizing public transportation could have an important effect on choice of transportation modality, which as noted earlier, has major effects on health. • Walking, riding bicycles, and using public transportation can be promoted by economic policies that, in addition Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 843 ©2006 The International Bank for Reconstruction and Development / The World Bank 63 Box 44.11 Poland: A Dramatic Decline in Heart Disease After Poland’s transition to a democratic government in the early 1990s, the government removed large subsidies for butter and lard, and consumption of nonhydrogenated vegetable fat increased rapidly (Zatonski, McMichael, and Powles 1998). The ratio of dietary polyunsaturated to saturated fat increased from 0.33 in 1990 to 0.56 in 1999, and during this period mortality rates from CAD dropped by 28 percent (data provided by W. Zatonski). Changes in smoking and in the consumption of fruits and vegetables probably played a minor role in this decrease (see figure). Risk of Coronary Heart Disease According to Polyunsaturated to Saturated Fat Ratio Rate ratio for coronary heart disease in Nurses’ Health Study 1.2 Poland 1992 1990 1.0 Rate ratio for coronary heart disease mortality in Poland 1.0 1994 0.8 1996 1999 0.6 0.7 0.4 0.2 0.33 0 0.2 0.56 0.3 0.4 0.5 0.6 Dietary polyunsaturated to saturated fat ratio Notes: Squares represent data for Poland from 1990 to 1999. Circles are for deciles of polyunsaturated fat to saturated fat and for risk of coronary heart disease in the Nurses’ Health Study (Hu 1999), which closely predict the observed changes in Poland. to providing better infrastructure, include discounts on transportation fares, provide secure bicycle parking, and reduce health insurance premiums. COST-EFFECTIVENESS OF INTERVENTIONS Only a few studies have described interventions for lifestyle diseases in developing countries. Modeling Likely Interventions Primary targets for reducing lifestyle diseases include changing the fat composition of the diet, limiting sodium intake, and engaging in regular physical activity. Using available data, we calculated a range of estimates under given assumptions for the cost-effectiveness of replacing dietary saturated fat with monounsaturated fat, replacing trans fat with polyunsaturated fat, and reducing salt intake. An increase in moderate physical activity by three to five hours per week is considered likely to lower the risk of many diseases, but data to model the cost-effectiveness of this intervention are not currently available. For further details of methods and assumptions underlying the analyses presented here, see the Web site version of this book. Reducing Saturated Fat Content. In the base case, assuming a 3 percent drop in cholesterol and a US$6 per person cost of the intervention, averting one disability-adjusted life year (DALY) would cost as little as US$1,865 in South Asia and as much as US$4,012 in the Middle East and North Africa. The intervention’s effectiveness could be increased by replacing part of the saturated fat with polyunsaturated fat, which has additional beneficial effects mediated by mechanisms other than LDL cholesterol (see tables 44.2 and 44.3). Replacing Dietary Trans Fat from Partial Hydrogenation with Polyunsaturated Fat. We could not use the model for saturated fat to estimate the effects of replacing trans fat with polyunsaturated fat because only a small part of the benefit is attributable to reducing LDL cholesterol (F. B. Hu and Willett 2002). Trans fats also adversely affect high-density lipoprotein (HDL) cholesterol, triglycerides, endothelial function, and inflammatory markers. In addition, increases in polyunsaturated fat (assuming a mix of N-6 and omega-3 fatty acids) will reduce LDL cholesterol, insulin resistance, and probably fatal cardiac arrhythmias. In calculations that are based only on the adverse effects on LDL and HDL, replacing 2 percent of the energy from trans fat 844 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 64 Table 44.2 Incremental Cost-Effectiveness Ratios, Selected Interventions, by Region (US$/DALY averted) Substituting 2 percent of energy from trans fat with polyunsaturated fat Media campaign to reduce saturated fat content Region East Asia and the Pacific 7 percent CAD reduction Intervention Intervention cost of cost of US$0.50/adulta US$6.00/adult 2,769 73 1,583 40 percent CAD reduction Intervention Intervention cost of cost of US$0.50/adulta US$6.00/adult Cost saving 227 Reducing salt content by means of legislation plus public education 2,056 Europe and Central Asia 2,929 65 1,670 Cost saving 228 2,170 Latin America and the Caribbean 3,297 40 1,865 Cost saving 225 2,476 Middle East and North Africa 4,012 25 2,259 Cost saving 252 3,056 South Asia 1,865 38 1,014 Cost saving 138 1,325 Sub-Saharan Africa 2,356 53 1,344 Cost saving 184 1,766 Source: Authors’ calculations. a. Based on the U.S. Food and Drug Administration’s analysis of the costs of the intervention in the United States. Table 44.3 Two-Way Sensitivity Analysis of the Costs of the Intervention to Reduce Saturated Fat Content and of the Relative Risk Reduction in CAD Events, South Asia (US$/DALY averted) Relative risk reduction in CAD events (percent) US$0.25 10 Cost saving Cost per individual US$3.00 318 US$6.00a 680 5 Cost saving 680 1,403 4b Cost saving 911 1,865 1 258 3,572 7,188 Source: Authors’ calculations. a. Threshold analysis reveals that at the base assumption of US$6 for the intervention, no level in the range of assumed CAD reduction is cost saving. b. Threshold analysis reveals that at a cost below US$0.36 per individual and a 4 percent reduction in CAD (base assumption), the intervention is cost saving. with polyunsaturated fat was estimated to reduce CAD by 7 to 8 percent (Grundy 1992; Willett and Ascherio 1994). Epidemiological studies, which include the contributions of the additional causal pathways, suggest a much greater reduction, from about 25 to 40 percent (F. B. Hu and others 1997; Oomen and others 2001). Another likely benefit is a reduction in the incidence of type 2 diabetes: estimates indicate that the same 2 percent reduction would reduce incidence by 40 percent (Salmeron and others 2001). Because voluntary action by industry (as has nearly been achieved in the Netherlands) or by regulation (as occurred in Denmark) can eliminate partially hydrogenated fat from the diet, this initiative does not require consumer education, and the costs can be extremely low. In an analysis required before implementing food labeling, the U.S. Food and Drug Administration (2003) estimated that trans fat labeling would be highly cost-effective. Even though the effect of labeling itself was estimated to have only a modest effect on consumer behavior, as noted earlier, it is having a major effect on manufacturers’ behavior. The potential for reducing CVD rates by replacing trans fats with polyunsaturated fats will depend on the diets of specific populations. Whereas the intake of trans fat is low in China, it is likely to be high in parts of India, Pakistan, and other Asian countries because of the extraordinarily high content in commonly used cooking fats. Table 44.2 presents the results of a cost-effectiveness analysis assuming the two different estimates for CAD reduction: 7 percent and 40 percent. We used costs of US$0.50 per adult per year, which was the maximal cost in the U.S. Food and Drug Administration analysis, and of US$6.00 per adult per year using traditional health education approaches. The lower estimate— or one even lower—is possible because trans fat can be eliminated at the source rather than depending entirely on changes in individual behavior. With the lower cost, the smaller effect estimate leads to a cost-effectiveness ratio of between US$25 and US$73 per DALY averted, depending on the region, and with the higher-effect estimate, the intervention can be cost saving. Reducing the Salt Content of Manufactured Foods through Legislation and an Accompanying Education Campaign. Table 44.2 shows the base-case cost-effectiveness of a legislated reduction in salt content. The intervention appears to be relatively cost-effective, with a cost per DALY averted of US$1,325 in South Asia to US$3,056 in the Middle East and North Africa. Those regional variations are attributable to differing risk profiles across regions as well as to price differentials for the costs of treating disease sequelae. The actual blood pressure reduction from lower salt consumption could vary from the base-case assumption, as could the costs of the education campaign. Table 44.4 shows the Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 845 ©2006 The International Bank for Reconstruction and Development / The World Bank 65 Table 44.4 Two-Way Sensitivity Analysis of the Costs of the Intervention to Reduce Salt Content and Its Effectiveness, South Asia (US$/DALY averted) Blood pressure reduction (millimeters of mercury) Cost per individual US$1a US$3 US$6 4 9 308 608 3 49 448 847 2b 129 727 1,326 1 368 1,565 2,761 Source: Authors’ calculations. a. Threshold analysis reveals that at a cost of US$1 per individual, a blood pressure reduction would have to be greater than 5 millimeters of mercury for the intervention to be cost saving. At the base-case assumption of a cost of US$6 for the intervention, there is no cost saving threshold level of reduction. b. Threshold analysis reveals that at a cost of less than US$0.47 per individual the intervention is cost-saving. results of lower costs of the education campaign and higher or lower effects of the intervention on blood pressure. These results may argue for initial efforts to focus on reductions in the use of salt during the manufacturing process with no public education campaign. The cost-effectiveness of such a change is high and could be augmented with a public education campaign only if needed to support the legislated change. At lower implementation costs, the intervention is highly cost-effective, even with half the assumed effect on blood pressure. Adopting Physical Activity Interventions. Even though health experts believe that physical activity interventions are effective in reducing the risk of lifestyle diseases, no studies of their cost-effectiveness are available from developing countries. If people walk voluntarily (the model assumes no opportunity cost), a net economic benefit would accrue to all segments of the U.S. population. If we project the economic benefits to the entire U.S. population and assume 25 percent compliance by the sedentary population, the voluntary program would generate US$6.8 billion in savings (in 2001 U.S. dollars). Aggregate Costs of Obesity and Unhealthy Lifestyles A series of U.S. studies appears to confirm that the avoidable costs of chronic diseases are substantial, although many developing countries have not yet experienced the full demands on their health sectors resulting from these conditions. Colditz (1999) estimates that obesity is responsible for 7 percent of all U.S. direct health care costs and that inactivity is responsible for an additional 2.4 percent of all health care costs. Indirect costs associated with obesity and inactivity account for another 5 percent of health care costs. Pronk and others (1999) assess the difference in health care costs between adult patients with and without risk factors for noncommunicable diseases (physical activity, BMI, and smoking status) and find that a healthier lifestyle of physical activity three times per week, a moderate BMI, and nonsmoking status reduce health care costs by 49 percent compared with an unhealthy lifestyle. Cost-Effectiveness of Community-Based Interventions Populationwide and community-based interventions appear to be cost-effective if they reach large populations, address highmortality and high-morbidity diseases, and are multipronged and integrated efforts. The full costs of achieving changes in behavior and policy are often complex and difficult to estimate. Interventions may yield additional spinoff benefits. For instance, decisions to reduce children’s television viewing could easily improve school outcomes as well as reduce childhood obesity. Similarly, increasing walking and bicycle riding for transportation could reduce air pollution. RESEARCH AND DEVELOPMENT PRIORITIES A number of research and development priorities have been identified: • Conduct randomized trials of the use of folic acid and alpha-linoleic acid to prevent CAD in developing countries. These interventions cost little, and the potential benefits are large and rapid. • Develop prospective cohort studies of dietary and lifestyle factors in developing and transition countries to refine the understanding of risk factors in those contexts. To date, almost all such studies have taken place in Europe and North America. • Develop surveillance systems for chronic diseases and for major risk factors, such as obesity, in developing countries. • Develop additional multifaceted, community-based demonstration programs in developing countries to document the feasibility of lifestyle changes and to learn more about effective strategies. • Conduct detailed cost-effectiveness analyses of various prevention strategies to modify dietary and lifestyle factors. RECOMMENDED PRIORITY INTERVENTIONS An overall objective is to develop comprehensive national and local plans that take advantage of every opportunity to encourage and promote healthy eating and active living. These plans would involve health care providers; worksites; schools; media; urban planners; all levels of food production, processing, and preparation; and governments. The goal is cultural change in the direction of healthy living. An important element in 846 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 66 cultural change is national leadership by individuals and by professional organizations. Specific interventions will depend on local physical and cultural conditions and should be based on careful analysis of existing dietary and activity patterns and their determinants; however, the following interventions can be considered (specific interventions for control of smoking are discussed elsewhere): • Physical activity: º Develop transportation policies and a physical environment to promote walking and riding bicycles. This intervention includes constructing sidewalks and protected bicycle paths and lanes that are attractive, safe, welllighted, and functional with regard to destinations. º Adopt policies that promote livable, walker-friendly communities that include parks and are centered around access to public transportation. º Encourage the use of public transportation and discourage overdependence on private automobiles. º Promote the use of stairs. Building codes can require the inclusion of accessible and attractive stairways. • Healthy diets: º Develop comprehensive school programs that integrate nutrition into core curricula and healthy nutrition into school food services. Regional or national standards to promote healthy eating should be developed for school food services. Programs should also aim at limiting television watching, in part by promoting attractive alternatives. º Work with the agriculture sector and food industries to replace unhealthy fats with healthy fats, including adequate amounts of omega-3 fatty acids. This goal can be achieved through a combination of education, regulation, and incentives. Specific actions will depend on local sources of fat and on regional production and distribution. For example, in areas where palm oil is dominant, research could focus on developing strains that are lower in saturated fat and higher in unsaturated fat through selective breeding or genetic alteration. Labeling requirements or regulation can be used to discourage or eliminate the use of partially hydrogenated vegetable oils and to promote the use of nonhydrogenated unsaturated oils instead. º Require clear labeling of energy content for all packaged foods, including fast food. º Use tax policies to encourage the consumption of healthier foods. For example, high-sugar sodas could be fully taxed and not subsidized in the same way as healthier foods. º Emphasize the production and consumption of healthy food products in agriculture support and extension programs. º Implement folic acid fortification if folic acid intake is low. º Ensure that health providers regularly weigh both children and adult patients, track their weights over time, and provide counseling regarding diet and activity if they are already overweight or if unhealthy weight gain is occurring during adulthood. Those activities should be integrated with programs that address undernutrition. Health care providers should be encouraged to set a good example by not smoking, by exercising regularly, and by eating healthy diets. º Promote healthy foods at worksite food services. Worksites can also promote physical activity by providing financial incentives for using public transportation or riding bicycles (and by not subsidizing automobiles by providing free parking). Providing areas for exercise during work breaks and showers may be useful. º Set standards that restrict the promotion of foods high in sugar, refined starch, and saturated and trans fats to children on television and elsewhere. º Set national standards for the amount of sodium in processed foods. • National campaigns: º Invest in developing locally appropriate health messages related to diet, physical activity, and weight control. This effort is best done in cooperation with government agencies, nongovernmental organizations, and professional organizations so that consistent messages can be used on television and radio; at health care settings, schools, and worksites; and elsewhere. This effort should use the best social-marketing techniques available, with messages continuously evaluated for effectiveness. º Develop a sustainable surveillance system that monitors weight and height, physical activity, and key dietary variables. Implementation of the recommended policies to promote health and well-being is often not straightforward because of opposition by powerful and well-funded political and economic forces, such as those involved in the tobacco, automobile, food, and oil industries (Nestle 2002). The solutions will depend on a country’s specific political landscape. However, experiences in many countries indicate that alliances of public interest groups, professional organizations, and motivated individuals can overcome such powerful interests. Strategies should start with sound science and can use a mix of mass media, lobbying efforts, and lawsuits. Also, the food industry is far from monolithic, and elements can often be identified whose interests coincide with health promotion, which can create valuable partnerships. As an example, the willingness of some margarine manufacturers to invest in developing products free of trans fatty acids greatly helped the effort to reduce these fats, because these producers Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 847 ©2006 The International Bank for Reconstruction and Development / The World Bank 67 then became proponents for labeling the trans fat content of foods. Protection of children can be a powerful lever because of almost universal concern about their welfare and the recognition that they cannot be responsible for the long-term consequences of their diet and lifestyle choices. CONCLUSIONS Many of the ongoing diet and lifestyle interventions in lowand middle-income countries are relatively recent, and few have documented reductions in the rates of major chronic diseases. However, the successes of Finland, Singapore, and many other high-income countries in reducing rates of CAD, stroke, and smoking-related cancers strongly suggest that similar benefits will emerge in the developing countries. ACKNOWLEDGMENTS Studied Cohort of U.S. Adults.” New England Journal of Medicine 348: 1625–38. Coitinho, D., C. A. Monteiro, and B. M. Popkin. 2002. “What Brazil Is Doing to Promote Healthy Diets and Active Lifestyles.” Public Health Nutrition 5: 263–67. Colditz, G. A. 1999. “Economic Costs of Obesity and Inactivity.” Medicine Science and Sports Exercise 31: S663–67. Conlin, P. R. 1999. “The Dietary Approaches to Stop Hypertension (Dash) Clinical Trial: Implications for Lifestyle Modifications in the Treatment of Hypertensive Patients.” Cardiology Review 7: 284–88. Cutter, J., B. Y. Tan, and S. K. Chew. 2001. “Levels of Cardiovascular Disease Risk Factors in Singapore Following a National Intervention Programme.” Bulletin of the World Health Organization 79: 908–15. de Lorgeril, M., S. Renaud, N. Mamelle, P. Salen, J. L. Martin, I. Monjaud, and others. 1994. “Mediterranean Alpha-Linolenic Acid–Rich Diet in Secondary Prevention of Coronary artery disease.” Lancet 343: 1454–59. (Erratum in Lancet 1995, 345: 738.) Dickinson, S., S. Colagiuri, E. Faramus, P. Petocz, and J. C. Brand-Miller. 2002. “Postprandial Hyperglycemia and Insulin Sensitivity Differ among Lean Young Adults of Different Ethnicities.” Journal of Nutrition 132: 2574–79. The authors appreciate Hilary Farmer’s assistance in preparing this manuscript. Ebbeling, C. B., M. M. Leidig, K. B. Sinclair, J. P. Hangen, and D. S. Ludwig. 2003. “A Reduced-Glycemic Load Diet in the Treatment of Adolescent Obesity.” Archives of Pediatric and Adolescent Medicine 157: 773–79. REFERENCES Ewing, R, R. Schieber, and C. Zegeer. 2003. “Urban Sprawl as a Risk Factor in Motor Vehicle Occupant and Pedestrian Facilities.” American Journal of Public Health 93: 1541–45. Ascherio, A., E. Cho, K. Walsh, F. M. Sacks, W. C. Willett, and A. Faruqui. 1996. “Premature Coronary Deaths in Asians” (letter). British Medical Journal 312: 508. Fox, K. R. 1999. “The Influence of Physical Activity on Mental Well-Being.” Public Health Nutrition 2: 411–18. Bacon, C. G., M. A. Mittleman, I. Kawachi, E. Giovannucci, D. B. Glasser, and E. B. Rimm. 2003. “Sexual Function in Men Older Than 50 Years of Age: Results from the Health Professionals Follow-up Study.” Annals of Internal Medicine 139: 161–68. Ball, D., S. Ellison, J. Adamy, and G. Fowler. 2004. “Recipes without Borders?” Wall Street Journal, August 18, 2004, p. 1. Baylin, A., E. K. Kabagambe, A. Ascherio, D. Spiegelman, and H. Campos. 2003. “Adipose Tissue Alpha-Linolenic Acid and Nonfatal Acute Myocardial Infarction in Costa Rica.” Circulation 107: 1586–91. Bell, A. C., K. Ge, and B. M. Popkin. 2002. “The Road to Obesity or the Path to Prevention: Motorized Transportation and Obesity in China.” Obesity Research 10: 277–83. Bell, E. A., L. S. Roe, and B. J. Rolls. 2003. “Sensory-Specific Satiety Is Affected More by Volume Than by Energy Content of a Liquid Food.” Physiology and Behavior 78: 593–600. Blendon, R. J. 2002. Welfare of Children in America. Cambridge, MA: Cogent Research. Bly, J. L., R. C. Jones, and J. E. Richardson. 1986. “Impact of Worksite Health Promotion on Health Care Costs and Utilization. Evaluation of Johnson & Johnson’s Live for Life Program.” Journal of the American Medical Association 256: 3235–40. Boutelle, K, R. Jeffery, D. McMurray, and K. Schmitz. 2001. “Using Signs, Artwork, and Music to Promote Stair Use in a Public Building.” American Journal of Public Health 91: 2004–6. Bunn, F., T. Collier, C. Frost, K. Ker, I. Roberts, and R. Wentz. 2003. “Traffic Calming for the Prevention of Road Traffic Injuries: Systematic Review and Meta-Analysis.” Injury Prevention 9: 200–4. Buss, D. 2004. “Is the Food Industry the Problem or the Solution?” New York Times, August 29, 2004, p. 5. Calle, E. E., C. Rodriguez, K. Walker-Thurmond, and M. J. Thun. 2003. “Overweight, Obesity, and Mortality from Cancer in a Prospectively Frumkin, H. 2002. “Urban Sprawl and Public Health.” Public Health Reports 117: 201–17. Giovannucci, E. 2002. “Epidemiologic Studies of Folate and Colorectal Neoplasia: A Review.” Journal of Nutrition 132: 2350–55S. Gortmaker, S. L., K. Peterson, J. Wiecha, A. M. Sobol, S. Dixit, M. K. Fox, and N. Laird. 1999. “Reducing Obesity via a School-Based Interdisciplinary Intervention among Youth: Planet Health.” Archives of Pediatric and Adolescent Medicine 153: 409–18. Grundy, S. M. 1992. “How Much Does Diet Contribute to Premature Coronary Heart Disease?” In Atherosclerosis IX: Proceedings of the Ninth International Symposium on Atherosclerosis, ed. O. Stein, S. Eisenberg, and Y. Stein, 471–78. Tel Aviv: Creative Communications. Handy, S. L., M. G. Boarnet, R. Ewing, and R. E. Killingsworth. 2002. “How the Built Environment Affects Physical Activity: Views from Urban Planning.” American Journal of Preventive Medicine 23: 64–73. Hao, L., J. Ma, M. J. Stampfer, A. Ren, Y. Tian, Y. Tang, and others. 2003. “Geographical, Seasonal, and Gender Differences in Folate Status among Chinese Adults.” Journal of Nutrition 133: 3630–35. Harris, M. I., K. M. Flegal, C. C. Cowie, M. S. Eberhardt, D. E. Goldstein, R. R. Little, and others. 1998. “Prevalence of Diabetes, Impaired Fasting Glucose, and Impaired Glucose Tolerance in U.S. Adults: The Third National Health and Nutrition Examination Survey, 1988–1994.” Diabetes Care 21: 518–24. Hernandez, B., S. L. Gortmaker, G. A. Colditz, K. E. Peterson, N. M. Laird, and S. Parra-Cabrera. 1999. “Association of Obesity with Physical Activity, Television Programs, and Other Forms of Video Viewing among Children in Mexico City.” International Journal of Obesity and Relational Metabolism Disorders 23: 845–54. Honein, M. A., L. J. Paulozzi, T. J. Mathews, J. D. Erickson, and L. Y. Wong. 2001. “Impact of Folic Acid Fortification of the U.S. Food Supply on the Occurrence of Neural Tube Defects.” Journal of the American Medical Association 285: 2981–86. 848 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 68 Hu. 2002. Hu, F. B., M. F. Leitzmann, M. J. Stampfer, G. A. Colditz, W. C. Willett, and E. B. Rimm. 2001. “Physical Activity and Television Watching in Relation to Risk for Type 2 Diabetes Mellitus in Men.” Archives of Internal Medicine 161: 1542–48. Hu, F. B., M. J. Stampfer, J. E. Manson, E. B. Rimm, A. Wolk, G. A. Colditz, and others. 1999. “Dietary Intake of Alpha-Linolenic Acid and Risk of Fatal Ischemic Heart Disease among Women.” American Journal of Clinical Nutrition 69: 890–97. Ludwig, D. S., K. E. Peterson, and S. L. Gortmaker. 2001. “Relation between Consumption of Sugar-Sweetened Drinks and Childhood Obesity: A Prospective, Observational Analysis.” Lancet 357: 505–8. Mak, K. H., K. S. Chia, J. D. Kark, T. Chua, C. Tan, B. H. Foong, and others. 2003. “Ethnic Differences in Acute Myocardial Infarction in Singapore.” European Heart Journal 24: 151–60. Matsudo, V., S. Matsudo, D. Andrade, T. Araujo, E. Andrade, L. Carlos de Oliveira, and G. Braggion. 2002. “Promotion of Physical Activity in a Developing Country: The Agita São Paulo Experience.” Public Health Nutrition 5: 253–61. Hu, F. B., M. J. Stampfer, J. E. Manson, E. Rimm, G. A. Colditz, B. A. Rosner, and others. 1997. “Dietary Fat Intake and the Risk of Coronary Heart Disease in Women.” New England Journal of Medicine 337: 1491–99. Nestle, M. 2002. Food Politics: How the Food Industry Influences Nutrition and Health. Berkeley. CA: University of California Press. Hu, F. B., R. M. van Dam, and S. Liu. 2001. “Diet and Risk of Type 2 Diabetes: The Role of Types of Fat and Carbohydrate.” Diabetologia 44: 805–17. Nissinen, A., X. Berrios, and P. Puska. 2001. “Community-Based Noncommunicable Disease Interventions: Lessons from Developed Countries for Developing Ones.” Bulletin of the World Health Organization 79: 963–70. Hu, F. B., B. Wang, C. Chen, Y. Jin, J. Yang, M. J. Stampfer, and X. Xu. 2000. “Body Mass Index and Cardiovascular Risk Factors in a Rural Chinese Population.” American Journal of Epidemiology 151: 88–97. Hu, F. B., and W. C. Willett. 2002. “Optimal Diets for Prevention of Coronary Heart Disease.” Journal of the American Medical Association 288: 2569–78. Hu, G., H. Pekkarinen, O. Hanninen, Z. J. Yu, H. G. Tian, Z. Y. Guo, and A. Nissinen. 2002. “Physical Activity during Leisure and Commuting in Tianjin, China.” Bulletin of the World Health Organization 80: 933–38. IARC (International Agency for Research on Cancer). 2002. Weight Control and Physical Activity. Lyon, France: IARC Press. Institute of Medicine. 2002. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients): A Report of the Panel on Macronutrients, Subcommittees on Upper Reference Levels of Nutrients and Interpretation and Uses of Dietary Reference Intakes, and the Standing Committee on the Scientific Evaluation of Dietary Reference Intakes. Washington, DC: National Academy of Sciences. http://www.nap.edu/catalog/ 10490.html. Jones, T. F., and C. B. Eaton. 1994. “Cost-Benefit Analysis of Walking to Prevent Coronary Heart Disease.” Archives of Family Medicine 3: 703–10. Keeler, E. B., W. G. Manning, J. P. Newhouse, E. M. Sloss, and J. Wasserman. 1989. “The External Costs of a Sedentary Life-Style.” American Journal of Public Health 79: 975–81. Knowler, W. C., E. Barrett-Connor, S. E. Fowler, R. F. Hamman, J. M. Lachin, E. A. Walker, and D. M. Nathan. 2002. “Reduction in the Incidence of Type 2 Diabetes with Lifestyle Intervention or Metformin.” New England Journal of Medicine 346 (6): 393–403. Koplan, J. P., and W. H. Dietz. 1999. “Caloric Imbalance and Public Health Policy.” Journal of the American Medical Association 282: 1579–81. Lambert, E. V., I. Bohlmann, and T. Kolbe-Alexander. 2001. “‘Be Active’: Physical Activity for Health in South Africa.” South African Journal of Clinical Nutrition 14: S12–16. Law, M. R., C. D. Frost, and N. J. Wald. 1991. “By How Much Does Dietary Salt Reduction Lower Blood Pressure? III—Analysis of Data from Trials of Salt Reduction.” British Medical Journal 302 (6780): 819–24. Leitzmann, M. F., E. B. Rimm, W. C. Willett, D. Spiegelman, F. Grodstein, M. J. Stampfer, and others. 1999. “Recreational Physical Activity and the Risk of Cholecystectomy in Women.” New England Journal of Medicine 341: 777–84. Leyden, K. 2003. “Social Capital and the Built Environment: The Importance of Walkable Neighborhoods.” American Journal of Public Health 93: 1546–51. Lopez, R. 2004. “Urban Sprawl and Risk for Being Overweight or Obese.” American Journal of Public Health 94: 1574–79. Office of the Administrator. 2001. What Is Smart Growth? Washington, DC: Environmental Protection Agency. http://www.epa.gov/ smartgrowth/pdf/whtissg4v2.pdf. Oomen, C. M., M. C. Ocke, E. J. Feskens, M. A. van Erp-Baart, F. J. Kok, and D. Kromhout. 2001. “Association between Trans Fatty Acid Intake and 10-Year Risk of Coronary Heart Disease in the Zutphen Elderly Study: A Prospective Population-Based Study.” Lancet 357: 746–51. Pendakur, V. S. 2000. World Bank Urban Transport Strategy Review. Yokohama, Japan: Pacific Policy and Planning Associates. Petrella, R., J. Koval, D. Cunningham, and D. Paterson. 2003. “Can Primary Care Doctors Prescribe Exercise to Improve Fitness: The Step Test Exercise Prescription (STEP) Project.” American Journal of Preventive Medicine 24: 316–22. Pignone, M. P., A. Ammerman, L. Fernandez, C. T. Orleans, N. Pender, S. Woolf, and others. 2003. “Counseling to Promote a Healthy Diet in Adults: A Summary of the Evidence for the U.S. Preventive Services Task Force.” American Journal of Preventive Medicine 24: 75–92. Pratt, M., C. A. Macera, and G. Wang. 2000. “Higher Direct Medical Costs Associated with Physical Inactivity.” Physician and Sports Medicine 28:63–70. http://www.physsportsmed.com/issues/2000/10_00/pratt. htm. Pronk, N. P., M. J. Goodman, P. J. O’Connor, and B. C. Martinson. 1999. “Relationship between Modifiable Health Risks and Short-Term Health Care Charges.” Journal of the American Medical Association 282: 2235–39. Pucher, J., and L. Dijkstra. 2003. “Promoting Safe Walking and Cycling to Improve Public Health: Lessons from the Netherlands and Germany.” American Journal of Public Health 93: 1509–16. Puska, P., E. Vartiainen, J. Tuomilehto, V. Salomaa, and A. Nissinen. 1998. “Changes in Premature Deaths in Finland: Successful Long-Term Prevention of Cardiovascular Diseases.” Bulletin of the World Health Organization 76: 419–25. Raben, A., T. H. Vasilaras, A. C. Moller, and A. Astrup. 2002. “Sucrose Compared with Artificial Sweeteners: Different Effects on Ad Libitum Food Intake and Body Weight after 10 Weeks of Supplementation in Overweight Subjects.” American Journal of Clinical Nutrition 76: 721–29. Robinson, T. N. 1999. “Reducing Children’s Television Viewing to Prevent Obesity: A Randomized Controlled Trial.” Journal of the American Medical Association 282: 1561–67. Rossouw, J. E., P. L. Jooste, D. O. Chalton, E. R. Jordaan, M. L. Langenhoven, P. C. Jordaan, and others. 1993. “Community-Based Intervention: The Coronary Risk Factor Study (Coris).” International Journal of Epidemiology 22: 428–38. Ruangdaraganon, N., N. Kotchabhakdi, U. Udomsubpayakul, C. Kunanusont, and P. Suriyawongpaisal. 2002. “The Association between Television Viewing and Childhood Obesity: A National Survey in Prevention of Chronic Disease by Means of Diet and Lifestyle Changes | 849 ©2006 The International Bank for Reconstruction and Development / The World Bank 69 Thailand.” Journal of the Medical Association of Thailand 85 (Suppl. 4): S1075–80. Where Do They Feature in the Health Research Agenda?” Bulletin of the World Health Organization 79: 947–53. Saelens, B. E., J. F. Sallis, and L. D. Frank. 2003. “Environmental Correlates of Walking and Cycling: Findings from the Transportation, Urban Design, and Planning Literatures.” Annals of Behavioral Medicine 25: 80–91. U.S. Food and Drug Administration, Center for Food and Safety and Applied Nutrition. 2003. “Food Labeling: Trans Fatty Acids in Nutrition.” Federal Register 68, no. 133, 41433–506 (July 11, 2003). http://www.cfsan.fda.gov/~lrd/fr03711a.html (see also http://vm. cfsan.fda.gov/~lrd/fr991117.html). Salmeron, J., F. B. Hu, J. E. Manson, M. J. Stampfer, G. A. Colditz, E. B. Rimm, and W. C. Willett. 2001. “Dietary Fat Intake and Risk of Type 2 Diabetes in Women.” American Journal of Clinical Nutrition 73: 1019–26. Schulze, M. B., J. E. Manson, D. S. Ludwig, G. A. Colditz, M. J. Stampfer, W. C. Willett, and F. B. Hu. 2004. “Sugar-Sweetened Beverages, Weight Gain, and Incidence of Type 2 Diabetes in Young and Middle-Aged Women.” Journal of the American Medical Association 292: 927–34. Selmer, R., I. Kristiansen, A. Haglerod, S. Graff-Iverson, H. Larsen, H. Meyer, and others. 2000. “Cost and Health Consequences of Reducing the Population Intake of Salt.” Journal of Epidemiology and Community Health 54: 697–702. Sobngwi, E., J. F. Gautier, and J. C. Mbanya. 2003. “Exercise and the Prevention of Cardiovascular Events in Women” (author reply). New England Journal of Medicine 348: 77–79. Staunton, C., D. Hubsmith, and W. Kallins. 2003. “Promoting Safe Walking and Biking to School: The Marin County Success Story.” American Journal of Public Health 93: 1431–34. Surface Transportation Policy Project. 1999. “Aggressive Driving: Where You Live Matters.” Washington, DC. http://www.transact.org/report. asp?id=56. Uusitalo, U., E. J. Feskens, J. Tuomilehto, G. Dowse, U. Haw, D. Fareed, and others. 1996. “Fall in Total Cholesterol Concentration over Five Years in Association with Changes in Fatty Acid Composition of Cooking Oil in Mauritius: Cross-Sectional Survey.” British Medical Journal 313: 1044–46. Wang, G., C. Macera, B. Scudder-Soucie, T. Schmid, M. Pratt, and D. Buchner. 2004. “Cost Effectiveness of a Bicycle/Pedestrian Trail Development in Health Promotion.” Preventive Medicine 38: 237–42. Wang, S. Y., G. B. Chi, C. X. Jing, X. M. Dong, C. P. Wu, and L. P. Li. 2003. “Trends in Road Traffic Crashes and Associated Injury and Fatality in the People’s Republic of China, 1951–1999.” Injury Control and Safety Promotion 10: 83–87. WHO (World Health Organization). 2000. Obesity: Preventing and Managing the Global Epidemic. WHO Technical Report 894. Geneva: WHO. ———. 2002. The World Health Organization Report 2002: Reducing Risks, Promoting Healthy Life. Geneva: WHO. WHO and FAO (World Health Organization and Food and Agriculture Organization of the United Nations). 2003. Diet, Nutrition, and the Prevention of Chronic Diseases: Report of a Joint WHO/FAO Expert Consultation. Report 916. Geneva: WHO. Swinburn, B. A., I. Caterson, J. C. Seidell, and W. P. James. 2004. “Diet, Nutrition, and the Prevention of Excess Weight Gain and Obesity.” Public Health Nutrition 7: 123–46. Willett, W. C. 2002. “Balancing Lifestyle and Genomics Research for Disease Prevention.” Science 296: 695–98. Toh, C. M., J. Cutter, and S. K. Chew. 2002. “School-Based Intervention Has Reduced Obesity in Singapore.” British Medical Journal 324: 427. Willett, W. C., and A. Ascherio. 1994. “Trans-Fatty Acids: Are the Effects Only Marginal?” American Journal of Public Health 84: 722–24. Tosteson, A., M. Weinstein, M. Hunink, M. A. Mittleman, L. Williams, P. Goldman, and L. Goldman. 1997. “Cost-Effectiveness of PopulationWide Educational Approaches to Reduce Serum Cholesterol Levels.” Circulation 95: 24–30. Willett, W. C., W. H. Dietz, and G. A. Colditz. 1999.“Guidelines for Healthy Weight.” New England Journal of Medicine 341: 427–34. Tudor-Locke, C., B. E. Ainsworth, L. S. Adair, S. Du, and B. M. Popkin. 2003. “Physical Activity and Inactivity in Chinese School-Aged Youth: The China Health and Nutrition Survey.” International Journal of Obesity 27: 1093–99. Tuomilehto, J., J. Lindstrom, J. G. Eriksson, T. T. Valle, H. Hamalainen, P. Ilanne-Parikka, and others. 2001. “Prevention of Type 2 Diabetes Mellitus by Changes in Lifestyle among Subjects with Impaired Glucose Tolerance.” New England Journal of Medicine 344: 1343–50. Willett, W. C., and R. L. Leibel. 2002. “Dietary Fat Is Not a Major Determinant of Body Fat.” American Journal of Medicine 113 (Suppl. 9B): 47–59S. World Bank. 2003. Noncommunicable Diseases in Pacific Island Countries: Disease Burden, Economic Cost, and Policy Options. Nouméa, New Caledonia: World Bank. Zatonski, W. A., A. J. McMichael, and J. W. Powles. 1998. “Ecological Study of Reasons for Sharp Decline in Mortality from Ischaemic Heart Disease in Poland since 1991.” British Medical Journal 316: 1047–51. Unwin, N., P. Setel, S. Rashid, F. Mugusi, J. C. Mbanya, H. Kitange, and others. 2001. “Noncommunicable Diseases in Sub-Saharan Africa: 850 | Disease Control Priorities in Developing Countries | Walter C. Willett, Jeffrey P. Koplan, Rachel Nugent, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 70 Chapter 46 Tobacco Addiction Prabhat Jha, Frank J. Chaloupka, James Moore, Vendhan Gajalakshmi, Prakash C. Gupta, Richard Peck, Samira Asma, and Witold Zatonski Cigarette smoking and other forms of tobacco use impose a large and growing global public health burden. Worldwide, tobacco use is estimated to kill about 5 million people annually, accounting for 1 in every 5 male deaths and 1 in 20 female deaths of those over age 30. On current smoking patterns, annual tobacco deaths will rise to 10 million by 2030. The 21st century is likely to see 1 billion tobacco deaths, most of them in low-income countries. In contrast, the 20th century saw 100 million tobacco deaths, most of them in Western countries and the former socialist economies. Hundreds of millions of premature tobacco deaths could be avoided if effective interventions were widely applied in low- and middle-income countries. Numerous studies from high-income countries and a growing number from low- and middle-income countries provide robust evidence that tobacco tax increases, timely dissemination of information about the health risks of smoking, restrictions on smoking in public and workplaces, comprehensive bans on advertising and promotion, and increased access to cessation therapies are effective in reducing tobacco use and its consequences. Cessation by the 1.1 billion current smokers is central to meaningful reductions in tobacco deaths over the next five decades. New analyses presented here find that higher tobacco taxes could prevent 3 million tobacco deaths by 2030 among smokers alive today. Reduced uptake of smoking by children would yield benefits chiefly after 2050. Price and non-price interventions are, for the most part, highly cost-effective. This chapter begins with an overview of smoking trends and tobacco’s health consequences, followed by a discussion of the economic rationale for government intervention, with a focus on the uniquely addictive properties of nicotine. A review of the effectiveness of tobacco-control policies in reducing tobacco initiation and in increasing cessation follows. A cost-effectiveness analysis of these interventions is provided. Finally, the constraints to implementing tobacco-control policies are discussed. SMOKING TRENDS Tobacco use, in both smoked and nonsmoked forms, is common worldwide. This chapter focuses on smoked tobacco, chiefly cigarettes and bidis (tobacco hand rolled in the leaf of another plant, temburi, which is popular in India and parts of Southeast Asia), because smoked tobacco is more common— accounting for about 65 to 85 percent of all tobacco produced worldwide (WHO 1997)—and causes more disease and more diverse types of disease than does oral tobacco use. Prevalence A systematic review of 139 studies on adult smoking prevalence (Jha and others 2002) found that more than 1.1 billion people worldwide smoke, with about 82 percent of smokers residing in low- and middle-income countries. Table 46.1 provides an update of these estimates for the population in 2000. Globally, male smoking far exceeds female smoking, with a smaller gender difference in high-income countries. Smoking prevalence is highest in Europe and Central Asia, where 35 percent of all adults are smokers. While overall smoking prevalence continues to increase in many low- and middle-income countries, many high-income countries have witnessed decreases, most clearly in men. A 869 ©2006 The International Bank for Reconstruction and Development / The World Bank 71 Table 46.1 Estimated Smoking Prevalence (by Gender) and Number of Smokers, 15 Years of Age and Older, 2000 Smoking prevalence (percent) World Bank region Males Total smokers Females Overall Millions Percentage of all smokers East Asia and the Pacific 63 5 34 429 38 Europe and Central Asia 56 17 35 122 11 Latin America and the Caribbean 40 24 32 98 9 Middle East and North Africa 36 5 21 37 3 South Asia 32 6 20 178 15 Sub-Saharan Africa 29 8 18 56 6 Low- and middle-income economies 49 8 29 920 82 High-income economies 37 21 29 202 18 Source: Authors. study in 36 mostly Western countries, from early 1980 to the mid 1990s, suggested that the decrease in smoking prevalence observed among men was caused by the higher prevalence in younger age groups of those who have never smoked. Among women, there was little overall change in smoking prevalence because the increasing prevalence of smokers in younger cohorts counterbalanced increasing cessation in older age groups (Molarius and others 2001). Cessation Ex-smoking rates are a good measure of cessation at a population level. In some high-income countries, the prevalence rates of ex-smokers have increased over the past two to three decades. For example, in the United Kingdom, smoking prevalence among males over age 30 fell from 70 percent in the 1950s to 30 percent in 2000; female smoking prevalence fell from 40 to 20 percent over the same period. Much of the decrease arose from cessation. Today, two times as many ex-smokers as smokers exist among those age 50 or over. Currently, 30 percent of the U.K. male population is made up of former smokers (Peto and others 2000). Polish male cessation rates have also increased, partly because of control programs. One of every four adult Polish males described himself as an ex-smoker (Zatonski and Jha 2000). In contrast, the prevalence of male exsmokers in most developing countries is low: 10 percent in Vietnam, 5 percent in India, and 2 percent in China (Jha and others 2002). Even those low figures may be falsely elevated because they include people who quit because either they were too ill to continue or they had early symptoms of tobaccorelated illness (Martinson and others 2003). health and population policy. Thus, the salient aspects of tobacco epidemiology are outlined in this section. Key Messages for the Individual Smoker More than 50 years of epidemiology on smoking-related diseases have led to three key messages for individual smokers worldwide (Doll and others 2004; Peto and others 2003). • The eventual risk of death from smoking is high, with about one-half to two-thirds of long-term smokers eventually being killed by their addiction. • These deaths involve a substantial number of life years forgone. About half of all tobacco deaths occur at ages 35 to 69, resulting in the loss of about 20 to 25 years of life, compared with the life expectancy of nonsmokers. • Cessation works: those adults who quit before middle age avoid almost all the excess hazards of continued smoking. Worldwide, about 80 percent of deaths among the 2.7 billion adults over age 30 involve vascular, respiratory, or neoplastic disease. Smoking is associated with an increase in the frequency of many of these diseases, although important differences exist between and across populations. The following discussion focuses on the consequences of smoking on adult mortality. Detailed epidemiological reviews of worldwide mortality from smoking are found elsewhere (C. Gajalakshmi and others 2000; V. Gajalakshmi and others 2003; Gupta and Mehta 2000; Liu and others 1998; Niu and others 1998; Peto and others 1994). Current Mortality and Disability from Smoking HEALTH CONSEQUENCES OF SMOKING The health consequences of smoking are often assumed to be widely understood. In fact, ignorance of the magnitude of tobacco hazards is widespread in terms of both individual Recent updates of indirect estimates of global tobacco mortality (Ezzati and Lopez 2003; M. Ezzati, personal communication, November 2004) indicate that in 2000, 5.0 million premature deaths were caused by tobacco. About half (2.6 million) of those deaths were in low-income countries. Males accounted 870 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 72 Table 46.2 Tobacco Mortality and Total DALYs by Gender, 2000 (thousands) Tobacco deaths World Bank region Total DALYs Males Females Males Females 829 274 13,116 4,128 Europe and Central Asia 754 161 12,407 2,686 Latin America and the Caribbean 177 97 2,789 1,613 East Asia and the Pacific Middle East and North Africa South Asia Sub-Saharan Africa Low- and middle-income economies 28 1,676 554 187 12,397 3,285 105 66 1,659 1,091 2,730 813 44,044 13,357 929 548 12,304 6,866 3,659 1,361 56,347 20,222 High-income economies World 97 768 Source: Ezzati and Lopez 2003; Mathers and others 2006. Note: The terms high-income and former socialist economies as used in the text correspond roughly to high-income and Europe and Central Asia regions using the World Bank classification. Low-income countries corresponds roughly to East Asia and the Pacific, Latin America and the Caribbean, Middle East and North Africa, South Asia, and Sub-Saharan Africa. for 3.7 million deaths, or 72 percent of all tobacco deaths. About 60 percent of male and 40 percent of female tobacco deaths were of middle-aged persons (ages 35 to 69). In high-income countries and former socialist economies, the 1 million middle-aged male tobacco deaths were largely composed of cardiovascular disease (0.45 million) and lung cancer (0.21 million). In contrast, in low-income countries, the leading causes of death among the 1.3 million male tobacco deaths were cardiovascular disease (0.4 million), chronic obstructive pulmonary disease (0.2 million), other respiratory disease (chiefly tuberculosis, 0.2 million), and lung cancer (0.18 million). The specific numbers of deaths from tobacco and of total disability-adjusted life years (DALYs) by gender and World Bank region are shown in table 46.2. Disability estimates are not discussed here; however, disability is highly correlated with mortality in most settings. Past and Future Trends in Mortality In high-income and former socialist economies with more complete and reliable mortality statistics, one can measure the effects of increased smoking prevalence and subsequent decreases that have been observed among large numbers of adults. These changes are best documented by examining lung cancer mortality rates among young adults because lung cancer is not often misclassified with other causes of death at young ages and it is almost entirely attributable to smoking. Age-Standardized Lung Cancer Mortality Rates Age-standardized male lung cancer rates at ages 35 to 44 per 100,000 men in the United Kingdom had fallen from 18 in 1950 to 4 by 2000. In contrast, comparable French male lung cancer rates show the reverse pattern (Peto and others 2003; figure 46.1). In France,the increase in smoking occurred some decades later than in the United Kingdom, and declines in smoking began only after 1990. Similarly, a large increase in female lung cancer at young ages was avoided in the United Kingdom, but female lung cancer at young ages continues to rise in France. Future increases in tobacco deaths worldwide are expected to arise from increased smoking by males in developing countries and by women worldwide. Such increases are a product of population growth and increased age-specific tobacco mortality rates, the latter relating to both smoking duration and the amount of tobacco smoked. Peto and others (1994) have made the following calculation: if the proportion of young people taking up smoking continues to be about half of men and onetenth of young women, there will be about 30 million new long-term smokers each year. As previously noted, epidemiological studies in developed and developing countries suggest that half of these smokers will eventually die from smoking. However, if we conservatively assume that “only” about onethird of smokers die as a result of smoking, then smoking will eventually kill about 10 million people a year. Thus, for the 25-year period from 2000 to 2025, there would be about 150 million tobacco deaths, or about 6 million deaths per year on average; from 2025 to 2050, there would be about 300 million tobacco deaths, or about 12 million deaths per year. Further estimations are more uncertain, but current smoking trends and projected population growth indicate that from 2050 to 2100 there will be an additional 500 million tobacco deaths. These projections for the next three to four decades are comparable to retrospective and early prospective epidemiological studies in China (Liu and others 1998; Niu and others 1998), which suggest that annual tobacco deaths will rise to 1 million before 2010 and to 2 million by 2025, when the young adult smokers of today reach old age. Similarly, results from a large retrospective study in India suggest that 1 million annual deaths can be expected from male smokers by 2025 (V. Gajalakshmi and others 2003). With other populations in Asia, Eastern Tobacco Addiction | 871 ©2006 The International Bank for Reconstruction and Development / The World Bank 73 a. United Kingdom b. France Death rate/100,000 men, age standardized a Death rate/100,000 men, age standardized a 18 18 16 16 Males 14 14 12 12 10 10 8 8 6 6 4 4 Females 2 0 1950 Males Females 2 1960 1970 1980 1990 2000 0 1950 1960 1970 1980 1990 2000 Source: Peto and others 2003. a. Mean of annual rates in component five-year age groups (35–39, 40–44). Figure 46.1 Changes in Lung Cancer Mortality at Age 35 to 44 in the United Kingdom and France, 1950–99 Europe, Latin America, the Middle East, and (less certainly) Sub-Saharan Africa showing similar growth in population and age-specific tobacco death rates, the estimate of some 450 million tobacco deaths over the next five decades appears plausible. Almost all of these deaths will be among current smokers. Tobacco deaths (million) 520 500 500 400 340 300 Benefits of Cessation Current tobacco mortality statistics reflect past smoking behavior, given the long delay between the onset of smoking and the development of disease. The prevention of a substantial proportion of these tobacco deaths before 2050 requires adult cessation. For example, halving the per capita adult consumption of tobacco by 2020 (akin to the declines in adult smoking in the United Kingdom) would avert about 180 million tobacco deaths. Continuing to reduce the percentage of children who start to smoke will prevent many deaths, but its main effect will be on mortality rates in 2050 and beyond (figure 46.2; Jha and Chaloupka 2000a; Peto and Lopez 2001). Substantial evidence indicates that smoking cessation reduces the risk of death from tobacco-related diseases. Among doctors in the United Kingdom, those who quit smoking before the onset of major disease avoided most of the excess hazards of smoking (Doll and others 2004). The benefits of quitting were largest in those who quit before middle age (between ages 25 and 34 years) but were still significant in those who quit later (between ages 45 and 54 years). Cessation before middle age avoids more than 90 percent of the lung cancer risk attributable to tobacco, with quitters possessing a pattern of survival similar to that of persons who 220 200 190 100 70 0 1950 2000 Baseline If adult consumption halves by 2020 2025 2050 If proportion of young adults taking up smoking halves by 2020 Source: Jha and Chaloupka 2000a; Peto and Lopez 2001. Note: Peto and others (1994) estimate 60 million tobacco deaths between 1950 and 2000 in industrial countries. This figure estimates an additional 10 million tobacco deaths between 1990 and 2000 in developing countries. The figure also assumes no tobacco deaths before 1990 in developing countries and minimal tobacco deaths worldwide before 1950. Projections for deaths from 2000 to 2050 are based on Peto and Lopez (2001). Figure 46.2 Tobacco Deaths in the Next 50 Years under Current Smoking Patterns have never smoked. In the United Kingdom, among those who stopped smoking, the risk of lung cancer fell steeply with time since cessation. For men who stopped at ages 60, 50, 40, and 30, the cumulative risks of lung cancer by age 75 were 872 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 74 Lung cancer mortality (percent) Continued smoking 15 10 • The addictive nature of tobacco is underappreciated and poorly understood. Although general awareness of risks is better in high-income countries, many people still underestimate tobacco’s danger relative to other health risks, and many smokers fail to fully internalize these risks (Weinstein 1998). • Smokers may impose costs on others from passive tobacco smoke or, more controversially, from higher health care costs (Lightwood and others 2000; Warner 2003). Stopped age 50 5 Stopped age 30 Never smoked 0 45 55 65 75 Age Source: Peto and others 2000. Figure 46.3 Stopping Works: Cumulative Risk of Lung Cancer Mortality in U.K. Males, 1990 rates The reader is referred to more detailed discussions on the welfare economics of tobacco (Barnum 1994; Jha and others 2000; Peck and others 2000; Warner and others 1995; and several background papers in the Disease Control Priorities Project Working Paper Series). We discuss nicotine addiction because this newer evidence has profound implications for explaining smoking behavior and for devising control policies. Nicotine Addiction In addition to the public health burden caused by tobacco, an economic rationale exists for government to intervene to reduce tobacco use: Before the landmark 1988 U.S. Surgeon General’s report, which suggested that cigarettes and other forms of tobacco are addictive and that nicotine is the major agent in tobacco responsible for addiction, the prevailing view was that tobacco use was largely a voluntary behavior or personal choice (Koop 2003). Since that time, clinicians, behavioral scientists, researchers, and public health experts have increasingly recognized manufactured tobacco products as some of the most addictive and deadly dependence-producing substances available. Although numerous factors have been identified that can contribute to the reinforcement of the smoking habit—for example, the synergistic and independent effects of other compounds in tobacco smoke (such as tar and acetaldehyde) or the sensory and environmental stimuli associated with smoking (such as tobacco advertising)—little debate exists that nicotine is a significant contributor to the development and maintenance of the smoking habit (Markou and Henningfield 2003). In most aspects of dependence, nicotine is on par with other powerfully addictive drugs, such as heroin and cocaine. Newer evidence has converged on the following key points. • Consumers have inadequate information about the health consequences of tobacco use (Jha and others 2000; Warner and others 1995). Specifically, the decision to initiate smoking is made primarily by youths, whose ability to make fully informed, appropriately forward-looking decisions is questioned by society in many different contexts (minimum ages for drinking, driving, and voting, for instance). In industrial countries, about 80 percent of adult smokers begin smoking before age 20. Even if children and young adults have information on future risks, they tend to discount that future risk greatly. Biological Aspects. Nicotine is a psychoactive drug that triggers a cascade of neurobiological events in the reward areas of the brain and throughout the body that can, in turn, act in concert to reinforce tobacco use (Markou and Henningfield 2003). Even a short-term exposure to nicotine has been shown to induce long-lasting changes of the excitatory input into the brain’s reward system, which may be an important early step in the path to addiction (Laviolette and van der Kooy 2004). Notably, in some experimental models, if nicotine’s neurobiological effects are blocked pharmacologically, or if nicotine is removed from cigarette smoke, smoking eventually ceases 10 percent, 6 percent, 3 percent, and 2 percent, respectively (Peto and others 2000; figure 46.3). These results have been supported by a recent multicenter study of men in four European countries; for men who quit smoking at age 40, the study found that the excess lung cancer risk avoided was 85 percent, 91 percent, and 80 percent in the United Kingdom, Germany, and Italy, respectively (Crispo and others 2004). Smoking cessation is uncommon in most developing countries, but some evidence exists that, among Chinese men, quitting also reduces the risks of total and vascular mortality (Lam and others 2002). RATIONALE FOR GOVERNMENT INTERVENTION Tobacco Addiction | 873 ©2006 The International Bank for Reconstruction and Development / The World Bank 75 (Jarvis 2004). The overwhelming property of nicotine that leads to its frequent use is the occurrence of nicotine withdrawal, for which cigarette smoke provides rapid relief. Though each individual differs greatly in his or her sensitivity to nicotine dependence, evidence suggests that most adults are susceptible to the biological effects of nicotine and tobacco (Picciotto 2003). Psychological Aspects. In addition to the unique neurobiology of nicotine, the ready availability of tobacco influences the uptake of smoking as well as the development and maintenance of dependence. With illicit drugs, legal and social barriers constantly test a user’s drive to consume the drug. In contrast, a smoker is presented with nearly ubiquitous opportunities and frequent cues to both purchase and use tobacco because of mass marketing and promotion of tobacco (Shiffman and West 2003). Young people, who are attracted to many risk behaviors, such as fast driving or binge drinking, do not “learn” from early smoking in the way that most young people become safer drivers and moderate drinkers as adults (Jha and others 2000; O’Malley, Bachman, and Johnston 1988). Economics. The traditional economic formulation of costs and benefits tends not to take into account the unique properties of addiction (see Chaloupka, Tauras, and Grossman 2000 for a review). Newer models have begun to incorporate factors such as lack of information, regret, and addiction itself. One key innovation by Gruber and Koszegi (2001, 2002) permits smokers to be time inconsistent, meaning that, given preferences, smokers would make different decisions at different points in time. This approach, now widely used within the new field of behavioral economics, admits conflict between what smokers would like for themselves today and what they would like for themselves in the future. Implications for Control Programs. These newer economic models have several implications for control programs. First is the need for much more aggressive tobacco taxation to deter the development of tobacco smoking. Estimates suggest that, in the United States, optimal taxation taking into account smoking initiation and nicotine addiction would be at least US$1 higher per pack (Gruber 2003; Gruber and Koszegi 2002; Gruber and Mullainathan 2002). The second implication is that the usual assumption that higher taxes reduce the welfare or satisfaction of continuing smokers may not be true. Higher taxes enhance welfare by acting like an external control device over the time-inconsistent preferences of smokers, which would reduce the likelihood of smoking initiation. The third implication is that the overall economic benefits of tobacco control, taking into account addiction, are likely to be substantially positive. Earlier cost-benefit analyses have shown that if even modest costs are assigned to uninformed smoking choices, the net economic costs of tobacco are profoundly negative (Barnum 1994; Peck and others 2000). While some of the methods for such costing have been disputed, newer economic evidence supports the idea that widespread hazards of tobacco use lead to major economic costs. Jamison, Lau, and Wang (2005) have outlined that male survival explains income growth independent of changes in physical capital, education, fertility, economic openness, and technical progress. Thus, if adult male survival in the former socialist economies of Europe had been that of high-income countries, annual growth rates over the past three decades would have been about 1.4 percent rather than 1 percent, making 1990 per capita income about 12 percent higher, or an absolute value of US$140 billion. The chief determinant of the mortality gap between the former socialist economies and high-income countries from 1960 to 1990 is smoking (Peto and others 1994; Zatonski and Jha 2000). More recent economic studies that have put a value on “statistical life” suggest that smoking cessation generates huge benefits. For example, Murphy and Topel (2003) find that in the United States, the value of reduced mortality from all causes between 1970 and 1998 amounted to US$2.6 trillion per year, or half of gross domestic product (GDP) growth during the period. Fully US$1.1 trillion per year arose from reduced heart disease, of which at least one-third was from reduced smoking and saturated fat in diets (Cutler and Kadiyala 2003; see chapter 15 for a fuller discussion on the economic benefits of disease control). INTERVENTIONS TO REDUCE DEMAND FOR TOBACCO Numerous studies, mostly from high-income countries, have examined the effect of interventions aimed at reducing the demand for tobacco products on smoking and other kinds of tobacco use. The small but growing number of studies from low- and middle-income countries provide useful lessons about differences in the effect of these interventions between these countries and high-income countries. The following is a review of the effect of price and non-price interventions in reducing demand for smoking, including a discussion of each intervention’s effect on initiation and cessation. A more complete study of the effectiveness of various interventions is available elsewhere (Jha and Chaloupka 2000b). Tobacco Taxation Nearly all governments tax tobacco products. However, significant differences exist across countries in levels of tobacco taxation. Some of these taxes are specific or per unit taxes; others are expressed as a percentage of wholesale or retail prices (ad valorem taxes). As illustrated in figure 46.4, taxes tend to be absolutely higher and account for a greater share of the retail price (two-thirds or more) in high-income countries. In 874 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 76 Average price or tax per pack (US$) Percentage of tax share 3.5 80 3.0 70 60 2.5 50 2.0 40 1.5 30 1.0 20 0.5 0 10 High-income countries Upper-middle- Lower-middleincome income countries countries Average price Average tax Low-income countries 0 Percentage of tax share Source: Authors. Figure 46.4 Average Cigarette Price, Tax, and Percentage of Tax Share per Pack, by Income Group, 1996 low- and middle-income countries, taxes are generally much lower and account for less than half of the final price of cigarettes. In the United States, federal and state excise taxes on cigarettes were one-third lower, in real terms, in 1995 than their peak level of the mid 1960s. However, taxes rose sharply over the next eight years and stood at US$1.12 per pack as of 2002. Well over 100 studies from high-income countries clearly demonstrate that increases in taxes on cigarettes and other tobacco products lead to significant reductions in cigarette smoking and other tobacco use (Chaloupka, Hu, and others 2000). These reductions reflect the combination of increased smoking cessation, reduced relapse, lower smoking initiation, and decreased consumption among continuing tobacco users. Studies from Canada, the United Kingdom, the United States, and many other high-income countries generally estimate that the price elasticity of cigarette demand ranges from ⫺0.25 to ⫺0.50, indicating that a 10 percent increase in cigarette prices will reduce overall cigarette smoking by 2.5 to 5.0 percent (Chaloupka, Hu, and others 2000; U.S. DHHS 2000). Estimates from a limited number of studies from lowand middle-income countries suggest a greater price elasticity of ⫺0.8 in such countries. Recent studies using survey data have concluded that half or more of the effect of price on overall cigarette demand results from reducing the number of current smokers (CDC 1994; Wasserman and others 1991). Higher taxes increase both the number of attempts at quitting smoking and the success of those attempts (Tauras 1999; Tauras and Chaloupka 2003). A study in the United States (Taurus 1999) suggested that a 10 percent increase in price would result in 11 to 13 percent shorter smoking duration or a 3.4 percent higher probability of cessation. Many recent studies from the United States have used individual-level data to explore differences in the price elasticity of cigarette demand by age, with a particular emphasis on youth and young adults (Chaloupka, Hu, and others 2000; U.S. DHHS 2000). Given that most smoking behavior becomes firmly established during teenage years and young adulthood, interventions that are effective in preventing smoking initiation and the transition to regular, addicted smoking will have significant long-term public health benefits. Estimates from these recent studies conclude that an inverse relationship exists between price elasticity and age, with estimates for youth price elasticity of demand up to three times those obtained for adults (Gruber 2003; Ross, Chaloupka, and Wakefield 2001). Several recent studies have begun to explore the differential effect of cigarette prices on youth smoking uptake, concluding that higher cigarette prices are particularly effective in preventing young smokers from moving beyond experimentation into regular, addicted smoking (Emery, White, and Pierce 2001; Ross, Chaloupka, and Wakefield 2001). In the United Kingdom and the United States, increases in the price of cigarettes have had the greatest effect on smoking among the lowest-income and least educated populations (CDC 1994; Townsend, Roderick, and Cooper 1998). Furthermore, it was estimated that smokers in U.S. households below median income level are four times more responsive to price increases than smokers in households above median income level. In general, estimates of price elasticity for lowand middle-income countries are about double those estimated for high-income countries, implying that significant increases in tobacco taxes in these countries would be effective in reducing tobacco use. Restrictions on Smoking Over the past three decades, as the quantity and quality of information about the health consequences of exposure to passive smoking have increased, many governments, especially in high-income countries, have enacted legislation restricting smoking in a variety of public places and private worksites. In addition, increased awareness of the consequences of passive smoke exposure, particularly to children, has led many workplaces and households to adopt voluntary restrictions on smoking. Although the intent of those restrictions is to reduce nonsmokers’ exposure to passive tobacco smoke, the policies also reduce smokers’ opportunities to smoke. Additional reductions in smoking, especially among youths, will result from the changes in social norms that are introduced by adopting these policies (U.S. DHHS 1994). In Western populations, comprehensive restrictions on cigarette smoking have been estimated to reduce population Tobacco Addiction | 875 ©2006 The International Bank for Reconstruction and Development / The World Bank 77 smoking rates by 5 to 15 percent (see review by Woolery, Asma, and Sharp 2000) and can also lead to changes in social norms regarding smoking behavior, especially among youths. As with higher taxes, these restrictions reduce both the prevalence of smoking and cigarette consumption among current smokers. Smoking bans in workplaces generally reduce the quantity of cigarettes smoked by 5 to 25 percent and reduce prevalence rates by up to 20 percent (Levy, Friend, and Polishchuk 2001). No-smoking policies were most effective when strong social norms against smoking helped make smoking restrictions selfenforcing. Health Information and Counteradvertising The 1962 report by the British Royal College of Physicians and the 1964 U.S. Surgeon General’s Report were landmark tobaccocontrol events in high-income countries. These publications resulted in the first widespread press coverage of the scientific links between smoking and lung cancer. The reports were followed, in many countries, by policies requiring health warning labels on tobacco products, which were later extended to tobacco advertising. Research from high-income countries indicates that these initial reports and the publicity that followed about the health consequences of smoking led to significant reductions in consumption, with initial declines of between 4 and 9 percent and longer-term cumulative declines of 15 to 30 percent (Kenkel and Chen 2000; Townsend 1993). Efforts to disseminate information about the risks of smoking and of other tobacco use in low- and middle-income countries have led to similar declines in tobacco use in those countries (Kenkel and Chen 2000). In addition, mass media antismoking campaigns, in many cases funded by earmarked tobacco taxes, have generated reductions in cigarette smoking and other tobacco use (Kenkel and Chen 2000; Saffer 2000). Decreases in smoking prevalence were largest in Western countries, where the public is constantly and consistently reminded of the dangers of smoking by extensive coverage of issues related to tobacco in the news media (Molarius and others 2001). In many low- and middle-income countries, a lack of awareness continues to exist about the risks of mortality and disease posed by smoking. For example, a national survey in China in 1996 found that 61 percent of smokers thought that tobacco did them “little or no harm” (Chinese Academy of Preventive Medicine 1997). In high-income countries, smokers are aware of the risks, but a recent review of psychological studies found that few smokers judge the size of these risks to be higher and more established than do nonsmokers, and that smokers minimize the personal relevance of these risks (Weinstein 1998). Bans on Advertising and Promotion Cigarettes are among the most heavily advertised and promoted products in the world. In 2002, cigarette companies spent US$12.5 billion on advertising and promotion in the United States alone, the highest spending level reported to date (U.S. Federal Trade Commission 2004). Tobacco advertising efforts worldwide include traditional forms of advertising on television, radio, and billboards and in magazines and newspapers as well as favorable product placement; price-related promotions, such as coupons and multipack discounts; and sponsorship of highly visible sporting and cultural events. Numerous econometric studies, largely from the United Kingdom and the United States, have explored the relationship between cigarette advertising and promotional expenditure and cigarette demand. In general, these studies have resulted in mixed findings, with most studies concluding that advertising has a small positive effect on demand (Chaloupka, Hu, and others 2000; Townsend 1993). However, critics of these studies note that econometric methods, which estimate the effect of a marginal change in advertising expenditures on smoking, are ill suited for studying the effect of advertising (Chaloupka, Hu, and others 2000; U.S. Federal Trade Commission 2004; Townsend 1993). Approaches used by other disciplines, including survey research and experiments that assess reactions to and recall of cigarette advertising, do support the hypothesis that increases in cigarette advertising and promotion directly and indirectly increase cigarette demand and smoking initiation (U.S. DHHS 1994; U.K. Department of Health 1992). These studies conclude that cigarette advertising is effective in getting and retaining children’s attention, with the strength of the association strongly correlated with current smoking behavior, smoking initiation, and smoking intentions. Comprehensive advertising and promotion bans on cigarettes provide more direct evidence on the effect of advertising these products (Saffer 2000). One study using data from 22 high-income countries for the period 1970 through 1992 provides strong evidence that comprehensive bans on cigarette advertising and promotion led to significant reductions in cigarette smoking. The study predicts that a comprehensive set of tobacco advertising bans in high-income countries could reduce tobacco consumption by more than 6 percent, taking into account price and non-price control interventions (Saffer and Chaloupka 2000). However, the study concludes that partial bans have little effect on smoking behavior, given that the tobacco industry can shift its resources from banned media to other media that are not banned. Smoking Cessation Treatments Near-term reductions in smoking-related mortality depend heavily on smoking cessation. Numerous behavioral smoking cessation treatments are available, including self-help manuals, community-based programs, and minimal or intensive clinical interventions (U.S. DHHS 2000). In clinical settings, pharmacological treatments, including nicotine replacement 876 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 78 therapies (NRT) and bupropion, have become much more widely available in recent years in high-income countries through deregulation of some NRT from prescription to overthe-counter status (Novotny and others 2000; U.S. DHHS 2000). The evidence is strong and consistent that pharmacological treatments significantly improve the likelihood of quitting, with success rates two to three times those when pharmacological treatments are not used (Novotny and others 2000; Raw, McNeill, and West 1999; U.S. DHHS 2000). The effectiveness of all commercially available NRT seems to be largely independent of the duration of therapy, the setting in which the therapy is provided, regulatory status (over-the-counter versus prescribed therapy), and the type of provider (Novotny and others 2000). Over-the-counter NRT without physician oversight have been used in many countries for a number of years with good success. Although NRT are successful in treating nicotine addiction, the markets for NRT and other pharmacological therapies are more highly regulated and less affordable than are nicotine-containing tobacco products. Recent evidence indicates that the demand for NRT is related to economic factors, including price (Tauras and Chaloupka 2003). Policies that decrease the cost of NRT and increase availability—such as mandating private health insurance coverage of NRT, including such coverage in public health insurance programs, and subsidizing NRT for uninsured or underinsured individuals— would likely lead to substantial increases in the use of these products. Given the demonstrated efficacy of NRT in treating smoking, these policies could generate significant increases in smoking cessation. ment of antismuggling laws, and stronger penalties for those caught violating these laws (Joossens and others 2000). Recent analysis suggests that, even in the presence of smuggling, tax increases will reduce consumption and increase revenue (Merriman, Yurekli, and Chaloupka 2000). In contrast to the effectiveness of demand-side interventions, there is much less evidence that interventions aimed at reducing the supply of tobacco products are as effective in reducing cigarette smoking (Jha and Chaloupka 1999, 2000a). The U.S. experience provides mixed evidence about the effectiveness of limiting youth access to tobacco products in reducing youth tobacco use (U.S. DHHS 2000; Woolery, Asma, and Sharp 2000). In addition, the effective implementation and enforcement of these policies may require infrastructure and resources that do not exist in many low- and middleincome countries. A preliminary discussion is occurring in Canada about reducing its number of retail outlets for tobacco from the current 65,000. Neither the effect of such a move nor its enforcement costs are well known. Crop substitution and diversification programs are often proposed as a means of reducing the supply of tobacco. However, little evidence exists that such programs would significantly reduce the supply of tobacco, given that the incentives for growing tobacco tend to attract new farmers who would replace those who abandon tobacco farming (Jacobs and others 2000). Similarly, direct prohibition of tobacco production is not likely to be politically feasible, effective, or economically optimal. Finally, although trade liberalization has contributed to increases in tobacco use (particularly in low- and middle-income countries), restrictions on trade in tobacco and tobacco products that violate international trade agreements or draw retaliatory measures (or both) may be more harmful (Taylor and others 2000). INTERVENTIONS TO REDUCE THE SUPPLY OF TOBACCO The key intervention on the supply side is the control of smuggling. Recent estimates suggest that 6 to 8 percent of cigarettes consumed globally are smuggled (Merriman, Yurekli, and Chaloupka 2000). Of note, the tobacco industry itself has an economic incentive to smuggle, in part to increase market share and decrease tax rates (Joossens and others 2000; Merriman, Yurekli, and Chaloupka 2000). Although differences in taxes and prices across countries create a motive for smuggling, a recent analysis comparing the degree of corruption in individual countries with price and tax levels found that corruption within countries is a stronger predictor of smuggling than is price (Merriman, Yurekli, and Chaloupka 2000). Several governments are adopting policies aimed at controlling smuggling. In addition to harmonizing price differentials between countries, effective measures include prominent tax stamps and warning labels in local languages, better methods for tracking cigarettes through the distribution chain, aggressive enforce- EFFECTIVENESS AND COST-EFFECTIVENESS OF TOBACCO-CONTROL INTERVENTIONS Using a static model of the cohort of smokers alive in 2000, we estimate the number of deaths attributable to smoking over the next few decades that could be averted by (a) price increases, (b) NRT, and (c) a package of non-price interventions other than NRT. Cost-effectiveness of these policy interventions was calculated by weighing the approximate public sector costs against the years of healthy life saved, measured in DALYs. The details of an earlier version of this model have been published previously (Ranson and others 2002). Results of Model Projections The following is an updated analysis, using higher price increases and a greater effectiveness for NRT than did the Tobacco Addiction | 877 ©2006 The International Bank for Reconstruction and Development / The World Bank 79 Table 46.3 Reductions in Future Tobacco Deaths among Smokers Alive in 2000 from Price Increases of 10 Percent, 33 Percent, 50 Percent, and 70 Percent by World Bank Region Reduction in number of deaths (millions) World Bank region East Asia and the Pacific Baseline smoking-attributable deaths (millions) 173 (percent) Europe and Central Asia 51 (percent) Latin America and the Caribbean 40 (percent) Middle East and North Africa 13 (percent) South Asia 62 (percent) Sub-Saharan Africa 23 (percent) Low- and middle-income countries 362 (percent) High-income countries 81 (percent) World (percent) 443 10 percent price increase 33 percent price increase 50 percent price increase 70 percent price increase Low High Low High Low High Low High 2.9 8.7 9.6 27.5 14.5 37.5 20.3 46.2 (1.7) (5.0) (5.5) (15.9) (8.4) (21.7) (11.7) (26.8) 0.9 2.6 2.8 8.1 4.3 11.2 6.0 13.8 (1.7) (5.1) (5.6) (16.0) (8.5) (22.0) (11.8) (27.2) 0.7 2.1 2.3 6.7 3.5 9.5 4.9 11.6 (1.8) (5.3) (5.8) (16.8) (8.8) (23.7) (12.3) (29.1) 0.2 0.7 0.8 2.2 1.2 3.1 1.6 3.8 (1.7) (5.2) (5.8) (16.6) (8.7) (23.2) (12.2) (28.5) 0.9 2.6 2.9 8.5 4.4 12.5 6.2 16.0 (2.4) (8.6) (9.5) (27.7) (14.3) (40.6) (20.1) (52) 0.4 1.1 1.3 3.7 1.9 5.5 2.7 6.6 (1.6) (4.9) (5.4) (15.9) (8.2) (23.6) (11.5) (28.5) 6.0 17.9 19.7 56.8 29.8 79.2 41.7 98.2 (1.6) (4.9) (5.4) (15.7) (8.2) (21.9) (11.5) (27.1) 0.6 2.6 2.1 8.5 3.2 12.2 4.5 16.2 (0.8) (3.2) (2.6) (10.6) (4.0) (15.1) (5.6) (20.0) 6.6 20.5 21.8 65.3 33.0 91.5 46.2 114.3 (1.5) (4.6) (4.9) (14.7) (7.5) (20.7) (10.4) (25.8) Source: Authors’ calculations. original (Ranson and others 2002). This analysis is conservative in its assumptions about effectiveness and generous in its assumptions about the costs of tobacco control. Potential Effect of Price Increases. With a price increase of 33 percent, the model predicts that 22 million to 65 million smoking-attributable deaths will be averted worldwide, which is approximately equivalent to 5 to 15 percent of all smokingattributable deaths expected among those who smoke in 2000 (see table 46.3). Low- and middle-income countries account for about 90 percent of averted deaths. East Asia and the Pacific alone will account for roughly 40 percent of averted deaths. Total smoking-attributable deaths averted worldwide range from 33 million to 92 million for a 50 percent price increase and 46 million to 114 million for a 70 percent price increase. A 70 percent price increase would avert 10 to 26 percent of all smoking-attributable deaths worldwide. Of the tobacco-related deaths that would be averted by a price increase, 80 percent would be male, reflecting the higher overall prevalence of smoking in men. The greatest relative effect of a price increase on deaths averted is among younger cohorts. Note that these projections use conservative price increases. In certain countries, such as Poland and South Africa, recent tax increases have doubled the real price of cigarettes (Guindon, Tobin, and Yach 2002). Potential Effect of Nicotine Replacement Therapies. Provision of NRT with an effectiveness of 1 percent is predicted to result in the avoidance of about 3.5 million smokingattributable deaths (table 46.4). NRT of 5 percent effectiveness would have about five times the effect. Again, low- and middleincome countries would account for roughly 80 percent of the averted deaths. The relative effect of NRT (of 2.5 percent effectiveness) on deaths averted is 2 to 3 percent among individuals age 15 to 59 and lower among those age 60 and older (results not shown). Potential Effect of Non-price Interventions Other Than NRT. A package of non-price interventions, other than NRT, that decreases the prevalence of smoking by 2 percent is predicted to prevent about 7 million smoking-attributable deaths (more than 1.6 percent of all smoking-attributable deaths among those who smoked in 2000; see table 46.4). A package of interventions that decreases the prevalence of smoking by 10 percent would have an effect five times greater. Low- and middleincome countries would account for approximately four-fifths 878 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 80 Table 46.4 Reductions in Future Tobacco Deaths among Smokers Alive in 2000 from Price Increases of 33 Percent, Increased NRT Use, and a Package of Non-price Measures by World Bank Region Reduction in number of deaths (millions) World Bank region East Asia and the Pacific Baseline smoking-attributable deaths (millions) 173 (percent) Europe and Central Asia 51 (percent) Latin America and the Caribbean 40 (percent) Middle East and North Africa 13 (percent) South Asia 62 (percent) Sub-Saharan Africa 23 (percent) Low- and middle-income countries 362 (percent) High-income countries 81 (percent) World (percent) 443 33 percent price increase NRT effectiveness Low elasticity High elasticity 1 percent 9.6 27.5 1.4 (5.5) (15.9) 2.8 8.1 (5.6) 5 percent Non-price intervention effectiveness 2 percent 10 percent 6.9 2.8 13.8 (0.8) (4.0) (1.6) (8.0) 0.4 2.1 0.8 4.1 (16.0) (0.8) (4.0) (1.6) (8.1) 2.3 6.7 0.3 1.7 0.7 3.4 (5.8) (16.8) (0.8) (4.2) (1.7) (8.5) 0.8 2.2 0.11 0.6 0.2 1.1 (5.8) (16.6) (0.8) (4.2) (1.7) (8.4) 2.9 8.5 0.4 2.2 0.9 4.3 (9.5) (27.7) (1.4) (7.2) (2.8) (13.9) 1.3 3.7 0.2 0.9 0.4 1.8 (5.4) (15.9) (0.8) (4.0) (1.6) (7.9) 19.7 56.8 2.9 14.3 5.7 28.6 (5.4) (15.7) (0.8) (4.0) (1.6) (7.9) 2.1 8.5 0.6 3.1 1.2 6.1 (2.6) (10.6) (0.8) (3.8) (1.5) (7.6) 21.8 65.3 3.5 17.4 6.9 34.7 (4.9) (14.8) (0.8) (3.9) (1.6) (7.8) Source: Authors. of quitters and averted deaths. The greatest relative effect of non-price interventions on deaths averted would be among younger cohorts. Figure 46.5 summarizes the potential effect of a set of independent tobacco-control interventions, using 33 and 70 percent price increases (using a high elasticity of ⫺1.2 for low- and middle-income regions and ⫺0.8 for high-income regions), a 5 percent effectiveness of NRT, and a 10 percent reduction from non-price interventions other than NRT. In this cohort of smokers alive in 2000, approximately 443 million are expected to die in the next 50 years in the absence of interventions. A substantial fraction of these tobacco deaths are avoidable with interventions. Price increases have the greatest effect on tobacco mortality, with the most aggressive price increase of 70 percent having the potential to avert almost one-quarter of all tobacco deaths. Even a modest price increase of 33 percent could potentially prevent 66 million tobacco deaths over the course of the next 50 years. Although NRT and other non-price interventions are less effective than price increases, they can still avert a substantial number of tobacco deaths (18 million and 35 million deaths, respectively). The greatest effect of these tobacco-control interventions would occur after 2010, but a substantial number of deaths could be avoided even before then. Tobacco deaths (millions) 400 443 425 408 377 Death year 2030: 10 million deaths per year 300 328 versus 7 million deaths per year 200 100 0 2000 2010 2020 Baseline NRT with 5 percent effectiveness Non-price interventions with 10 percent reduction 2030 2040 2050 33 percent price increase 70 percent price increase Source: Authors. Note: Price increases assume a high price elasticity (–1.2 for low- and middle-income countries and –0.8 for high-income countries). Figure 46.5 Potential Effect of Tax Increases, NRT, and Non-price Interventions on Tobacco Mortality, 2000–50 Tobacco Addiction | 879 ©2006 The International Bank for Reconstruction and Development / The World Bank 81 Note that no attempt has been made in this analysis to examine the effect of combining the various packages of interventions (for example, price increases with NRT, or NRT and other non-price interventions). A number of studies have compared the effect of price and non-price interventions; few empirical attempts have been made to assess how these interventions might interact. Although price increases have been found in this analysis to be the most cost-effective antismoking intervention, policy makers should use both price and nonprice interventions to counter smoking. Non-price measures may be required to affect the most heavily dependent smokers, for whom medical and social support in stopping will be necessary. Furthermore, these non-price measures may be effective in increasing social acceptance and support of tobacco price increases. Cost-Effectiveness of Antismoking Interventions. In general, price increases are found to be the most cost-effective antismoking intervention. A 33 percent price increase (our base case scenario) could be achieved for a cost of US$13 to US$195 per DALY saved globally, or US$3 to US$42 in low-income countries and US$85 to US$1,773 in high-income countries. Wider access to NRT could be achieved for between US$75 and US$1,250 per DALY saved, depending on which assumptions are used. Non-price interventions other than NRT could be implemented for between US$233 and US$2,916 per DALY saved (table 46.5). Thus, NRT and other non-price measures are slightly less cost-effective than price increases but remain cost-effective in many settings. The cost-effectiveness of NRT is highly sensitive to the actual price of the NRT. NRT with a price of US$25 have a cost-effectiveness of US$75 per DALY compared with US$329 for an NRT price of US$150 (data not shown). For a given set of assumptions, the variation in the costeffectiveness of each intervention between low- and middleincome regions is relatively small and sensitive to the discount rate (data not shown). All three interventions are most costeffective in South Asia and Sub-Saharan Africa. The difference between low- and middle-income countries and high-income countries is more pronounced. For NRT, the cost per year of healthy life gained is 3 to 10 times higher in high-income countries than elsewhere. For non-price interventions other than NRT, the cost in high-income countries is 22 times higher, and for price increases, almost 42 times higher. Of note, the estimates of cost-effectiveness are given as wide ranges,which reflect the range of assumptions used. For price increases, the high-end estimates are roughly 25 times the low-end estimates, and this difference is consistent among the regions. For NRT, the high-end estimates are 2.5 to 10 times the low-end estimates, varying among the regions. For non-price interventions other than NRT, the high-end estimates are 20 times the low-end estimates, and this difference is consistent among the regions. The cost-effectiveness results can be compared against existing studies only for high-income countries because of a lack of studies situated elsewhere. Our estimates of deaths avoided for a 10 percent price increase are conservative compared with those of Moore (1996) and Warner (1986). Table 46.5 Range of Cost-Effectiveness Values for Price Increase, NRT, and Non-price Interventions, 2000 (2002 U.S. dollars per DALY saved) World Bank region East Asia and the Pacific Baseline smoking-attributable deaths (millions) 173 33 percent price increase NRT with effectiveness of 1 to 5 percent Non-price interventions with effectiveness of 2 to 10 percent Low-end estimate High-end estimate Low-end estimate High-end estimate Low-end estimate High-end estimate 2 30 65 864 40 498 Europe and Central Asia 51 3 42 45 633 55 685 Latin America and the Caribbean 40 6 85 53 812 109 1,361 Middle East and North Africa 13 6 89 47 750 115 1,432 South Asia 62 2 27 54 716 34 431 Sub-Saharan Africa Low- and middle-income countries High-income countries World 23 2 26 42 570 33 417 362 3 42 55 761 54 674 81 85 1,773 175 3,781 1,166 14,572 443 13 195 75 1,250 233 2,916 Source: Authors. 880 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 82 COMPREHENSIVE TOBACCO-CONTROL PROGRAMS In recent years, several governments, mostly in high-income countries, have adopted comprehensive programs to reduce tobacco use, often funded by earmarked tobacco tax revenues. The programs generally have similar goals for reducing tobacco use: • • • • preventing initiation among youths and young adults promoting cessation among all smokers reducing exposure to passive tobacco smoke identifying and eliminating disparities among population subgroups (U.S. DHHS 1994). Furthermore, the programs have one or more of four key components: community interventions engaging a diverse set of local organizations; countermarketing and health information campaigns; program policies and regulations (such as taxes, restrictions on smoking, bans on tobacco advertising, and access to better cessation treatments); and surveillance and evaluation of potential issues, such as smuggling (U.S. DHHS 1994). Programs have placed differing emphasis on these four components, with substantial diversity among the types of activities supported within each component. Disaggregating current tobacco-control program spending reveals that the greatest effect can be achieved through a focus on macro-level changes, such as policy change. Recent analyses from the United Kingdom and United States clearly indicate that these comprehensive efforts have been successful in reducing tobacco use and in improving public health (Farrelly, Pechacek, and Chaloupka 2003; Townsend, Roderick, and Cooper 1998; U.S. DHHS 1994). In California, for example, the state’s comprehensive tobacco-control program has produced a rate of decline in tobacco use double that seen in the rest of the United States. The cost of implementing control programs is low. Table 46.6 provides the estimated total costs of implementing price and NRT interventions by World Bank region. Current estimates of the costs of implementing a comprehensive tobacco-control program range from US$2.50 to US$10 per capita in the United States. The U.S. Centers for Disease Control and Prevention recommends spending US$6 to US$16 per capita for a comprehensive tobacco-control program in the United States (CDC 1999). Canadian spending on tobaccocontrol programs was approximately US$1.70 per capita in 1996 (Pechmann, Dixon, and Layne 1998). At the highest recommended spending level (US$16 per capita) in the United States, annual funding for a comprehensive tobacco program would equal only 0.9 percent of U.S. public spending, per capita, on health. Evidence from the United States demonstrates that states with the greatest prevalence of smoking have a greater marginal effect with their tobacco-control spending, suggesting that the potential gains from modest investments in comprehensive tobacco-control measures are large. Each US$10 spent per capita on tobacco control annually has resulted in a 55 percent reduction (variation of 20 to 70 percent) in per capita cigarette consumption (Tauras and others 2005). In the United States, US$10 translates into 0.03 percent of per capita GDP in 2003. Table 46.6 Estimated Cost of Price Intervention and NRT Programs (2002 U.S. dollars) Cost of NRT (US$25 to US$150) (millions) Cost for price increase (millions) World Bank region East Asia and the Pacific GDP (billions) Low-end estimate (0.02 percent GDP) 1,802 360 High-end estimate (0.05 percent GDP) 901 To treat 1 percent of current smokers To treat 5 percent of current smokers US$25 US$50 US$150 US$25 US$50 US$150 1,079 2,158 6,474 5,395 10,791 32,372 Europe and Central Asia 1,136 227 568 318 635 1,906 1,588 3,176 9,529 Latin America and the Caribbean 1,673 335 836 250 500 1,500 1,250 2,500 7,499 Middle East and North Africa 694 139 347 84 169 506 422 843 2,530 South Asia 655 131 327 2,312 1,926 3,853 11,558 2,312 1,926 Sub-Saharan Africa 319 64 159 868 723 1,447 4,340 868 723 6,279 1,256 3,138 4,911 6,111 15,686 24,553 20,490 54,579 High-income countries 25,992 5,198 12,996 3,034 2,529 10,114 15,172 3,034 2,529 World 32,271 6,454 16,134 7,945 8,640 25,800 39,725 23,524 57,108 Low- and middle-income countries Source: Authors. Tobacco Addiction | 881 ©2006 The International Bank for Reconstruction and Development / The World Bank 83 CONSTRAINTS TO EFFECTIVE TOBACCO-CONTROL POLICIES Although substantial evidence exists concerning the effectiveness of numerous policy interventions to reduce tobacco use, the use of these interventions globally is uneven and limited (see a more formal analysis in Chaloupka and others 2001). World Bank data reveal that ample room exists to increase tobacco taxes. In 1995, the average percentage of all government revenue derived from tobacco tax was 0.63 percent. Middle-income countries averaged 0.51 percent of government revenue from tobacco taxes, while lower-income countries averaged only 0.42 percent. An increase in cigarette taxes of 10 percent globally would raise cigarette tax revenues by nearly 7 percent, with relatively larger increases in revenues in high-income countries and smaller increases in revenues in low- and middle-income countries (Sunley, Yurekli, and Chaloupka 2000). Despite this evidence, price increases have been underused. Guindon, Tobin, and Yach (2002) studied 80 countries and found that the real price of tobacco, adjusted for purchasing power, fell in most developing countries from 1990 to 2000. Why does so much variation exist in tobacco-control policies? The political economy of tobacco control has been inadequately studied. A few plausible areas of interest are outlined here. First, the recognition of tobacco as a major health hazard appears to be the impetus for most of the tobaccocontrol policies in many high-income countries. Some evidence shows that improved national capacity and local needs assessment could increase the likelihood that tobacco-control measures will be adopted. For example, econometric analyses in South Africa geared to local policy requirements substantially increased the willingness of the government to implement tobacco-control policies (Abedian and others 1998). Second, tobacco-control budgets are only a fraction of what is required. Funding is needed not so much to implement programs as to fight off tobacco industry tactics and to build popular support for control. Third, the most obvious constraint to tobacco control is political opposition, which is difficult to quantify. Opposition from the tobacco industry is well organized and well funded (Pollock 1996). A key tool for addressing political opposition is earmarking tobacco taxes. Earmarking has been successfully used in several countries, including Australia, Finland, Nepal, and Thailand. Of the 48 countries currently in the World Health Organization’s European region, 12 earmark taxes for tobacco control and other public health measures. The average level of allocation is less than 1 percent of total tax revenue (WHO 2002). Earmarking does introduce clear restrictions and inefficiencies on public finance, and for this reason alone most macroeconomists do not favor earmarking, no matter how worthy the cause. However, analysis suggests that the efficiency or “dead-weight losses” from earmarking tobacco taxes are minimal (Hu, Xu, and Keeler 1998). Furthermore, earmarking tobacco taxes can be justified if governments use the funds to benefit those who pay (the benefits principle), provide assured funding for tobacco-control policies and programs, and secure public support for new or higher tobacco taxes. Earmarked taxes also have a political function in that they help concentrate political winners of tobacco control and thus influence policy. Earmarked funds that support broad health and social services (such as other disease programs) broaden the political and civil society support base for tobacco control. In Australia, broad political support from the Ministries of Sports and Education helped convince the Ministry of Finance that raising tobacco taxes was possible. Indeed, after an earmarked tax was passed, the Ministry of Finance went on to raise tobacco taxes further without earmarking (Galbally 1997). Additionally, targeting revenue from tobacco taxes to other health programs for the poorest socioeconomic groups could produce double health gains—reduced tobacco consumption combined with increased access to and use of health services. In China, a 10 percent increase in cigarette taxes would decrease consumption by 5 percent and would increase government revenue by 5 percent. The increased earnings could finance a package of essential health services for one-third of China’s poorest 100 million citizens in 1990 (Saxenian and McGreevey 1996). Finally, a key pillar in tobacco control that can help overcome some of these constraints is the Framework Convention on Tobacco Control (FCTC). The World Health Assembly of the World Health Organization adopted the FCTC in May 2003. It consists of a series of negotiated protocols within a general framework. The first three protocols are negotiations covering smuggling, advertising, and treatment of tobacco addiction. Countries agreeing to the negotiated protocols are to adopt appropriate legislation and, if necessary, implement the appropriate measures. As of February 2005, 168 countries had signed the FCTC, 57 had ratified it, and it had come into force on February 27, 2005. CONCLUSION Worldwide, only two large and growing causes of death exist. One is HIV-1 infection, and the other is tobacco. On current consumption patterns, about 1 billion people in the 21st century will be killed by their addiction to tobacco. Strong evidence shows that tobacco tax increases, the dissemination of information about health risks from smoking, restrictions on smoking in public places and workplaces, comprehensive bans on advertising and promotion, and increased access to cessation therapies are effective both in reducing tobacco use and in improving the health of populations. Despite this evidence, these policies, especially higher taxes, have been 882 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 84 applied aggressively only in a few high-income countries, covering a small proportion of the world’s smokers. Limited implementation of effective tobacco control in developing countries is due to political constraints as well as the lack of awareness of the unique effectiveness and cost-effectiveness of these interventions. ACKNOWLEDGMENTS We thank Allison Gilbert for help with the cost-effectiveness analyses and Hellen Gelband, Andra Ghent, and Dhirendra Sinha for comments. REFERENCES Abedian, I., R. van der Merwe, N. Wilkins, and P. Jha, eds. 1998. The Economics of Tobacco Control: Towards an Optimal Policy Mix. Cape Town, South Africa: Applied Fiscal Research Centre, University of Cape Town. Barnum, H. 1994. “The Economic Burden of the Global Trade in Tobacco.” Tobacco Control 3 (4): 358–61. CDC (U.S. Centers for Disease Control and Prevention). 1994. “Response to Increases in Cigarette Prices by Race/Ethnicity, Income, and Age Groups—United States, 1976–1993.” Morbidity and Mortality Weekly Report 43 (26): 469–72. ———. 1999. “Best Practices for Comprehensive Tobacco Control Programs.” Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Chaloupka, F. J., T. W. Hu, K. E. Warner, R. Jacobs, and A. Yurekli. 2000. “The Taxation of Tobacco Products.” In Tobacco Control in Developing Countries, ed. P. Jha and F. Chaloupka. Oxford, U.K.: Oxford University Press. Chaloupka, F., J. P. Jha, M.A. Corrao, V. Costa e Silva, H. Ross, C. Czart, and D. Yach. 2001. “The Evidence Base for Reducing Mortality from Smoking in Low and Middle Income Countries.” Commission on Macroeconomics and Health Working Paper Series. http://www. cmhealth.org/docs/wg5_paper7.pdf. Chaloupka, F. J., J. A. Tauras, and M. Grossman. 2000. “The Economics of Addiction.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Chinese Academy of Preventive Medicine. 1997. Smoking in China: 1996 National Prevalence Survey of Smoking Pattern. Beijing: China Science and Technology Press. Crispo, A., P. Brennan, K. H. Jockel, A. Schaffrath-Rosario, H. E. Wichmann, F. Nyberg, and others. 2004. “The Cumulative Risk of Lung Cancer among Current, Ex- and Never-Smokers in European Men.” British Journal of Cancer 91 (7): 1280–86. Cutler, D. M., and S. Kadiyala. 2003. “The Return to Biomedical Research: Treatment and Behavioral Effects.” In Measuring the Gains of Medical Research: An Economic Approach, ed. K. M. Murphy and R. H. Topel. Chicago: University of Chicago. Doll, R., R. Peto, J. Boreham, and I. Sutherland. 2004. “Mortality in Relation to Smoking: 50 Years’ Observation on Male British Doctors.” British Medical Journal 328 (7455): 1519–28. Emery, S., M. M. White, and J. P. Pierce. 2001. “Does Cigarette Price Influence Adolescent Experimentation?” Journal of Health Economics 20 (2): 261–70. Ezzati, M., and A. D. Lopez. 2003. “Estimates of Global Mortality Attributable to Smoking in 2000.” Lancet 362 (9387): 847–52. Farrelly, M. C., T. F. Pechacek, and F. J. Chaloupka. 2003. “The Impact of Tobacco Control Program Expenditures on Aggregate Cigarette Sales: 1981–2000.” Journal of Health Economics 22 (5): 843–59. Gajalakshmi, C. K., P. Jha, K. Ranson, and S. Nguyen. 2000. “Global Patterns of Smoking and Smoking-Attributable Mortality.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Gajalakshmi, V., R. Peto, T. S. Kanaka, and P. Jha. 2003. “Smoking and Mortality from Tuberculosis and Other Diseases in India: Retrospective Study of 43,000 Adult Male Deaths and 35,000 Controls.” Lancet 362 (9883): 507–15. Galbally, R. L. 1997. “Health-Promoting Environments: Who Will Miss Out?” Australia and New Zealand Journal of Public Health 21 (4 Spec. No.): 429–30. Gruber, J. 2003. “Government Policy toward Smoking: A New View from Economics.” Disease Control Priorities Project Working Paper Series. Paper presented at the Disease Control Priorities Project Nicotine Addiction Workshop, Mumbai, India, September 2003. Gruber, J., and B. Koszegi. 2001. “Is Addiction Rational? Theory and Evidence.” Quarterly Journal of Economics 116 (4): 1261–303. ———. 2002. “A Theory of Government Regulation of Addictive Bads: Optimal Tax Levels and Tax Incidence for Cigarette Taxation.” NBER Working Paper 8777. Cambridge, MA: National Bureau of Economic Research. Gruber, J., and S. Mullainathan. 2002. “Do Cigarette Taxes Make Smokers Happier?” NBER Working Paper 8872. Cambridge, MA: National Bureau of Economic Research. Guindon, G. E., S. Tobin, and D. Yach. 2002. “Trends and Affordability of Cigarette Prices: Ample Room for Tax Increases and Related Health Gains.” Tobacco Control 11 (1): 35–43. Gupta, P. C., and H. C. Mehta. 2000. “Cohort Study of All-Cause Mortality amongst Tobacco Users in Mumbai, India.” Bulletin of the World Health Organization 78 (7): 877–83. Hu, T. W., X. Xu, and T. Keeler. 1998. “Earmarked Tobacco Taxes: Lessons Learned.” In The Economics of Tobacco Control, ed. I. Abedian, R. van der Merwe, N. Wilkins, and P. Jha. Cape Town, South Africa: Applied Fiscal Research Centre, University of Cape Town. Jacobs, R., H. F. Gale, T. C. Capehart, P. Zhang, and P. Jha. 2000. “The Supply-Side Effects of Tobacco-Control Policies.” In Tobacco Control in Developing Countries, ed. P. Jha, and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Jamison, D. T., L. J. Lau, and J. Wang. 2005. “Health’s Contribution to Economic Growth in an Environment of Partially Endogenous Technical Progress.” In Health and Economic Growth: Findings and Policy Implications, eds. G. Lopez-Casasnovas, B. Rivera, and L. Currais. Cambridge: MIT Press, 67–91. Jarvis, M. J. 2004. “ABC of Smoking Cessation: Why People Smoke.” British Medical Journal 328 (7434): 277–79. Jha, P., and F. J. Chaloupka. 1999. Curbing the Epidemic: Governments and the Economics of Tobacco Control. Washington, DC: World Bank. ———. 2000a. “The Economics of Global Tobacco Control.” British Medical Journal 321 (7257): 358–61. ———, eds. 2000b. Tobacco Control in Developing Countries. Oxford, U.K.: Oxford University Press. Jha, P., P. Musgrove, F. J. Chaloupka, and A. Yurekli. 2000. “The Economic Rationale for Intervention in the Tobacco Market.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Jha, P., M. K. Ranson, S. N. Nguyen, and D. Yach. 2002. “Estimates of Global and Regional Smoking Prevalence in 1995 by Age and Sex.” American Journal of Public Health 92 (6): 1002–6. Tobacco Addiction | 883 ©2006 The International Bank for Reconstruction and Development / The World Bank 85 Joossens, L., F. J. Chaloupka, D. Merriman, and A.Yurekli. 2000. “Issues in the Smuggling of Tobacco Products.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Kenkel, D., and L. Chen. 2000. “Consumer Information and Tobacco Use.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Pechmann, C., P. Dixon, and N. Layne. 1998. “An Assessment of U.S. and Canadian Smoking Reduction Objectives for the Year 2000.” American Journal of Public Health 88 (9): 1362–67. Peck, R., F. J. Chaloupka, P. Jha, J. Lightwood. 2000. “Welfare Analyses of Tobacco.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka, 131–52. Oxford, U.K.: Oxford University Press. Koop, C. E. 2003. “Tobacco Addiction: Accomplishments and Challenges in Science, Health, and Policy.” Nicotine and Tobacco Research 5 (5): 613–19. Peto, R., S. Darby, H. Deo, P. Silcocks, E. Whitley, and R. Doll. 2000. “Smoking, Smoking Cessation, and Lung Cancer in the U.K. since 1950: Combination of National Statistics with Two Case-Control Studies.” British Medical Journal 321 (7257): 323–29. Lam, T. H., Y. He, Q. L. Shi, J. Y. Huang, F. Zhang, Z. H. Wan, and others. 2002. “Smoking, Quitting, and Mortality in a Chinese Cohort of Retired Men.” Annals of Epidemiology 12 (5): 316–20. Peto, R., and A. D. Lopez. 2001. “The Future Worldwide Health Effects of Current Smoking Patterns.” In Critical Issues in Global Health, ed. C. E. Koop, C. E. Pearson, and M. R. Schwarz. New York: Jossey-Bass. Laviolette, S. R., and D. van der Kooy. 2004. “The Neurobiology of Nicotine Addiction: Bridging the Gap from Molecules to Behavior.” Nature Reviews 5 (1): 55–65. Peto, R., A. D. Lopez, J. Boreham, and M. Thun. 2003. Mortality from Smoking in Developed Countries. 2nd ed. Oxford, U.K.: Oxford University Press. Levy, D. T., K. Friend, and E. Polishchuk. 2001. “Effect of Clean Indoor Air Laws on Smokers: The Clean Air Module of the SimSmoke Computer Simulation Model.” Tobacco Control 10 (4): 345–51. Peto, R., A. D. Lopez, J. Boreham, M. Thun, and C. Heath, Jr. 1994. Mortality from Smoking in Developed Countries, 1950–2000. Oxford, U.K.: Oxford University Press. Lightwood, J., D. Collins, H. Lapsley, and T. E. Novotny. 2000. “Estimating the Costs of Tobacco Use.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Picciotto, M. R. 2003. “Nicotine as a Modulator of Behavior: Beyond the Inverted U.” Trends in Pharmacological Sciences 24 (9): 493–99. Liu, B. Q., R. Peto, Z. M. Chen, J. Boreham, Y. P. Wu, J. Y. Li, and others. 1998. “Emerging Tobacco Hazards in China: 1. Retrospective Proportional Mortality Study of One Million Deaths.” British Medical Journal 317 (7170): 1411–22. Markou, A., and J. E. Henningfield. 2003. “Background Paper on the Neurobiology of Nicotine Addiction.” Paper presented at the Disease Control Priorities Project Nicotine Addiction Workshop, Mumbai, India, September 2003. Martinson, B. C., P. J. O’Connor, N. P. Pronk, and S. J. Rolnick. 2003. “Smoking Cessation Attempts in Relation to Prior Health Care Changes: The Effect of Antecedent Smoking-Related Symptoms?” American Journal of Health Promotion 18 (2): 125–32. Merriman, D., A. Yurekli, and F. J. Chaloupka. 2000. “How Big Is the Worldwide Cigarette Smuggling Problem?” In Tobacco Control in Developing Countries, ed. P. Jha, and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Pollock, D. 1996. “Forty Years On: A War to Recognise and Win—How the Tobacco Industry Has Survived the Revelations on Smoking and Health.” British Medical Bulletin 52 (1): 174–82. Ranson, M. K., P. Jha, F. J. Chaloupka, and S. N. Nguyen. 2002. “Global and Regional Estimates of the Effectiveness and Cost-Effectiveness of Price Increases and Other Tobacco Control Policies.” Nicotine and Tobacco Research 4 (3): 311–19. Raw, M., A. McNeill, and R. West. 1999. “Smoking Cessation: EvidenceBased Recommendations for the Healthcare System.” British Medical Journal 318 (7177): 182–85. Ross, H., F. J. Chaloupka, and M. Wakefield. 2001. “Youth Smoking Uptake Progress: Price and Public Policy Effects.” Research Paper 11. ImpacTeen, Health Research and Policy Centers, University of Illinois at Chicago. Saffer, H. 2000. “Tobacco Advertising and Promotion.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Molarius, A., R. W. Parsons, A. J. Dobson, A. Evans, S. P. Fortmann, K. Jamrozik, and others. 2001. “Trends in Cigarette Smoking in 36 Populations from the Early 1980s to the Mid-1990s: Findings from the WHO MONICA Project.” American Journal of Public Health 91 (2): 206–12. Saffer, H., and F. Chaloupka. 2000. “Tobacco Advertising: Economic Theory and International Evidence.” Journal of Health Economics 19 (6): 1117–37. Moore, M. J. 1996. “Death and Tobacco Taxes.” RAND Journal of Economics 27 (2): 415–28. Shiffman, S., and R. West. 2003. “Background Paper on the Psychology of Nicotine Addiction.” Paper presented at the Disease Control Priorities Project Nicotine Addiction Workshop, Mumbai, India, September 2003. Murphy, K. M., and R. H. Topel. 2003. “The Economic Value of Medical Research.” In Measuring the Gains of Medical Research: An Economic Approach, ed. K. M. Murphy and R. H. Topel, 41–73. Chicago: University of Chicago. Niu, S. R., G. H. Yang, Z. M. Chen, J. L. Wang, G. H. Wang, X. Z. He, and others. 1998. “Emerging Tobacco Hazards in China: 2. Early Mortality Results from a Prospective Study.” British Medical Journal 317 (7170): 1423–24. Novotny, T. E., J. C. Cohen, A. Yurekli, D. Sweaner, and J. de Beyer. 2000. “Smoking Cessation and Nicotine-Replacement Therapies.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. O’Malley, P. M., J. G. Bachman, and L. D. Johnston. 1988. “Period, Age, and Cohort Effects on Substance Use among Young Americans: A Decade of Change, 1976–86.” American Journal of Public Health 78 (10): 1315–21. Saxenian, H., and B. McGreevey. 1996. “China: Issues and Options in Health Financing.” Report 15278-CHA, World Bank, Washington, DC. Sunley, E. M., A. Yurekli, and F. J. Chaloupka. 2000. “The Design, Administration and Potential Revenue of Tobacco Excises.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Tauras, J. A. 1999. “The Transition to Smoking Cessation: Evidence from Multiple Failure Duration Analysis.” NBER Working Paper 7412. Cambridge, MA: National Bureau of Economic Research. Tauras, J. A., and F. J. Chaloupka. 2003. “The Demand for Nicotine Replacement Therapies.” Nicotine and Tobacco Research 5 (2): 237–43. Tauras, J. A., F. J. Chaloupka, M. Farrelly, G. A. Giovino, M. Wakefield, L. D. Johnston, and others. 2005. “State Tobacco Control Spending and Youth Smoking.” American Journal of Public Health 95 (2): 338–44. 884 | Disease Control Priorities in Developing Countries | Prabhat Jha, Frank J. Chaloupka, James Moore, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 86 Taylor, A. L., F. J. Chaloupka, E. Guindon, and M. Corbett. 2000. “The Impact of Trade Liberalization on Tobacco Consumption.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka, 343–64. Oxford, U.K.: Oxford University Press. ———. 2003. “The Costs of Benefits: Smoking and Health Care Expenditures.” American Journal of Health Promotion 18 (2): 123–24. Townsend, J. L. 1993. “Policies to Halve Smoking Deaths.” Addiction 88 (1): 43–52. Warner, K. E., F. J. Chaloupka, P. J. Cook, W. G. Manning, J. P. Newhouse, T. E. Novotny, and others. 1995. “Criteria for Determining an Optimal Cigarette Tax: The Economist’s Perspective.” Tobacco Control 4: 380–86. Townsend, J. L., P. Roderick, and J. Cooper. 1998. “Cigarette Smoking by Socio-Economic Group, Sex, and Age: Effects of Price, Income, and Health Publicity.” British Medical Journal 309 (6959): 923–26. Wasserman, J., W. G. Manning, J. P. Newhouse, and J. D. Winkler. 1991. “The Effects of Excise Taxes and Regulations on Cigarette Smoking.” Journal of Health Economics 10 (1): 43–64. U.K. Department of Health. 1992. “Effect of Tobacco Advertising on Tobacco Consumption: A Discussion Document Reviewing the Evidence.” London: U.K. Department of Health, Economics and Operational Research Division. Weinstein, N. D. 1998. “Accuracy of Smokers’ Risk Perceptions.” Annals of Behavioral Medicine 20 (2): 135–40. U.S. DHHS (United States Department of Health and Human Services). 1994. Preventing Tobacco Use amongst Young People. A Report of the Surgeon General. Atlanta: U.S. DHHS, Public Health Service, Centers for Disease Control, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. ———. 2000. Reducing Tobacco Use: A Report of the Surgeon General. Atlanta: U.S. DHHS, Public Health Service, Centers for Disease Control, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. U.S. Federal Trade Commission. 2004. Cigarette Report for 2002. Washington, DC: U.S. Federal Trade Commission. http://www.ftc. gov/reports/cigarette/041022cigaretterpt.pdf. WHO (World Health Organization). 1997. Tobacco or Heath: A Global Status Report. Geneva: WHO. ———. 2002. “The European Report on Tobacco Control Policy.” Paper presented at the WHO European Ministerial Conference for a Tobacco-free Europe, Warsaw. Document EUR/01/5020906/8, WHO Regional Office for Europe, Copenhagen. Woolery, T., S. Asma, and D. Sharp. 2000. “Clean Indoor-Air Laws and Youth Access.” In Tobacco Control in Developing Countries, ed. P. Jha and F. J. Chaloupka. Oxford, U.K.: Oxford University Press. Zatonski, W., and P. Jha. 2000. “The Health Transformation in Eastern Europe after 1990: A Second Look.” Warsaw: Marie Skeodowska-Curie Cancer Center and Institute of Oncology. Warner, K. E. 1986. “Smoking and Health Implications of a Change in the Federal Cigarette Excise Tax.” Journal of the American Medical Association 255 (8): 1028–32. Tobacco Addiction | 885 ©2006 The International Bank for Reconstruction and Development / The World Bank 87 ©2006 The International Bank for Reconstruction and Development / The World Bank 88 Chapter 47 Alcohol Jürgen Rehm, Dan Chisholm, Robin Room, and Alan D. Lopez This chapter provides an overview of epidemiology of alcohol use and health consequences as well as introducing costeffectiveness interventions to reduce alcohol-related harm. EPIDEMIOLOGY OF ALCOHOL USE AND ALCOHOL-RELATED DISEASE CONDITIONS Alcoholic beverages and the problems they engender have been familiar fixtures in human societies since the beginning of recorded history. Because alcohol is causally related to more than 60 International Classification of Diseases codes (Rehm, Room, Graham, and others 2003), disease outcomes are among the most important alcohol-related problems. Depending on the pattern of consumption, alcohol is also protective against diseases, most important among them, coronary heart disease (Rehm, Sempos, and Trevisan 2003). However, the net effect is negative, and 4 percent of the global burden of disease is attributable to alcohol, or about as much death and disability globally as is attributable to tobacco and hypertension (Ezzati and others 2002; WHO 2002). Alcohol thus constitutes a serious public health problem (Room, Babor, and Rehm 2005). Evidence-based preventive measures are available at both the individual and the population levels, with alcohol taxes, restrictions on alcohol availability, and drinking-and-driving countermeasures among the most effective policy options (Babor and others 2003). This chapter reviews the cost-effectiveness of different interventions in developing regions of the world. Dimensions of Alcohol Related to Disease The relationship between alcohol consumption and health and social outcomes is complex and multidimensional (Rehm and others 2004). As figure 47.1 shows, alcohol consumption is linked to acute and long-term health and social consequences through three intermediate mechanisms—toxic and beneficial biochemical effects, intoxication, and dependence (Babor and others 2003; Rehm, Room, Graham, and others 2003)—as follows: • Toxic and beneficial biochemical effects. These effects of alcohol consumption may influence chronic disease in either beneficial or harmful ways. Accepted beneficial effects include the influence of moderate drinking on coronary heart disease through reduction of plaque deposits in arteries, protection against blood clot formation, and promotion of blood clot dissolution (Zakhari 1997). Examples of harmful effects include increased risk for high blood pressure and for liver damage (Rehm, Room, Graham, and others 2003) and direct toxic effects on acinar cells triggering pancreatic damage (Apte, Wilson, and Korsten 1997) or hormonal disturbances (Emanuele and Emanuele 1997). These are just examples, because alcohol exposure is associated with a multitude of toxic effects on different organs. • Intoxication. Alcohol intoxication is a powerful mediator for acute health outcomes, such as accidental or intentional injuries or deaths, although intoxication can also be implicated in chronic health and social problems and in certain forms of heart disease. The subjective feeling of intoxication is mainly caused by the effects of alcohol on the central nervous system, and these effects are felt and can be measured even at light to moderate consumption levels (Eckardt and others 1998). • Dependence. Alcohol dependence is a clinical disorder in its own right, but it is also a powerful mechanism sustaining alcohol consumption and mediating its impact on both 887 ©2006 The International Bank for Reconstruction and Development / The World Bank 89 Patterns of drinking Average volume in terms of high-risk drinking occasions and also in terms of drinking in public settings and the proportion of drinking that occurs outside of meals (for further details, see Rehm and others 2004). Epidemiology of High-Risk Alcohol Use Toxic and beneficial biochemical effects Chronic disease Intoxication Dependence Accidents and injuries (acute disease) Acute social and psychological problems Chronic social and psychological problems Source: Adapted from Babor and others 2003. Figure 47.1 Model of Alcohol Consumption, Intermediate Outcomes, and Long-Term Consequences chronic and acute physiological and social consequences of alcohol (Drummond 1990). In the quantitative analyses reported in this chapter, alcohol dependence—and alcoholuse disorders (AUDs) in general—will be considered only as a health outcome related to high-risk alcohol use. This chapter, including the section on the cost-effectiveness of interventions, focuses primarily on health consequences, although later it briefly discusses the social consequences of high-risk drinking and recommended interventions. The epidemiological calculations are taken from Ezzati and others’ (2002) comparative risk analysis (CRA) and the World Health Organization (WHO) assessment of the global burden of disease (WHO 2002). (For further information, see Mathers and others 2003; Rehm, Rehn, and others 2003; Rehm, Room, Graham, and others 2003; Rehm, Room, Monteiro and others 2003; Rehm and others 2004). The CRA defines alcohol exposure using two measures: the average volume of alcohol consumption and patterns of drinking (figure 47.1). It then relates these exposure measures to disease outcomes. The average volume of consumption has been the conventional measure of exposure in alcohol epidemiology (Bruun and others 1975) and has been linked to many disease categories following the seminal work of English and others (1995; see also Rehm, Room, Graham, and others 2003). Patterns of drinking have been linked mainly to two categories of disease outcome: acute effects of alcohol (such as accidental and intentional injuries) and cardiovascular outcomes (mainly coronary heart disease). The CRA defines patterns of drinking primarily The intervention analyses presented in this chapter focus on average high-risk drinking, although patterns of drinking were also incorporated into the disease burden calculations. Highrisk drinking is defined in sex-specific terms as drinking 20 grams per day or more of pure alcohol on average for females and 40 grams per day or more of pure alcohol on average for males (a bottle of table wine contains about 70 grams of pure alcohol). This definition of high-risk drinking is fairly standard in alcohol epidemiology and was first introduced by English and others (1995) on the basis of Australian guidelines. Originally, English and others (1995) used two categories: hazardous drinking (defined as drinking between 20 and 40 grams per day of pure alcohol on average for females and between 40 and 60 grams per day of pure alcohol for males) and harmful drinking (defined as drinking 40 grams per day or more of pure alcohol on average for females and 60 grams per day or more of pure alcohol on average for males). These categories have been used in almost every comprehensive meta-analysis on alcohol and disease since 1995 (see Rehm, Room, Graham, and others 2003 for an overview). However, critics asserted that the terms hazardous drinking and harmful drinking were not neutral; thus, the CRA uses drinking categories II and III, referring to the term high-risk drinking when both categories are considered together. High-risk drinking thresholds differ by sex because the risk for chronic disease is related to lower volumes of drinking for women than for men; thus, the thresholds for high-risk drinking were set to reflect an approximately similar risk of chronic disease. Table 47.1 shows the distribution of high-risk drinking by age and by World Bank region. The table excludes the Middle East and North Africa because prevalence rates of high-risk drinking are considerably lower than 1 percent and this situation is unlikely to change in the near future. Calculating the burden of high-risk alcohol use that is avertable by means of effective interventions requires additional epidemiological data—in particular, rates of incidence to and remission from high-risk alcohol use and the relative fatality of high-risk alcohol users compared with non-high-risk alcohol users. We derived remission rates from studies of natural recovery from alcohol problems, which found an average of 10.9 years to remission (Sobell, Ellingstad, and Sobell 2000), with an adjustment of plus 20 percent for older age groups and minus 20 percent for younger age groups. We set the relative risk of mortality for high-risk alcohol users age 15 to 44 at 2.5 and the relative risk for older age groups at 1.3 for men and 1.4 for women (Gmel, Gutjahr, and Rehm 2003; Rehm, Gutjahr, and Gmel 888 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 90 Table 47.1 Prevalence of High-Risk Drinking by Gender, Age Group, and Region, 2000 (percentage of the population) Age group (years) Region Gender 15–29 30–44 45–59 60–69 70ⴙ Europe and Central Asia Male Female 20.8 11.2 18.7 10.4 21.4 11.5 15.2 7.9 8.1 5.7 Latin America and the Caribbean Male Female 9.7 6.8 11.1 7.5 10.6 6.5 7.9 5.8 3.4 3.1 Sub-Saharan Africa Male Female 10.4 3.1 14.3 4.7 12.9 5.1 11.3 3.2 8.4 2.2 East Asia and the Pacific Male Female 6.2 0.3 7.5 0.2 7.1 0.1 6.5 0.1 5.0 0.0 South Asia Male Female 0.8 1.2 2.5 0.4 0.3 0.4 0.1 0.0 0.0 0.0 High-income countries Male Female 18.0 10.9 17.9 8.7 16.2 9.8 10.9 6.8 7.6 5.4 Source: Authors’ calculations based on Rehm, Rehn, and others 2003 and Rehm and others 2004. Note: The criteria for high-risk drinking were set sex specific (for details see text). 2001). Using WHO disease-modeling software, we derived an internally consistent epidemiological profile of current high-risk alcohol use in each region, including specifications of incidence and the relative risk of mortality, with currently observed rates of prevalence, remission, and risk of mortality as inputs. A final input parameter is the disability level for high-risk alcohol use, which we estimated at 0.154 (where zero equals no disability); this is a weighted average based on the severity breakdown of high-risk drinkers from the CRA (80 percent category II, or hazardous; 20 percent category III, or harmful). The preference values for these health states of 0.11 and 0.33, respectively, are derived from Stouthard, Essink-Bot, and Bonsel (2000). Relationship between High-Risk Drinking and AUDs Assessing the relationship between high-risk drinking and AUDs is not a straightforward exercise. Even though high-risk drinking over a long period entails the risk of AUDs, that all people with AUDs are also high-risk drinkers does not automatically follow. First, neither the definition of alcohol dependence nor WHO’s (1993) definition of harmful use includes actual consumption levels. An individual is considered dependent if at least three of the following criteria apply: • strong desire or compulsion to take the substance • impaired control and physiological withdrawal if the substance is reduced or ceased • tolerance to the effects of the substance • preoccupation with use of the substance • persistent use despite clear evidence of harmful consequences. By contrast, harmful alcohol use is defined as a pattern of use that is causing damage to physical or mental health. Thus, whereas many of these criteria are associated with high-risk alcohol use, no strict classificatory rule indicates that people with AUDs are a subcategory of high-risk drinkers. Second, the prevalence of AUDs is often derived from surveys, where the operationalization usually requires that three symptoms be present in a lifetime and at least one of these criteria be present within the past 12 months (see, for example, Demyttenaere and others 2004, table 2). Thus individuals may be categorized as alcohol dependent even if they are currently abstaining from alcohol. Third, qualitative studies across a wide range of cultures have found that the criteria used for diagnosing AUDs often have different meanings and implications in different cultural settings (Room and others 1996; Schmidt and Room 1999). For instance, in the United States over the past decade, the level of reported AUDs increased despite decreases in high-risk drinking (Grant and others 2004). This fact has been explained in terms of changes in drinking norms and social attitudes during a period when the United States has become a “drier” culture. Thus, the measurement of AUDs is quite complex and culturally dependent. Moreover, AUDs are only one outcome of alcohol consumption and, in many parts of the world, not the most important one. As a result, we decided to focus on high-risk alcohol consumption rather than AUDs. Relationship between Alcohol Use and Disease Categories The exact procedures for quantifying the risk of disease attributable to alcohol are described in detail elsewhere (Rehm, Room, Graham, and others 2003; Rehm and others 2004). For most chronic disease categories, investigators have derived alcohol-attributable fractions of disease by combining Alcohol | 889 ©2006 The International Bank for Reconstruction and Development / The World Bank 91 prevalence and relative risk estimates based on meta-analyses (Corrao and others 2000; English and others 1995; Gutjahr, Gmel, and Rehm 2001; Ridolfo and Stevenson 2001; Single and others 1996, 1999). For depression, we drew alcoholattributable fractions from mental health surveys, looking at the rates of comorbidity and the order of onset of depression and alcohol disorders. For coronary heart disease, we modeled the interaction of average volumes and patterns of drinking based on multilevel analyses that include temporal information as covariates (Gmel, Rehm, and Frick 2003; Rehm and others 2004). For the final estimates, we based alcoholattributable fractions on these multilevel results for all countries, except for developed countries with relatively favorable drinking patterns (Australia, Japan, and countries in North America and Western Europe), which are not discussed here because the focus is on developing countries. For injuries, we took a similar multilevel approach to quantify the interaction of the average volume of consumption and patterns of drinking in determining alcohol-attributable fractions (Rehm and others 2004). Thus the analysis includes the following major disease categories: • chronic disease º cancer (mouth and oropharyngeal, esophageal, liver, female breast) º neuropsychiatric diseases (AUDs, unipolar major depression, epilepsy) º diabetes º cardiovascular diseases (hypertensive diseases, coronary heart disease, stroke) º gastrointestinal diseases (cirrhosis of the liver) º conditions arising during the perinatal period (low birthweight) • injury º unintentional injury (motor vehicle accidents, drowning, falls, poisonings, other unintentional injuries) º intentional injury (self-inflicted injuries, homicide, other intentional injuries). We did not include other disease categories that are clearly alcohol-related, such as fetal alcohol syndrome, because the current analysis was based on the CRA and was, thus, limited to the global burden-of-disease categories. Social Determinants of Exposure and Risk Alcohol-specific risks to health are in part determined and modified by social determinants. For example, Harrison and Gardiner (1999) find that for men age 25 to 69 in England and Wales in 1988–94, those in the lowest socioeconomic status category, unskilled labor, had a 15-fold greater risk for alcoholrelated mortality than professionals in the highest category had. These differences cannot be explained by the overall volume of drinking, which actually tended to be greater for those in higher socioeconomic groups. Rather, the differences can be explained by the fact that more of the drinking of those in lower socioeconomic status categories is in high-risk patterns; that is, depending on the use values for drinking in the culture, poor drinkers may see little point in wasting resources on drinking that is not to intoxication. Poorer drinkers are also likely to be less protected physically and socially from possible harm arising from drinking, such as injuries and chronic and infectious diseases. Mäkelä (1999) finds that multiple dimensions of socioeconomic status are required to capture all the adverse interactions of socioeconomic status with alcoholrelated mortality. A critical macroeconomic question is how a country’s level of economic development is related to alcohol-related risks to health. The impact of alcohol on disease and mortality may be more potent in countries with greater poverty and nutritional deficiencies (Isichei, Ikwuagu, and Egbuta 1993; Room and others 2002, 119–30). However, most of the risk relationships between alcohol and disease have been derived from studies in established market economies, and the extent of systematic research is currently insufficient to allow quantification of this phenomenon. As a result, the estimated disease burden cited here may be considered as a lower-bound estimate of the actual alcohol-attributable disease burden in developing countries. BURDEN OF DISEASE RELATED TO HIGH-RISK ALCOHOL USE In the following sections, the procedures to estimate alcoholrelated burden of disease are described, as well as the limitations of the used approach. Determining the Alcohol-Related Burden of Disease Table 47.2 breaks down alcohol-attributable disability-adjusted life years (DALYs) by disease category and World Bank region using a constant 3 percent per year discount rate, but with no age weighting. Results differ from those of the CRA (Ezzati and others 2002; Rehm and others 2004; WHO 2002) because of the use of non-age-weighted DALYs.1 Determining the Burden of Disease Related to High-Risk Alcohol Consumption In determining the burden of disease related to high-risk alcohol consumption, we first divided the burden of disease between chronic and acute disease. For chronic disease, we assume that almost the entire disease burden reported in the CRA is associated with high-risk alcohol use. Indeed, the overall disease burden in the CRA is an underestimate, because drinking up to 20 grams per day of pure alcohol by females and 890 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 92 Table 47.2 Alcohol-Attributable DALYs by Disease Category and World Bank Region, 2001 (thousands of DALYs) Disease category Europe and Central Asia Latin America and the Caribbean Sub-Saharan Africa Middle East and North Africa East Asia and the Pacific South Asia High-income countries World Chronic disease Maternal and perinatal conditions 12 7 39 1 2 29 6 105 526 296 635 25 2,820 189 1,103 5,594 Neuropsychiatric 2,159 3,315 1,035 89 4,726 1,444 4,752 17,600 Vascular 2,639 926 556 40 1,751 1,199 ⫺2,488 5,209 Other noncommunicable diseases 1,175 739 504 27 997 306 1,153 5,126 Subtotal chronic disease 6,511 5,283 2,769 182 10,296 3,167 4,526 33,634 Unintentional 4,127 1,984 2,308 135 3,613 2,222 1,753 15,619 Intentional 1,822 1,872 1,074 9 927 567 571 6,755 Cancer Injury Subtotal injury Total DALYs attributable to alcohol Total DALYs from all diseases Proportion of DALYs attributable to alcohol (percent) 5,949 3,856 3,382 144 4,540 2,789 2,324 22,374 12,460 9,139 6,151 326 14,836 5,956 6,850 56,008 116,502 104,287 344,754 65,570 346,225 408,655 149,161 1,535,871 10.7 8.8 1.8 0.5 4.3 1.5 4.6 3.6 Source: Authors’ calculations based on Rehm and others 2004 and WHO 2002. Note: Negative DALYs can occur because certain patterns of alcohol have cardio-protective effects. up to 40 grams per day of pure alcohol by males is globally associated with a net beneficial effect in relation to chronic disease. However, this effect occurs mainly in countries with moderate drinking patterns (Rehm, Sempos, and Trevisan 2003), which tend to be high-income countries (Rehm, Rehn, and others 2003). Although high-risk but regular drinking patterns may also have some beneficial effects, such effects are not important in countries with binge drinking patterns. (For the association between alcohol and coronary heart disease, see McKee and Britton 1998; Puddey and others 1999; Rehm, Sempos, and Trevisan 2003; for consequences on modeling the regional burden of disease, see Rehm and others 2004.) For injuries, which are considered to be acute outcomes, we started by separating out the proportion of injury not caused by high-risk drinking, which we accomplished by assuming that injuries are linearly related to per capita consumption (Rehm and others 2004).2 This assumption is probably conservative, because high-risk drinkers in countries with binge drinking patterns are likely to have more frequent and intensive drinking occasions, and the risk of injury usually rises logarithmically with the amount of drinking on a specific occasion (see, for example, National Highway Traffic Safety Administration 1992). Following this initial calculation, we could calculate the proportion of per capita consumption related to high-risk drinking in each region, thereby determining the proportion of injury caused by high-risk drinking (table 47.3). Together with our calculation of the chronic disease burden attributable to high-risk alcohol use, this percentage enabled us to estimate the overall disease burden attributable to high-risk alcohol use: whereas 3.6 percent of the global burden was attributable to alcohol drinking generally, 2.8 percent was attributable to high-risk drinking. Limitations of the CRA Approach The CRA’s estimates of the global and regional alcohol-related burden of disease are based on a number of assumptions, of which the following are the most crucial: • The estimates of per capita consumption and unrecorded consumption for different countries do not contain substantial measurement error. • The distribution of consumption as derived from surveys is similar to actual distribution in the population. • The relationships between alcohol and chronic disease derived from meta-analyses of cohort and case-control studies are stable among countries and regions. Alcohol | 891 ©2006 The International Bank for Reconstruction and Development / The World Bank 93 Table 47.3 DALYs Attributable to High-Risk Average Alcohol Consumption by Disease Category and Region, 2001 (thousands of DALYs) Disease category Europe and Central Asia Latin America and the Caribbean Sub-Saharan Africa East Asia and the Pacific South Asia High-income countries World Total chronic disease 6,510 5,283 2,770 10,296 3,167 4,526 33,634 Total injury 3,149 1,500 1,693 1,532 514 1,092 9,207 Total DALYs attributable to high-risk alcohol consumption 9,659 6,783 4,463 11,828 3,681 5,618 42,841 116,502 104,287 344,754 346,225 408,655 149,161 1,535,871 8.3 6.5 1.3 3.4 0.9 3.8 2.8 Total DALYs from all diseases Proportion of DALYs attributable to high-risk alcohol consumption Source: Authors’ calculations based on Rehm and others 2004 and WHO 2002. Some evidence indicates that per capita consumption can be reliably estimated, and information on this indicator is available for the vast majority of countries (Rehm, Rehn, and others 2003). With respect to survey information, reliability and worldwide coverage are lower. However, because the overall volume of consumption and, thus, the average volume per capita are based on production and sales estimates, the measure of the volume of drinking overall can be considered reliable. These factors leave the stability of relationships between alcohol and chronic disease as the most crucial part of our estimates. Some indications suggest that relative risks may not be the same in developing countries as in developed countries (for example, for tobacco and lung cancers, see Liu and others 1998). Thus, the CRA’s estimates may be biased, most likely toward an overestimation of the impact of alcohol. One additional problem pertains to the usual epidemiological approach as applied to alcohol. Most information about alcohol and chronic disease is derived from cohorts. Because cohorts are frequently not representative of the population as a whole, specific patterns of consumption such as binge drinking are often not represented, and thus their influence cannot be analyzed (Rehm, Gmel, and others 2003). Unfortunately, the patterns most often missing are those that are the most detrimental with respect to health; thus, the impact of alcohol on chronic diseases that are influenced by patterns of drinking other than average volumes is underestimated. INTERVENTIONS FOR REDUCING HIGH-RISK DRINKING The next two sections estimate the burden of disease attributable to high-risk alcohol consumption that is currently being averted or could be averted by a range of personal and nonpersonal intervention strategies and calculate the expected costs and cost-effectiveness of such interventions. Methods and analyses draw on Chisholm and others (2004), adjusted as necessary to conform to the analytical standards of this volume, including the specification of all costs in U.S. dollars rather than international dollars. Population Model We determined intervention effectiveness using a state transition population model (Lauer and others 2003), which traces the development of a regional population taking into account births, deaths, and the specified risk factor—in this case, highrisk alcohol use. In addition to population size and structure, the population model uses a number of epidemiological parameters (incidence and prevalence, remission, and causespecific and residual rates of mortality) and assigns age- and gender-specific health state valuations to both the disease in question and to the nondiseased population. The output of the model is an estimate of the total healthy life years experienced by the population over a lifetime period (100 years). We ran the model for a number of possible scenarios, including no intervention at all (natural history), current intervention coverage, and scaled-up coverage of current and possible new interventions. For the intervention scenarios, we used an implementation period of 10 years for an intervention program (after which epidemiological rates return to their natural history levels), from which we derived the number of additional DALYs averted each year compared with the case for no intervention at all. We discounted DALYs at 3 percent but did not age weight them. Effectiveness A number of interventions have been evaluated and shown to be effective in reducing alcohol use, yet their level of 892 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 94 implementation remains low in all but a handful of countries and their potential effect on population-level health has rarely been assessed. By contrast, some interventions without clearly established effects continue to be widely used, including, for example, mass media public information campaigns and school-based education aimed at reducing alcohol consumption. Recent reviews of measures to reduce alcohol misuse have assessed the quality of the evidence for four types of interventions specifically aimed at reducing high-risk alcohol use (Babor and others 2003; Ludbrook and others 2002): • policy and legislative interventions, including taxation of alcohol sales, laws on drunk driving, restrictions on retail outlets, and controls on advertising • measures to better enforce these interventions, such as random breath testing of drivers • mass media and other awareness campaigns • brief interventions with individual high-risk drinkers. On the basis of these reviews, we included the following strategies and intervention effects in our analysis: drinkingand-driving legislation and random breath testing, taxation of alcoholic beverages, reduced hours of sale in retail outlets, and advertising bans (included as population-based interventions) and so-called brief interventions (included as interventions aimed at personal behavior). We considered including one other intervention strategy—mass media or school-based awareness campaigns—but omitted it in the final analysis on the grounds that evidence for its effectiveness was weak, both in terms of methodological quality and in terms of its effect on consumption (as opposed to transfer of information or knowledge alone) (Babor and others 2003; Edwards and others 1994; Foxcroft and others 2003; Foxcroft, Lister-Sharp, and Lowe 1997; Ludbrook and others 2002). Drunk-Driving Legislation and Random Breath Testing. Drunk-driving laws and reinforcement policies, such as random breath testing of drivers, influence fatal and nonfatal traffic injuries among both high-risk alcohol users and other members of the population, such as passengers and pedestrians. We assessed two independent effects on alcohol-related traffic injuries, but note that evidence for these effects comes from the developed countries, where road infrastructures and driving patterns may differ significantly from those in the developing world. The first intervention was drunk-driving laws, estimated to reduce traffic fatalities by 7 percent if widely implemented across a region. The second was enforcement by random breath testing, estimated to reduce fatalities by 6 to 10 percent in regions partially implementing such a strategy and by 18 percent with wide implementation. The effect on nonfatal injuries was estimated to be a reduction of 15 percent (Peek-Asa 1999; Shults and others 2001). In each region, we applied these estimated effects to the proportion of total deaths and of years lived with a disability attributed to alcohol-related traffic accidents (table 47.4). Taxation on Alcoholic Beverages. Excise taxation on alcoholic beverages primarily affects the incidence of drinking through reduced consumption. Effects are measured in terms of price elasticity, which relates the change in consumption to the size of the price increase (table 47.5). We derived price elasticities, adjusted downward by one-third to reflect possible reduced price responsiveness among high-risk drinkers, with reference to preferred type of alcoholic beverage (beer, wine, or spirits) by region, built up from country-level data (WHO 2003b). This downward adjustment of price elasticities for high-risk drinkers is a conservative approach; most of the literature found similar effects on high-risk and dependent drinkers as on social users (Babor and others 2003; see also Farrell, Manning, and Finch 2003). Price elasticities ranged from ⫺0.3 for the most preferred beverage category to ⫺1.5 for the least preferred (Babor and others 2003; Levy and Ornstein 1983). For a beer-drinking region where wine is the second-most preferred beverage type, for example, elasticities were set as follows: beer ⫺0.3, wine ⫺1.0, distilled spirits ⫺1.5. We performed sensitivity analysis around these elasticities. We evaluated three rates of excise tax on alcoholic beverages: the current rate of tax, a 25 percent increase over the current rate, and a 50 percent increase over the current rate. We adjusted estimated reductions in the incidence of high-risk alcohol use by the observed or expected level of unrecorded consumption resulting from illicit production and smuggling (for instance, an estimated 35 percent of alcohol consumption in Eastern Europe and Central Asia is unrecorded, a proportion that was modeled to increase by 10 to 15 percent with the tax increases). In regions with rates of unrecorded consumption already greater than 50 percent (South Asia and Sub-Saharan Africa), tax increases can actually have a regressive impact on incidence if accompanied by a rise in the already high level of unrecorded (and therefore untaxed) consumption. Reduced Hours of Sale in Retail Outlets. Access to and availability of alcohol can be dramatically reduced by prohibition or rationing, but implementing and sustaining such strategies without adverse effects, such as black markets and poisonings from home-produced alcohol, present considerable challenges. A more modest strategy is to reduce the hours of sale of retail outlets selling alcoholic beverages (for example, no sales for offpremise consumption for a 24-hour period at the weekend), which in Scandinavia has reduced consumption and alcoholrelated harm (Leppänen 1979; Nordlund 1984; Norström and Skog 2003). On the basis of these studies, we modeled a modest reduction of 1.5 to 3.0 percent in the incidence of high-risk drinking and 1.5 to 4.0 percent in alcohol-related traffic Alcohol | 893 ©2006 The International Bank for Reconstruction and Development / The World Bank 95 Table 47.4 Effectiveness of Drinking-and-Driving Legislation and Its Enforcement (per 100,000 population) Effectiveness of drinking-and-driving laws and random breath testing Attributable fractions (per 100,000 deaths) Deaths attributed to traffic accidentsa Deaths attributed to alcohol-related traffic accidentsa Reduced deaths (per 100,000) Reduced years lost due to disability (per 100,000) World Bank region WHO subregion Europe and Central Asia Europe B Male Female 1,473 542 657 74 141 16 77 6 Europe C Male Female 2,197 799 1,396 223 299 48 193 30 Americas B Male Female 4,358 1,514 2,053 220 439 47 148 12 Americas D Male Female 2,599 1,093 861 101 184 22 64 6 Africa D Male Female 2,159 1,079 417 90 89 19 43 9 Africa E Male Female 2,075 1,027 803 123 172 26 107 17 Southeast Asia B Male Female 7,809 2,343 1,993 127 427 27 164 8 Western Pacific B Male Female 3,629 1,790 723 157 155 34 66 12 Southeast Asia D Male Female 3,689 1,451 591 53 126 11 45 3 Latin America and the Caribbean Sub-Saharan Africa East Asia and the Pacific South Asia Sex Source: Deaths attributed to traffic accidents: WHO 2003a; deaths attributed to alcohol-related traffic accidents: Rehm and others 2004. B ⫽ low child mortality, low adult mortality; C ⫽ low child mortality, high adult mortality; D ⫽ high child mortality, high adult mortality; E ⫽ high child mortality, very high adult mortality. a. Percentages for all age groups combined shown here. fatalities, depending on the regional pattern of drinking, with the largest effects in regions with the highest levels of high-risk drinking occasions. Advertising Bans. Public health specialists are becoming increasingly interested in the effect of a comprehensive ban on alcohol advertising, including advertising on television and through radio and billboards. However, available evidence from econometric studies suggests a modest effect on consumption at best, even for a comprehensive ban, arguably because of the continuing presence of other alcohol marketing strategies, such as product placement or event sponsorship (Grube and Agostinelli 2000; Saffer 2000; Saffer and Dave 2002). Here we consider the potential effects of a comprehensive advertising ban (television, radio, and billboards) by modeling a 2 to 4 percent reduction in the incidence of high-risk alcohol use, depending on regional drinking patterns. Brief Interventions. We modeled brief interventions (such as physician advice provided in primary health care settings), which involve a small number of education sessions and psychosocial counseling, to influence the prevalence of high-risk drinking by increasing remission and reducing disability. Efficacy reviews of brief interventions reveal an estimated 13 to 34 percent net reduction in consumption among high-risk drinkers (Higgins-Biddle and Babor 1996; Moyer and others 2002; Whitlock and others 2004), which, if applied to the total population at risk, would reduce the overall prevalence of high-risk drinking by 35 to 50 percent, equivalent to a 14 to 18 percent improvement in the rate of recovery over no treatment at all. After taking into account adherence (70 percent) and potential treatment coverage in the population (50 percent of high-risk drinkers), however, we estimated remission rates to be between 4.9 and 6.4 percent higher than natural history rates. 894 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 96 Table 47.5 Effect of Taxation on the Incidence of High-Risk Alcohol Use Prevalence by preferred beverage (percent) ©2006 The International Bank for Reconstruction and Development / The World Bank 97 World Bank region WHO subregiona Most preferred Next preferred Least preferred Europe and Central Asia Europe B 0.45 (spirits) 0.30 (beer) Europe C 0.68 (spirits) Americas B Latin America and the Caribbean Sub-Saharan Africa East Asia and the Pacific South Asia Rate of taxation by preferred beverage (percent) Price increases (percent)b Elasticity Elasticity Elasticity ⴚ0.3, most ⴚ1.0, next ⴚ1.5, least preferred preferred preferred Most preferred Next preferred Least preferred 0.25 (wine) 0.29 0.36 0.44 0.13 0.16 0.20 0.12 0.15 0.18 (current rate) (25 percent increase) (50 percent increase) ⫺0.04 ⫺0.05 ⫺0.06 ⫺0.08 ⫺0.09 ⫺0.11 ⫺0.11 ⫺0.13 ⫺0.15 0.21 (beer) 0.11 (wine) 0.65 0.81 0.98 0.13 0.16 0.20 0.25 0.31 0.38 (current rate) (25 percent increase) (50 percent increase) ⫺0.08 ⫺0.09 ⫺0.10 ⫺0.08 ⫺0.09 ⫺0.11 0.53 (beer) 0.30 (spirits) 0.17 (wine) 0.16 0.20 0.24 0.49 0.61 0.74 0.22 0.28 0.33 (current rate) (25 percent increase) (50 percent increase) ⫺0.03 ⫺0.03 ⫺0.04 Americas D 0.58 (spirits) 0.39 (beer) 0.03 (wine) 0.26 0.33 0.39 0.21 0.26 0.32 0.25 0.31 0.38 (current rate) (25 percent increase) (50 percent increase) Africa D 0.79 (beer) 0.16 (spirits) 0.05 (wine) 0.36 0.45 0.54 0.41 0.51 0.62 0.35 0.44 0.53 Africa E 0.49 (beer) 0.30 (spirits) 0.21 (wine) 0.28 0.35 0.42 0.50 0.63 0.75 Southeast Asia B 0.88 (spirits) 0.12 (beer) 0.00 (wine) 0.30 0.38 0.45 Western Pacific B 0.88 (spirits) 0.11 (beer) 0.01 (wine) Southeast Asia D 0.89 (spirits) 0.11 (beer) 0.00 (wine) Nonrecorded or untaxed consumption (percent) Effect (percent)c Baseline Lowerd Uppere 0.34 (current rate) 0.37 (10 percent increase) 0.39 (15 percent increase) ⫺0.05 ⫺0.05 ⫺0.06 ⫺0.03 ⫺0.04 ⫺0.04 ⫺0.06 ⫺0.07 ⫺0.08 ⫺0.20 ⫺0.24 ⫺0.27 0.36 (current rate) 0.40 (10 percent increase) 0.42 (15 percent increase) ⫺0.06 ⫺0.06 ⫺0.07 ⫺0.04 ⫺0.05 ⫺0.05 ⫺0.08 ⫺0.09 ⫺0.09 ⫺0.22 ⫺0.25 ⫺0.28 ⫺0.18 ⫺0.22 ⫺0.25 0.29 (current rate) 0.32 (10 percent increase) 0.34 (15 percent increase) ⫺0.08 ⫺0.09 ⫺0.10 ⫺0.06 ⫺0.06 ⫺0.07 ⫺0.10 ⫺0.12 ⫺0.13 ⫺0.04 ⫺0.05 ⫺0.06 ⫺0.12 ⫺0.14 ⫺0.16 ⫺0.20 ⫺0.24 ⫺0.27 0.22 (current rate) 0.24 (10 percent increase) 0.25 (15 percent increase) ⫺0.06 ⫺0.07 ⫺0.08 ⫺0.04 ⫺0.05 ⫺0.05 ⫺0.08 ⫺0.09 ⫺0.10 (current rate) (25 percent increase) (50 percent increase) ⫺0.05 ⫺0.06 ⫺0.07 ⫺0.19 ⫺0.23 ⫺0.25 ⫺0.26 ⫺0.30 ⫺0.34 0.77 (current rate) 0.85 (10 percent increase) 0.89 (15 percent increase) ⫺0.02 ⫺0.01 ⫺0.01 ⫺0.01 ⫺0.01 ⫺0.01 ⫺0.03 ⫺0.02 ⫺0.02 0.38 0.48 0.57 (current rate) (25 percent increase) (50 percent increase) ⫺0.04 ⫺0.05 ⫺0.06 ⫺0.22 ⫺0.26 ⫺0.29 ⫺0.28 ⫺0.32 ⫺0.36 0.47 (current rate) 0.52 (10 percent increase) 0.55 (15 percent increase) ⫺0.08 ⫺0.08 ⫺0.09 ⫺0.06 ⫺0.06 ⫺0.06 ⫺0.10 ⫺0.11 ⫺0.11 0.40 0.50 0.60 0.00 0.00 0.00 (current rate) (25 percent increase) (50 percent increase) ⫺0.05 ⫺0.05 ⫺0.06 ⫺0.19 ⫺0.22 ⫺0.25 0.00 0.00 0.00 0.36 (current rate) 0.39 (10 percent increase) 0.41 (15 percent increase) ⫺0.04 ⫺0.05 ⫺0.05 ⫺0.03 ⫺0.03 ⫺0.03 ⫺0.05 ⫺0.06 ⫺0.07 0.17 0.21 0.26 0.09 0.11 0.14 0.11 0.14 0.17 (current rate) (25 percent increase) (50 percent increase) ⫺0.03 ⫺0.04 ⫺0.04 ⫺0.06 ⫺0.07 ⫺0.08 ⫺0.10 ⫺0.12 ⫺0.14 0.27 (current rate) 0.32 (10 percent increase) 0.31 (15 percent increase) ⫺0.02 ⫺0.03 ⫺0.03 ⫺0.02 ⫺0.02 ⫺0.02 ⫺0.03 ⫺0.04 ⫺0.04 0.40 0.50 0.60 0.25 0.31 0.38 0.00 0.00 0.00 (current rate) (25 percent increase) (50 percent increase) ⫺0.06 ⫺0.07 ⫺0.08 ⫺0.13 ⫺0.16 ⫺0.18 0.00 0.00 0.00 0.79 (current rate) 0.87 (10 percent increase) 0.91 (15 percent increase) ⫺0.01 ⫺0.01 ⫺0.01 ⫺0.01 ⫺0.01 ⫺0.01 ⫺0.02 ⫺0.01 ⫺0.01 Source: WHO 2003b. a. B ⫽ low child mortality, low adult mortality; C ⫽ low child mortality, high adult mortality; D ⫽ high child mortality, high adult mortality; E ⫽ high child mortality, very high adult mortality. b. Price rise caused by tax ⫽ [percentage of tax/(1 ⫹ percentage of tax)] ⫻ elasticity ⫻ 2/3 (high-risk drinkers less responsive). c. Effect ⫽ sum of (prevalence ⫻ price increase) for each beverage ⫻ (1 ⫺ percentage of unrecorded consumption). d. Lower-range elasticities ⫽ ⫺0.2, ⫺0.7, ⫺1.2. e. Upper-range elasticities ⫽ ⫺0.4, ⫺1.3, ⫺2.0. Costs Costs covered in the analysis include program-level costs associated with running the intervention (such as administration, training, and media costs) and patient-level costs (such as costs of primary care visits). Program-level costs include resource inputs used in the production of an intervention at a level above that of the patient or providing facility, such as central planning, policy, and administration functions, as well as resources devoted to preventive programs, such as the enforcement of drunk-driving legislation by police officers (Johns and others 2003). We derived estimated quantities of resources required to implement each intervention for 10 years at the national, provincial, and district levels with reference to the region’s prevailing characteristics—for example, the stability and efficiency of tax systems, the volume of traffic (for breath testing), and the strength of antidrinking sentiment as indicated by existing alcohol controls (advertising bans, restricted sales). In this analysis, patient-level resource inputs used in the provision of a given health care intervention (for example, hospital inpatient days, outpatient visits, medications, and laboratory tests) are relevant only to brief interventions. We estimated an average of four primary care visits per year for the intervention itself, plus an additional 0.33 outpatient visits (20 percent ⫻ 1.67 visits) and 0.25 inpatient days (5 percent ⫻ 5 days) (see, for example, Fleming and others 2000). We applied these patient-level resource inputs to the 50 percent of prevalent high-risk alcohol users in receipt of brief advice in year 1 and (because we model an enduring effect for 10 years) year 6 and to the 50 percent of incident cases in years 2 to 5 and 7 to 10. Note that, throughout, the costing does not include possible offsetting revenues for the government, for instance, from drunk-driving convictions and, in particular, from the revenues likely to result from increased alcohol taxes. Unit costs and prices of program- and patient-level resource inputs include the salaries of central administrators; the capital costs of vehicles, offices, and furniture; and the cost per outpatient visit (see chapter 7 for an overview of the costing methodology, plus prices by World Bank region). All costs are expressed in U.S. dollars for 2001 and are discounted at an annual rate of 3 percent. males (the proportion for random breath testing rises to 80 to 90 percent because of the higher proportion of deaths and injuries attributed to traffic accidents among men). A clear difference is also apparent between regions with relatively high rates of high-risk alcohol use (that is, prevalence in the total population greater than 5 percent) and regions with generally low levels of high-risk drinking (that is, less than 2 percent). As shown in table 47.6, in the three regions with a higher prevalence of high-risk alcohol use—Europe and Central Asia, Latin America and the Caribbean, and Sub-Saharan Africa— the most effective interventions were taxation and brief physician advice to individual high-risk drinkers, with each averting more than 500 DALYs per million population per year. The remaining control strategies—random breath testing, reduced access to alcoholic beverage retail outlets, and a comprehensive advertising ban—mainly produced effects in the range of 200 to 400 DALYs averted per million population per year. In the two regions with lower rates of high-risk drinking (particularly among the female population), by contrast, the burden that is avertable through taxation is very much reduced (10 to 100 DALYs averted per million population per year). In South Asia, the most effective intervention is enforcement of drinkingand-driving laws by means of random breath testing, because of the higher rate of traffic-related injuries than elsewhere as well as the low levels of high-risk drinking. Population-Level Costs Table 47.7 summarizes the costs and cost-effectiveness of each intervention and of two combination strategies by region. The most costly interventions to implement in all regions were random breath testing and brief physician advice in primary care. The higher costs of brief advice stem from a combination of patient-level costs in the provision of the intervention itself (an average annual cost of US$7 to US$20 per treated case), plus program costs associated with administration and training primary care providers (15 to 40 percent of total costs). Random breath testing is also a relatively resource-intensive intervention to implement because of the need for regular sobriety checkpoints administered by law enforcement officers. Other interventions, including taxation, had a per capita cost in the range US$0.02 to US$0.13, depending in part on the efficiency of the tax collection system and the degree of antidrinking sentiment. COST-EFFECTIVENESS OF INTERVENTIONS In the following section, we provide results relating to the population-level health effects, costs, and cost-effectiveness of the evidence-based interventions previously reviewed. Population-Level Effects Except for random breath testing, two-thirds of the total population-level health gain from these interventions was among Population-Level Cost-Effectiveness Compared with doing nothing, taxation is the most costeffective population-level strategy in Europe and Central Asia, Latin America and the Caribbean, and Sub-Saharan Africa, the three regions with a relatively high prevalence of high-risk drinking (table 47.7). At the current rate of tax, for example, each DALY averted costs US$104 to US$225, equivalent to 4,435 to 9,633 DALYs averted per US$1 million expenditure. 896 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 98 Table 47.6 Population-Level Effects of Interventions to Reduce High-Risk Alcohol Use by World Bank Region Coveragea (percent) Europe and Central Asia Latin America and the Caribbean Sub-Saharan Africa East Asia and the Pacific South Asia 20,241 12,894 6,685 6,263 2,652 Burden of disease (DALYs/million population) Total effect (DALYs averted/ million population/year) Excise tax on alcoholic beverages (current situation) 0.95 685 586 697 83 13 Excise tax on alcoholic beverages (25 percent increase) 0.95 756 654 724 96 10 Excise tax on alcoholic beverages (50 percent increase) 0.95 828 719 764 109 8 Reduced access to alcoholic beverage retail outlets 0.95 441 287 386 203 32 Comprehensive advertising ban on alcohol 0.95 395 243 406 226 20 Random breath testing of motor vehicle drivers 0.80 284 307 197 181 125 Brief advice to heavy drinkers by a primary care physician 0.50 1,328 713 539 362 80 Combination: highest tax ⫹ brief advice 2,048 1,360 1,237 447 83 Combination: highest tax ⫹ advertising ban ⫹ random breath testing ⫹ brief advice 2,551 1,784 1,715 790 210 Reduction in current burden (percent) Excise tax on alcoholic beverages (current situation) 0.95 0.03 0.05 0.10 0.01 0.01 Excise tax on alcoholic beverages (25 percent increase) 0.95 0.04 0.05 0.11 0.02 0.00 Excise tax on alcoholic beverages (50 percent increase) 0.95 0.04 0.06 0.11 0.02 0.00 Reduced access to alcoholic beverage retail outlets 0.95 0.02 0.02 0.06 0.03 0.01 Comprehensive advertising ban on alcohol 0.95 0.02 0.02 0.06 0.04 0.01 Random breath testing of motor vehicle drivers 0.80 0.01 0.02 0.03 0.03 0.05 Brief advice to heavy drinkers by a primary care physician 0.50 0.07 0.06 0.08 0.06 0.03 Combination: highest tax ⫹ brief advice 0.10 0.11 0.19 0.07 0.03 Combination: highest tax ⫹ advertising ban ⫹ random breath testing ⫹ brief advice 0.13 0.14 0.26 0.13 0.08 Source: Chisholm and others 2004. a. Refers to the modeled percentage of all high-risk drinkers exposed to the intervention. Advertising bans had a cost per unit of effect similar to that of reduced access to sales outlets, US$134 to US$380, equivalent to 2,631 to 7,442 averted DALYs per US$1 million dollars expenditure, whereas random breath testing had the highest estimated cost per DALY averted: US$973 to US$1,856 per DALY, approximately 500 to 1,000 DALYs averted per US$1 million dollars expenditure. Brief physician advice provided in primary care settings had an average cost-effectiveness in the range of US$204 to US$502 per DALY averted, or close to 2,000 to 5,000 averted DALYs for every US$1 million expenditure. Alcohol | 897 ©2006 The International Bank for Reconstruction and Development / The World Bank 99 Table 47.7 Costs and Cost-Effectiveness of Interventions to Reduce High-Risk Alcohol Use by World Bank Region Coveragea Europe and Latin America (percent) Central Asia and the Caribbean Sub-Saharan East Asia and Africa the Pacific South Asia Total cost (US$ million/year/million population) Excise tax on alcoholic beverages (current situation) 0.95 0.10 0.13 0.07 0.04 0.04 Excise tax on alcoholic beverages (25 percent increase) 0.95 0.10 0.13 0.07 0.04 0.04 Excise tax on alcoholic beverages (50 percent increase) 0.95 0.10 0.13 0.07 0.04 0.04 Reduced access to alcoholic beverage retail outlets 0.95 0.10 0.10 0.06 0.03 0.03 Comprehensive advertising ban on alcohol 0.95 0.07 0.09 0.05 0.03 0.02 Random breath testing of motor vehicle drivers 0.80 0.53 0.47 0.19 0.18 0.07 Brief advice to heavy drinkers by a primary care physician 0.50 0.36 0.36 0.11 0.08 0.04 Combination: highest tax ⫹ brief advice 0.44 0.48 0.18 0.12 0.07 Combination: highest tax ⫹ advertising ban ⫹ random breath testing ⫹ brief advice 0.97 0.97 0.39 0.30 0.15 Cost-effectiveness relative to no intervention (US$/DALY averted) Excise tax on alcoholic beverages (current situation) 0.95 141 225 104 516 2,671 Excise tax on alcoholic beverages (25 percent increase) 0.95 127 202 100 447 3,654 Excise tax on alcoholic beverages (50 percent increase) 0.95 116 184 95 394 4,641 Reduced access to alcoholic beverage retail outlets 0.95 216 340 152 146 827 Comprehensive advertising ban on alcohol 0.95 185 380 134 123 1,123 Random breath testing of motor vehicle drivers 0.80 1,856 1,542 973 984 531 Brief advice to heavy drinkers by a primary care physician 0.50 270 502 204 224 462 Combination: highest tax ⫹ brief advice 216 350 143 269 845 Combination: highest tax ⫹ advertising ban ⫹ random breath testing ⫹ brief advice 381 546 229 383 707 DALYs averted/US$ million expenditure Excise tax on alcoholic beverages (current situation) 0.95 7,107 4,435 9,633 1,937 374 Excise tax on alcoholic beverages (25 percent increase) 0.95 7,847 4,953 10,007 2,239 274 Excise tax on alcoholic beverages (50 percent increase) 0.95 8,590 5,442 10,553 2,536 215 Reduced access to alcoholic beverage retail outlets 0.95 4,638 2,940 6,580 6,856 1,209 Comprehensive advertising ban on alcohol 0.95 5,417 2,631 7,442 8,139 891 Random breath testing of motor vehicle drivers 0.80 539 648 1,027 1,016 1,882 Brief advice to heavy drinkers by a primary care physician 0.50 3,705 1,992 4,891 4,460 2,163 Combination: highest tax ⫹ brief advice 4,627 2,859 7,016 3,718 1,184 Combination: highest tax ⫹ advertising ban ⫹ random breath testing ⫹ brief advice 2,621 1,833 4,364 2,612 1,415 Source: Chisholm and others 2004. a. Refers to the modeled percentage of all high-risk drinkers exposed to the intervention. Starting from the current situation in these regions, the most efficient strategies for reducing high-risk alcohol use would be tax increases (additional gains are obtained at virtually no extra cost because the costs of tax administration and enforcement remain relatively constant whatever the rate of tax), followed by the introduction or escalation of comprehensive advertising bans on alcohol products, reduced access to retail outlets, and the provision of brief interventions such as physician advice in primary care. Even a multifaceted strategy made up of an increase in taxation plus full implementation of the other interventions considered here has a favorable ratio of costs to health benefits. In East Asia and the Pacific and South Asia, the two regions with lower rates of high-risk alcohol use, a comparison of intervention costs and effects to a no-intervention scenario reveals that current practice—namely, excise taxes on alcoholic 898 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 100 beverages—is not the most efficient response to the existing burden of alcohol use. The reduced efficiency of taxation in these lower-prevalence regions is related both to the distribution of the fixed costs of administering and enforcing alcohol tax legislation across a smaller target population of drinkers and to underlying drinking patterns: more than 85 percent of all alcohol consumption falls into a single preferred drink category, spirits, which therefore diminishes the scope for reducing the consumption of less preferred but more elastic categories of alcoholic beverages. In South Asia, targeted strategies such as brief physician advice and random breath testing have the lowest cost per DALY averted (around US$500), while taxation policies are the most expensive at more than US$2,500 per DALY averted. In East Asia and the Pacific, the most costeffective interventions are brief physician advice, a comprehensive ban on advertising, and reduced access to retail outlets (below US$250 per DALY averted). Implications and Limitations of Sectoral Cost-Effectiveness Analyses This cost-effectiveness analysis offers a new approach to generating economic evidence that can inform public health policy on alcohol in a wide range of cultural and epidemiological settings (Chisholm and others 2004). Resulting estimates of cost-effectiveness can inform policy makers not only by determining the efficiency of existing resource allocation and practices, but also by identifying priorities for future alcohol control strategies. Furthermore, use of a common methodology enables comparison with cost per DALY estimates for other risk factors or disease entities, which may constitute an important argument when considering priorities for the allocation of scarce health care resources. However, the application of a broad sectoral approach using entire regions as the unit of analysis clearly limits the approach’s use in specific country contexts, where demographic or epidemiological characteristics, as well as treatment costs and coverage, may not coincide with estimates for the region as a whole. In addition, extrapolation of the extent of intervention effects from relatively information-rich countries to other sociocultural settings lessens the precision of derived estimates of population-level health gains. Although an ongoing analytical step is to calibrate results at the country level, the primary purpose and utility of the sectoral approach is to identify interventions that are clearly cost-effective as opposed to those that clearly do not seem to offer good value for money. In this respect, the primary conclusion to be drawn from the analysis is that in regions with high or moderate rates of high-risk alcohol use, a number of intervention strategies can have a notable effect on population health, including both individual-based interventions, such as brief physician advice at the primary care level, as well as population-wide measures, such as taxation of alcoholic beverages. Of these, taxation has the most sizable and least resourceintensive effect on reducing the avertable burden of high-risk alcohol use. In regions where high-risk alcohol use represents less of a public health burden, targeted approaches such as brief physician advice as well as other intervention strategies that restrict the supply or promotion of alcoholic beverages appear to be the most cost-effective mechanisms, although greater empirical support for the efficacy of these interventions in these localities is clearly needed before considering their widespread implementation. Even though sectoral cost-effectiveness analysis pursues a societal perspective, considerable challenges remain in relation to the appropriate measurement of certain societal costs and effects that fall outside the boundaries of the health system. Therefore, this analysis has not been able to successfully capture potential reductions in workforce and household productivity losses among high-risk drinkers, nor does it incorporate the economic costs associated with alcohol-related crime, violence, and harm reduction. It also does not value the time spent by patients and informal caregivers in seeking or providing care and support. Including these modest additional costs and substantial incremental effects is likely to improve the costeffectiveness ratios of all interventions, but to a variable and currently unknown extent. ECONOMIC BENEFITS OF INTERVENTIONS By design, estimates of the burden of alcohol do not include most social harm and harm to people other than the drinker; however, the burden of social problems from drinking can be at least as significant as the health burden. The burden attributable to alcohol in the CRA estimates is actually a substantial underestimate of the full harm alcohol imposes on human welfare. The estimates reported earlier reflect primarily the chronic disease and injury effects of drinking. Because the CRA focused on disease and disability, the estimates were not designed to take account of the social harm and problems that are particular to alcohol and that result for the drinker and for others as a consequence of a person’s drinking (Klingemann and Gmel 2001). These problems are quite prevalent in many populations (Room and others 2003) and are also affected by the interventions listed earlier. Some information on the relative burden of alcohol for social services versus health services is available for a handful of societies. In an estimate of the staffing and service costs attributable to alcohol in different service systems in Scotland for fiscal year 2001/02, for instance, health services accounted for only 21 percent of the estimated costs, whereas social services accounted for 19 percent, and criminal justice and fire services accounted for 60 percent (Catalyst Health Consultants 2001, 3). Alcohol | 899 ©2006 The International Bank for Reconstruction and Development / The World Bank 101 If those estimates are used as a rough gauge of the burden to society, the illness and disability burden of alcohol may thus constitute half or less of the total burden when social problems are also taken into consideration. Thus, policies that affect the levels of alcohol-related health and social harm not only are a matter of intervening to save people from the detrimental effects of their own behavior, but also potentially have a broader effect on the health and wellbeing of families and of associates of drinkers. This issue is especially relevant for women: even though men predominate among high-risk drinkers worldwide (Rehm and others 2004; Room and others 2002), women bear much of the burden of harm from others’ drinking, not only in such forms as domestic violence, but also in such forms as diversion of family resources from greater needs. IMPLEMENTATION OF CONTROL STRATEGIES: LESSONS OF EXPERIENCE The following paragraphs provide a few concrete examples of interventions or policy changes that illustrate the actual implementation and effects of control strategies in developing societies (the examples are taken from Room and others 2002). Tax Rate Reduction and the Resulting Disease Burden in Mauritius Mauritius, an island nation in the Indian Ocean, has a population of about 1 million. These people are of Indian, African, European, and Chinese origin. By religious affiliation, 53 percent are Hindu, 29 percent are Christian, and 17 percent are Muslim. Tourism is the third-ranked industry in terms of hard currency earnings. In June 1994, the government drastically lowered customs duties on imported alcoholic beverages to 80 percent from rates that had ranged from 200 percent for wine to 600 percent for whisky and other spirits (Abdool 1998). The government made the change under pressure from the hotel industry, which claimed that tourists were not purchasing enough alcohol because of its high prices (Lee 2001). Other reasons given for the change were to reduce unofficial imports from abroad and to make better, more refined alcoholic beverages available to the local population. Despite little evidence to support the view, there were claims in the public discussion that better-quality alcohol would result in fewer health problems. The effects of the change were felt mainly by Mauritians rather than tourists, as follows: • Arrests for driving with blood alcohol over the legal limit made primarily in connection with traffic crashes increased by 23 percent between 1993 and 1997. • Admissions of alcoholism cases to the island’s psychiatric hospital shot up in 1994. The 1995 rate was more than twice the 1993 rate, and the rate rose again slightly in 1996 and 1997. Medical specialists in Mauritius agree that patients with alcohol problems account for an increasing portion of admissions in general medical wards and now represent between 40 and 50 percent of bed occupancy (Abdool 1998). • Age-adjusted death rates per 100,000 population for chronic liver disease and cirrhosis rose from 32.8 for males and 4.0 for females in 1993 to 42.7 for males and 5.3 for females in 1996 (WHO 1999, 2000). Even though available statistics are limited, the reduction in alcohol import taxes clearly had a substantial negative effect on the health of Mauritians. Thus, the government’s 1997 call for control measures for alcohol—specifically, new permits for licensed premises, increased excise duties on alcohol, and limitations on bars’ opening hours—was not surprising. Alcohol taxes were increased somewhat in the 1999/2000 budget (U.S. Department of State 1999). However, an analysis by World Bank staff that did not take health effects into account called for further reductions in maximum tariff rates, identifying Mauritius as having an antitrade bias on the basis of the structure of its alcohol and tobacco taxes (Hinkle and HerrouAragon 2001). Wallace and Bird (2003) suggest the following general principles for setting and collecting alcohol taxes in the context of developing societies from the perspective of revenue generation rather than public health (see also Tax Policy Chief Directorate 2002): • Countries around the world need revenues they can raise relatively efficiently, but this need is probably more critical in the case of developing nations. That said, alcohol taxes are probably a good bet for future revenues. • Excise taxes on alcohol should be set by alcohol content, rather than as a percentage of the price. • Tax rates should be logically defined so that alcoholic beverages with similar alcohol content are treated similarly, with stronger alcohol beverages taxed more heavily. • Analyses of revenue-maximizing rates should be conducted to determine a range of tax rates that is likely to maximize government revenues. • Tax systems should be designed to be as simple as possible to allow for the maximum efficiency of tax administration. Reduced Access through Locational Prohibition in Brazil The second example involves the institution of a new control on alcohol availability in an environment where it is likely to be combined with driving. Although we have modeled the effects of another, better studied availability control (namely, closing on a weekend day), a wide variety of possible 900 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 102 measures is available to control the time and place of alcohol purchase or drinking (Babor and others 2003; Room and others 2002). Even though in this case the particular control was extremely limited in scope, it appears to have had measurable effects. Traffic deaths are an important source of mortality in Brazil, amounting to 3.6 percent of overall mortality. The few available studies suggest that alcohol plays a significant role in traffic casualties. For instance, one study in São Paulo found positive blood alcohol levels in 72 percent of pedestrian deaths and 32 percent of driver and passenger deaths of persons age 13 and older (Carlini-Cotrim and Chasin 2000). In 1985, motivated by concern about alcohol and impaired driving and about the lax enforcement of drinking and driving laws, a conservative party politician from the state of São Paulo introduced legislation to prohibit alcohol sales in commercial facilities that had access to state highways. Even though the bill passed in the legislature, its implementation was delayed by the state’s alcohol producers and commercial and industrial federations, which claimed that the law would be a barrier to improved facilities for travelers, would encourage people to carry bottles in their cars, and would restrict individual freedoms. Discussion in the press was also generally unsympathetic. In August 1988, however, a new state governor from the same party implemented the law. At that time, the press was slightly more supportive. Since then, the law has been on the books, although site visits to restaurants and snack bars along a state highway in 1997 suggested a low level of compliance. In 1995, another legislator from the same party proposed repealing the law on the grounds that no studies proved that it lowered traffic accidents. The repeal passed the legislature without significant public debate, but the state governor vetoed it. Undaunted, the same legislator then proposed a law to criminalize buying as well as selling alcohol along state highways. That law passed but has not yet been implemented. A study by Carlini-Cotrim, Pinsky, and Serrano Barbosa (1998) assesses the effects of the intervention. Finding data for a controlled study comparing traffic casualties on state highways with casualties on federal highways, which were unaffected by the law, proved impossible. The best data available were on crashes and crashes resulting in injuries per 10,000 vehicles traveling on three short highway systems administered by a private agency. Linear regressions on those data for 1983–93 showed that the law had made a significant difference in the number of accidents resulting in injuries on all three roads and a significant difference in all accidents on two of the roads. A separate analysis on estimated accidents and accidents with injuries per 10,000 vehicles in two geographic areas of the state did not show significant effects of the law. Overall, the analyses do provide some support for the law having a beneficial effect on the rate of traffic casualties. Drunk-Driving Enforcement in South Africa No published studies are available of the implementation of random breath testing in a developing country. However, some data are available on a campaign to increase drunk-driving enforcement in South Africa, a strategy that has often shown some effects, although weaker and less lasting than those of random breath testing. The minister of finance launched a short-term campaign, ARRIVE ALIVE, for the period October 1997 to January 1998, in response to the high rate of traffic fatalities and injuries. The campaign’s main aim was to mobilize all available traffic policing, control, and education resources to reduce traffic accidents on South African roads by at least 5 percent, especially in the Western Cape, Gauteng, and KwaZulu Natal provinces, because 75 percent of all accidents occurred in those provinces. The ARRIVE ALIVE campaign targeted, in turn, what were considered the three critical factors having the greatest impact on injuries: failing to wear seat belts, drinking and driving, and speeding. Unofficially, the campaign came to be called “belts, booze, and bats out of hell.” As many of the parties interested in road safety as possible were involved, with funding drawn from a variety of government and business sources. The campaign included a number of components particularly relevant to alcohol use. New equipment purchased by the provinces included alcohol screening devices, alcohol evidentiary units, and so-called booze buses (vehicles containing all the technology needed to check breath and blood alcohol levels). Sentences were increased to underline the point that traffic violations are serious offenses, with a three-month suspension of a driver’s license and an increased maximum fine for a first conviction for drunk driving and with license suspension for one to five years for second offenders. Traffic supervisors underwent intensive training courses before the start of the campaign. Because the aim of the campaign included educating road users, advertisements covering aspects of the campaign were run on the radio, on television, and in movie theaters throughout the country. Supplements were published in national and provincial newspapers. Private companies, such as a supermarket chain and an automobile manufacturer, also promoted the campaign. A national transportation center, established to collect and collate data from local and provincial authorities, operated for 12 hours every day throughout the campaign. Traffic authorities staffed an additional 80 roadside communications points, and at selected points on certain routes, road signs were erected and updated to display the percentage of speed limit and drinking-and-driving violations and the rate of seat belt use in that area. A total of 776 enforcement points were set up on 195 strategic routes in the selected provinces. Posters, pamphlets, key rings, and license decals were produced for distribution and Alcohol | 901 ©2006 The International Bank for Reconstruction and Development / The World Bank 103 display at roadblocks in the three provinces. Between October 1, 1997, and January 17, 1998, 6,674 notices of prosecution were issued for alcohol-related traffic offenses, 83 percent of which were issued in the intervention provinces. Comparison studies showed a decrease in the drinking rate of drivers in the three provinces, whereas the other six provinces, as a group, showed an increase. KwaZulu Natal had the lowest drinking rate of all drivers throughout the campaign (3 to 7 percent), and the Western Cape had the most dramatic decrease (from 12.0 to 9.3 percent in October). Except for in Gauteng, the drinking rates for pedestrians decreased from more than 15 percent to less than 7 percent. Overall, during the months targeted, drinking-and-driving rates decreased by 2 to 4 percent, as measured by breath testing. The total number of crashes decreased by 8 percent, and fatalities dropped by 9 percent. The ratio of benefits to costs for the intervention was estimated as 4 to 1, based on an investment in the campaign of R50 million, or about US$4.4 million at 2002 rates (ARRIVE ALIVE Campaign 2000). Despite the potential inconvenience of roadblocks and other enforcement activities, the public generally perceived the campaign positively. The liquor retail and hospitality industries complained about decreased sales, and tow truck operators complained about reduced business. Even though driver behavior improved during the focus months, violations often increased after the focus was changed, for example, from drunk driving to seat belt use. This finding emphasizes the need for sustained enforcement as opposed to ad hoc campaigns. (This example was summarized from ARRIVE ALIVE Campaign 2000 and Cerff and Plüddemann 1998.) Implementation of Brief Interventions in Several Developing Countries In the first phase of the WHO Collaborative Project on Identification and Management of Alcohol Related Problems (Saunders and Aasland 1987), a screening measure suitable for use in both developing and developed countries—the alcoholuse disorders identification test—was developed to identify people at risk for alcohol problems among those attending primary health care services. In the second phase, a multicenter clinical trial of brief intervention procedures designed to reduce the health risks associated with hazardous alcohol use was carried out in primary health care settings in Australia, Bulgaria, Costa Rica, Kenya, Mexico, Norway, the Soviet Union, the United Kingdom, the United States, and Zimbabwe (Babor and others 1994). The project’s aims were to study the influence of simple advice and brief counseling, to examine the moderating role of reduced consumption on the prevention of alcohol-related problems, and to evaluate the cross-national generalizability of brief intervention techniques. The project’s hypothesis was that the amount of change in alcohol consumption over a nine-month period would be proportional to the intensity of the intervention provided by a trained primary health care professional. The results showed a significant effect of interventions on both consumption and intensity of drinking among males, but the intensity of the intervention was not related to the amount of change in drinking behavior; 5 minutes of simple advice turned out to be as effective as 20 minutes of brief counseling (Babor and Grant 1992). The female sample was too small for the results to attain significance, and the intervention did not significantly affect men’s frequency of dependence symptoms, problems related to alcohol, or concern expressed by others (WHO Brief Intervention Study Group 1996). The findings suggest that in a population of high-risk drinkers, behavior change is more a function of motivational factors and social influence than of the moderation skills and social learning techniques that behavioral self-control training packages typically use. Changes in drinking were not attributable solely to the small number of patients who achieved an abstinence goal, nor to the small number who gave up daily or almost daily drinking. Rather, changes seem to have been distributed across a broad spectrum of the drinkers who reduced their consumption by small, but clinically meaningful, amounts. RESEARCH AND DEVELOPMENT AGENDA Research and development needs in the area of alcohol consumption are large and multidimensional. The work reported in this chapter represents best estimates from the available data, some of which were developed to fill the needs of the analysis; however, we cite few figures for the developing world for which we can say that the underlying data are so good that they could not usefully be improved. Nevertheless, more and better data are available on alcohol than on many other health topics. The health and social burdens of alcohol are clearly extremely large in most developing societies. Thus, the most urgent focus should be on development and evaluation projects to study the outcomes of various policy and program interventions. The projects must necessarily be attuned to what is politically feasible in a particular time and place. They are likely to include natural experiment studies, where the research tracks the effects of changes that governments undertake, whether those changes are expected to increase or to decrease the extent of alcohol problems. Where possible, the projects should include experimental and quasi-experimental studies, whereby the effects of a change at intervention sites are studied in comparison to outcomes at control sites, with random 902 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 104 assignment where possible. Costing data should be included to permit cost-effectiveness analysis. Also important are process studies—that is, research on how policy makers decide on policy changes, how they implement them, and what the reactions and sequelae are. For example, deciding to introduce a new alcohol tax may be the easiest part of an initiative, but actually implementing it in a developing society with a great deal of unrecorded alcohol in the informal market and with poorly guarded borders may be much more difficult. Currently, no international mechanism or nexus exists for developing and disseminating practical knowledge about implementing effective alcohol control strategies between developing countries. At this time, nearly all studies of alcohol interventions come from a limited range of developed countries. Extending knowledge and experience in and between developing societies is urgently needed. A secondary need, but one that is also important, is to extend the epidemiological database in developing societies on levels and patterns of drinking and on the health and social consequences of drinking. To this end, better estimation of unrecorded alcohol consumption is needed. Which dimensions of drinking patterns matter for what kinds of outcomes needs to be studied in the context of different kinds of developing societies. Studies of the effects that the interaction of drinking levels and patterns with poverty and social exclusion have on the extent of alcohol-related problems are also necessary. Because most of our knowledge about the health effects of drinking concerns mortality, studies of alcohol’s role in various kinds of morbidity should be emphasized. Another area where data are lacking is the social harm arising from drinking, for which we cannot presently make the kinds of estimates that are possible to make for harm to health. Developing and reaching consensus on ways to measure the social harm caused by drinking is a substantial agenda for both the developed and the developing world. Developing the epidemiological database can provide clues to etiology to be pursued further by biomedical and social researchers and, thus, offers hope for the development of new treatments or preventive interventions. It can provide information on the distribution of drinking patterns and problems in subpopulations that can be used to guide targeting and prevention and treatment priorities. However, from a short-term policy perspective, the most important function of developing the epidemiological database in a particular country may be providing a base for creating political will for action. For example, the development of devices to measure blood and breath alcohol and the collection of data on drinking and driving that they made possible were prerequisites for developing the political will and support for implementing drinking-and-driving countermeasures in industrial countries. CONCLUSION The burden of disease attributable to alcohol in the developing world is considerable, and the social harm not accounted for in this analysis increases the costs. However, known interventions can reduce the burden by up to 25 percent, depending on the region of the world. Compared with other interventions in the health care field, these interventions are quite cost-effective, but given the nature of many of the interventions, caution is needed. In particular, the following recommendations can be given: • Interventions and research about their effectiveness are based mostly on experiences from established market economies; thus, the levels of effectiveness estimated in our analysis should be treated as broad indications. Depending on actual methods of implementation, individual interventions could be more or less effective. • Interventions should ideally be modeled on the basis of the specific environment (that is, countries or provinces) and on the harm distribution in the respective environment, including social harm. • General principles, such as restricting access to alcohol, should be attuned to local cultures and traditions when interventions are formulated. • Population measures must take into account the complex interplay of public opinion and balance the interests of different groups and stakeholders with conflicting values. One of these stakeholders is, of course, the alcohol industry. If policy makers keep these principles in mind, reducing the alcohol-related health burden could be one of the most costeffective targets of population-level health programs in developing countries. This target is even more attractive because the measures discussed will also reduce the alcohol-related social burden, thereby further contributing to development. NOTES 1. The global burden of disease attributable to alcohol is 4.0 percent using age-weighted DALYs and 3.6 percent using non-age-weighted DALYs. This difference can be explained by the many alcohol-attributable outcomes occurring during adolescence and young adulthood, when age weights are higher. 2. The CRA defined per capita consumption as average consumption of pure alcohol per person 15 years old or older. REFERENCES Abdool, R. 1998. “Alcohol Policy and Problems in Mauritius.” Paper prepared for the World Health Organization Alcohol Policy in Developing Societies Project, World Health Organization, Geneva. Apte, M. V., J. S. Wilson, and M. A. Korsten. 1997. “Alcohol-Related Pancreatic Damage: Mechanisms and Treatment.” Alcohol Health and Research World 21 (1): 13–20. Alcohol | 903 ©2006 The International Bank for Reconstruction and Development / The World Bank 105 ARRIVE ALIVE Campaign. 2000. “ARRIVE ALIVE Safety Campaign.” Pretoria: Department of Transport. http://www.transport.gov.za/ projects/arrive/presentation/slide41.htm. Babor, T., F. Caetano, S. Casswell, G. Edwards, N. Giesbrecht, K. Graham, and others. 2003. Alcohol: No Ordinary Commodity—A Consumer’s Guide to Public Policy. Oxford, U.K.: Oxford University Press. Babor, T. F., and M. Grant, eds. 1992. Project on Identification and Management of Alcohol-Related Problems: Report on Phase II—A Randomized Clinical Trial of Brief Interventions in Primary Health Care. Geneva: World Health Organization, Program on Substance Abuse. Babor, T. F., M. Grant, W. Acuda, F. H. Burns, C. Campillo, F. K. Del Boco, and others. 1994. “Randomized Clinical Trial of Brief Interventions in Primary Health Care: Summary of a WHO Project (with Commentaries and a Response).” Addiction 89: 657–78. Bruun, K., G. Edwards, M. Lumio, K. Mäkelä, L. Pan, R. E. Popham, and others. 1975. Alcohol Control Policies in Public Health Perspective. Helsinki: Finnish Foundation for Alcohol Studies. Carlini-Cotrim, B., and A. A. da M. Chasin. 2000. “Blood Alcohol Content and Death from Fatal Injury: A Study in the Metropolitan Area of São Paulo, Brazil.” Journal of Psychoactive Drugs 32: 269–75. Carlini-Cotrim, B., I. Pinsky, and M. T. Serrano Barbosa. 1998. Alcohol Availability Restrictions in Developing Societies: The Case of São Paulo Highways, Brazil. Report prepared for the WHO Alcohol Policy in Developing Societies Project, World Health Organization, Geneva. Catalyst Health Consultants. 2001. Alcohol Misuse in Scotland: Trends and Costs—Final Report. Northwort, U.K.: Catalyst Health Consultants. http://www.scotland.gov.uk/health/alcoholproblems/docs/trco.pdf. Cerff, P., and A. Plüddemann. 1998. “Brief on the ARRIVE ALIVE Campaign.” Prepared for the Alcohol Policy in Developing Societies project. Cape Town, South Africa: Medical Research Council, Urbanisation, and Health Research Programme. Chisholm, D., J. Rehm, M. van Ommeren, and M. Monteiro. 2004. “Reducing the Global Burden of Hazardous Alcohol Use: A Comparative Cost-Effectiveness Analysis.” Journal of Studies on Alcohol 65 (6): 782–93. “Selected Major Risk Factors and Global and Regional Burden of Disease.” Lancet 360: 1347–60. Farrell, S., W. G. Manning, and M. D. Finch. 2003. “Alcohol Dependence and the Price of Alcoholic Beverages.” Journal of Health Economics 22 (1): 117–47. Fleming, M. F., M. P. Mundt, M. T. French, L. B. Manwell, E. A. Stauffacher, K. L. Barry, and others. 2000. “Benefit-Cost Analysis of Brief Physician Advice with Problem Drinkers in Primary Care Settings.” Medical Care 38: 7–18. Foxcroft, D. R., D. Ireland, D. J. Lister-Sharp, G. Lowe, and R. Breen. 2003. “Longer-Term Primary Prevention for Alcohol Misuse in Young People: A Systematic Review.” Addiction 98: 397–411. Foxcroft, D. R., D. Lister-Sharp, and G. Lowe. 1997. “Alcohol Misuse Prevention for Young People: A Systematic Review Reveals Methodological Concerns and Lack of Reliable Evidence of Effectiveness.” Addiction 92: 531–38. Gmel, G., E. Gutjahr, and J. Rehm. 2003. “How Stable Is the Risk Curve between Alcohol and All-Cause Mortality and What Factors Influence the Shape? A Precision-Weighted Hierarchical Meta-Analysis.” European Journal of Epidemiology 18 (7): 631–42. Gmel, G., J. Rehm, and U. Frick. 2003. “Trinkmuster, Pro-Kopf-Konsum von Alkohol und koronare Mortalität.” Sucht 49 (2): 95–104. Grant, B. F., D. A. Dawson, F. S. Stinson, S. P. Chou, M. C. Dufour, R. P. Pickering, and others. 2004. “The 12-Month Prevalence and Trends in DSM-IV Alcohol Abuse and Dependence: United States, 1991–1992 and 2001–2002.” Drug and Alcohol Dependence 11 (3): 223–34. Grube, J., and G. Agostinelli. 2000. Alcohol Advertising and Alcohol Consumption: A Review of Recent Research. Berkeley, CA: Prevention Research Center. Gutjahr, E., G. Gmel, and J. Rehm. 2001. “Relation between Average Alcohol Consumption and Disease: An Overview.” European Addiction Research 7 (3): 117–27. Harrison, L., and E. Gardiner. 1999. “Do the Rich Really Die Young? Alcohol-Related Mortality and Social Class in Great Britain, 1988–94.” Addiction 94: 1871–80. Corrao, G., L. Rubbiati, V. Bagnardi, A. Zambon, and K. Poikolainen. 2000. “Alcohol and Coronary Heart Disease: A Meta-Analysis.” Addiction 94 (10):1501–23. Higgins-Biddle, J. C., and T. F. Babor. 1996. Reducing Risky Drinking. Report prepared for the Robert Wood Johnson Foundation, University of Connecticut Health Center, Farmington. Demyttenaere, K., R. Bruffaerts, J. Posada-Villa, I. Gasquet, V. Kovess, J. P. Lepine, and others. 2004. “Prevalence, Severity, and Unmet Need for Treatment of Mental Disorders in the World Health Organization World Mental Health Surveys.” Journal of the American Medical Association 291 (21): 2581–90. Hinkle, L. E., and A. Herrou-Aragon. 2001. “How Far Did Africa’s First Generation Trade Reforms Go?” World Bank, Washington, DC. http://www.uesiglo21.edu.ar/pdfs%20dpto%20economia/ES?002? ECO.pdf. Drummond, D. C. 1990. “The Relationship between Alcohol Dependence and Alcohol-Related Problems in a Clinical Population.” British Journal of Addiction 85 (3): 357–66. Eckardt, M. J., S. E. File, G. L. Gessa, K. A. Grant, C. Guerri, P. L. Hoffman, and others. 1998. “Effects of Moderate Alcohol Consumption in the Central Nervous System.” Alcoholism: Clinical and Experimental Research 22 (5): 998–1040. Edwards, G., P. Anderson, T. F. Babor, S. Casswell, R. Ferrence, N. Giesbrecht, and others. 1994. Alcohol Policy and the Public Good. Oxford, U.K.: Oxford University Press. Emanuele, N., and M. A. Emanuele. 1997. “The Endocrine System: Alcohol Alters Critical Hormonal Balance.” Alcohol Health and Research World 21 (1): 53–64. English, D. R., C. D. J. Holman, E. Milne, M. Winter, G. K. Hulse, G. Codde, and others. 1995. The Quantification of Drug Caused Morbidity and Morality in Australia. Canberra: Commonwealth Department of Human Services and Health. Ezzati, M., A. D. Lopez, A. Rodgers, S. Vander Horn, C. J. L. Murray, and the Comparative Risk Assessment Collaborating Group. 2002. Isichei, H. U., P. U. Ikwuagu, and J. O. Egbuta. 1993. “Pattern of Alcoholism in Jos, Nigeria, and Castrop-Rauxel, West Germany: A Comparative Study.” In Epidemiology and Control of Substance Use in Nigeria, ed. I. S. Obot, 123–27. Jos, Nigeria: Centre for Research and Information on Substance Abuse. Johns, B., R. Baltussen, T. Adam, and R. Hutubessy. 2003. “Programme Costs in the Economic Evaluation of Health Interventions.” Cost-Effectiveness and Resource Allocation 1: 1. http://www .resourceallocation.com. Klingemann, H., and G. Gmel, eds. 2001. Mapping the Social Consequences of Alcohol Consumption. Dordrecht, Netherlands: Kluwer Academic Publishers. Lauer, J. A., C. J. L. Murray, K. Roehrich, and H. Wirth. 2003. “PopMod: A Longitudinal Population Model with Two Interacting Disease States.” Cost Effectiveness and Resource Allocation 1: 6. http://www .resourceallocation.com. Lee, V. J. 2001. Tourism and Alcohol in the Developing World: Potential Effects on Alcohol Policies and Local Drinking Problems. Stockholm: Stockholm University, Centre for Social Research on Alcohol and Drugs. 904 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 106 Leppänen, K. 1979. “Valtakunnallisen lauantaisulkemiskokoeilun vaikutuksista alkoholijuomien myyntiin” (Effects of National Saturday Closing Experiment on Alcohol Sales). Alkoholipolitikka 44: 20–21. Levy, F., and S. I. Ornstein. 1983. “Price and Income Elasticities and the Demand for Alcoholic Beverages.” In Recent Developments in Alcoholism, ed. M. Galanter, 303–45. New York: Plenum. Liu, B. Q., R. Peto, Z. M. Chen, J. Boreham, Y. P. Wu, J. Y. Li, and others. 1998. “Emerging Tobacco Hazards in China: 1. Retrospective Proportional Mortality Study of One Million Deaths.” British Medical Journal 317 (7170): 1411–22. Ludbrook, A, C. Godfrey, L. Wyness, S. Parrot, S. Haw, M. Napper, and others. 2002. “Effective and Cost-Effective Measures to Reduce Alcohol Misuse in Scotland: A Literature Review.” Aberdeen, U.K.: Health Economics Research Unit. Mäkelä, P. 1999. “Alcohol-Related Mortality as a Function of SocioEconomic Status.” Addiction 94: 867–86. Mathers, C. D., C. Bernard, K. Moesgaard Iburg, M. Inoue, D. Ma Fat, K. Shibuya, and others. 2003. “Global Burden of Disease in 2002: Data Sources, Methods and Results.” Global Programme on Evidence for Health Policy Discussion Paper 54, Geneva, World Health Organization. McKee, M., and A. Britton. 1998. “The Positive Relationship between Alcohol and Heart Disease in Eastern Europe: Potential Physiological Mechanisms.” Journal Royal Society Medicine 91: 402–7. Moyer, A., J. W. Finney, C. E. Swearingen, and P. Vergun. 2002. “Brief Interventions for Alcohol Problems: A Meta-Analytic Review of Controlled Investigations in Treatment-Seeking and Non-TreatmentSeeking Populations.” Addiction 97: 279–92. National Highway Traffic Safety Administration. 1992. “Driving under the Influence: A Report to Congress on Alcohol Limits.” Washington, DC: U.S. Department of Transportation. Nordlund, S. 1984. “Effekten av lørdagsstengningen ved Vinmonopolets butikker” (Effects of Saturday Closing of the Wine/Liquor Monopoly Outlets). Alkoholpolitik—Tidsskrift for Nordisk Alkoholforskning 1: 221–29. Norström, T., and O. J. Skog. 2003. “Saturday Opening of Alcohol Retail Shops in Sweden: An Impact Analysis.” Journal of Studies on Alcohol 64: 393–401. Peek-Asa, C. 1999. “The Effect of Random Alcohol Screening in Reducing Motor Vehicle Crash Injuries.” American Journal of Preventive Medicine 16 (1 Suppl.): 57–67. Puddey, I. B., V. Rakic, S. B. Dimmitt, and L. J. Beilin. 1999. “Influence of Drinking on Cardiovascular Disease and Cardiovascular Risk Factors—A Review.” Addiction 94: 649–63. Rehm, J., G. Gmel, C. T. Sempos, and M. Trevisan. 2003. “Alcohol-Related Mortality and Morbidity.” Alcohol Research and Health 27 (1): 39–51. Rehm, J., E. Gutjahr, and G. Gmel. 2001. “Alcohol and All-Cause Mortality: A Pooled Analysis.” Contemporary Drug Problems 28: 337–61. Rehm, J., N. Rehn, R. Room, M. Monteiro, G. Gmel, D. Jernigan, and U. Frick. 2003. “The Global Distribution of Average Volume of Alcohol Consumption and Patterns of Drinking.” European Addiction Research 9 (4): 147–56. Rehm, J., R. Room, K. Graham, M. Monteiro, G. Gmel, and C. T. Sempos. 2003. “The Relationship of Average Volume of Alcohol Consumption and Patterns of Drinking to Burden of Disease: An Overview.” Addiction 98 (10): 1209–28. Rehm, J., R. Room, M. Monteiro, G. Gmel, K. Graham, N. Rehn, and others. 2003. “Alcohol as a Risk Factor for Global Burden of Disease.” European Addiction Research 9 (4):157–64. Rehm, J., R. Room, M. Monteiro, G. Gmel, K. Graham, T. Rehn, and others. 2004. “Alcohol.” In Comparative Quantification of Health Risks: Global and Regional Burden of Disease Due to Selected Major Risk Factors, ed. M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Murray, 959–1108. Geneva: World Health Organization. Rehm, J., C. T. Sempos, and M. Trevisan. 2003. “Average Volume of Alcohol Consumption, Patterns of Drinking, and Risk of Coronary Heart Disease: A Review.” Journal of Cardiovascular Risk 10 (1): 15–20. Ridolfo, B., and C. Stevenson. 2001. The Quantification of Drug-Caused Mortality and Morbidity in Australia 1998. Canberra: Australian Institute of Health and Welfare. Room, R., T. Babor, and J. Rehm. 2005. “Alcohol and Public Health: A Review.” Lancet 365 (February 5): 519–30. Room, R., K. Graham, J. Rehm, D. Jernigan, and M. Monteiro. 2003. “Drinking and Its Burden in a Global Perspective: Policy Considerations and Options.” European Addiction Research 9 (4): 165–75. Room, R., A. Janca, L. A. Bennett, L. Schmidt, N. Sartorius, and others. 1996. “WHO Cross-Cultural Applicability Research on Diagnosis and Assessment of Substance Use Disorders: An Overview of Methods and Selected Results.” Addiction 91 (2): 199–220. Room, R., D. Jernigan, B. Carlini-Marlatt, O. Gureje, K. Mäkelä, M. Marshall, and others. 2002. Alcohol in Developing Societies: A Public Health Approach. Helsinki: Finnish Foundation for Alcohol Studies. Saffer, H. 2000. “Alcohol Consumption and Alcohol Advertising Bans.” NBER Working Paper 7758, National Bureau of Economic Research, Cambridge, MA. Saffer, H., and D. Dave. 2002. “Alcohol Consumption and Alcohol Advertising Bans.” Applied Economics 34 (11): 1325–34. Saunders, J. B., and O. G. Aasland. 1987. WHO Collaborative Project on Identification and Treatment of Persons with Harmful Alcohol Consumption. WHO/MNH/DAT 86.3. Geneva: World Health Organization. Schmidt, L., and R. Room. 1999. “Cross-Cultural Applicability in International Classifications and Research on Alcohol Dependence.” Journal of Studies on Alcohol 60: 448–62. Shults, R. A, R. W. Elder, D. A. Sleet, J. L. Nichols, M. O. Alao, V. G. Carande-Kulis, and others. 2001. “Reviews of Evidence Regarding Interventions to Reduce Alcohol-Impaired Driving.” American Journal of Preventive Medicine 21 (Suppl. 4): 66–88. Single, E., L. Robson, J. Rehm, and X. Xie. 1999. “Morbidity and Mortality Attributable to Alcohol, Tobacco, and Illicit Drug Use in Canada.” American Journal of Public Health 89 (3): 385–90. Single, E., L. Robson, X. Xie, and J. Rehm. 1996. The Cost of Substance Abuse in Canada. Ottawa: Canadian Centre on Substance Abuse. Sobell, L. C., T. P. Ellingstad, and M. B. Sobell. 2000. “Natural Recovery from Alcohol and Drug Problems: Methodological Review of the Research with Suggestions for Future Directions.” Addiction 95: 749–64. Stouthard, M. E., M. L. Essink-Bot, and G. L. Bonsel. 2000. “On Behalf of the Dutch Disability Weights Group: Disability Weights for Diseases— A Modified Protocol and Results for a Western European Region.” European Journal of Public Health 10: 24–30. Tax Policy Chief Directorate. 2002. The Taxation of Alcoholic Beverages in South Africa and Its Impact on the Consumption Levels of Alcoholic Beverages. Pretoria: National Treasury. U.S. Department of State. 1999. “Mauritius: 1999–2000 Budget Increased.” U.S. Department of State, Washington, DC. http://www.tradeport.org/ ts/countries/mauritius/mrr/mark0003.html. Wallace, S., and R. Bird. 2003. “Taxing Alcohol in Africa: Reflections from International Experience.” Paper presented at the South Africa Conference on Excise Taxation, June 11–13, Gauteng, South Africa. http://www.iticnet.org/030707-PRESENTATION_Bird.pdf. Whitlock, E. P., M. R. Polen, C. A. Green, T. Orleans, and J. Klein. 2004. “Behavioral Counseling Interventions in Primary Care to Reduce Alcohol | 905 ©2006 The International Bank for Reconstruction and Development / The World Bank 107 Risky/Harmful Alcohol Use by Adults: A Summary of the Evidence for the U.S. Preventive Services Task Force.” Annals of Internal Medicine 140: 557–68. WHO (World Health Organization). 1993. The ICD-10 Classification of Mental and Behavioral Disorders: Diagnostic Criteria for Research. Geneva: WHO. ———. 1999. Global Status Report on Alcohol. WHO/HSC/SAB/99.11. Geneva: WHO, Substance Abuse Department. ———. 2000. 1997–1999 World Health Statistics Annual. Geneva: WHO. http://www.who.int/whosis/. ———. 2003a. “Burden of Disease Project.” WHO, Geneva. http:// www.who.int/evidence/bod. ———. 2003b. “WHO Global Alcohol Database.” WHO, Geneva. http://www3.who.int/whosis/alcohol. WHO Brief Intervention Study Group. 1996. “A Cross-National Trial of Brief Interventions with High Risk Drinkers.” American Journal of Public Health 86: 948–55. Zakhari, S. 1997. “Alcohol and the Cardiovascular System: Molecular Mechanisms for Beneficial and Harmful Action.” Alcohol Health and Research World 21 (1): 21–29. ———. 2002. World Health Report 2002: Reducing Risks, Promoting Healthy Life. Geneva: WHO. 906 | Disease Control Priorities in Developing Countries | Jürgen Rehm, Dan Chisholm, Robin Room, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 108 Chapter 48 Illicit Opiate Abuse Wayne Hall, Chris Doran, Louisa Degenhardt, and Donald Shepard Illicit drugs are those banned by international drug control treaties. They include cannabis products (for example, marijuana, hashish, and bhang); stimulant drugs (such as cocaine and methamphetamine); so-called dance-party drugs (such as 3, 4-methylenedioxymethamphetamine, also known as ecstasy or MDMA); and illicit opioids (for instance, heroin and opium) and diverted pharmaceutical opioids (such as buprenorphine, methadone, and morphine) (see annex 48.A). Worldwide, 185 million people were estimated to have used illicit drugs during 1998–2002 (UNODC 2004; UNODCCP 2002). Cannabis was the most widely used illicit drug, with 146.2 million users in 2002, or 3.7 percent of the global population over age 15. The stimulant drugs were the next most widely used illicit drugs: 29.6 million people worldwide used amphetamines; 13.3 million used cocaine; and 8.3 million used ecstasy. An estimated 15.3 million, or 0.4 percent of the world population age 15 to 64, used illicit opioids; more than half used heroin and the remainder used opium or diverted pharmaceutical opioids. Illicit opioids continue to be the major illicit drug problem in most regions of the world in terms of impact on public health and public order (UNODC 2004). Even though cannabis use accounts for about 80 percent of illicit drug use worldwide, the mortality and morbidity attributable to its use are not well understood, even in developed countries (W. Hall and Pacula 2003; Macleod and others 2004; WHO Programme on Substance Abuse 1997). The same is true of the morbidity and mortality attributable to cocaine and amphetamine-type stimulants (Macleod and others 2004). Dance-party drugs have been used for too short a time in most developed societies to enable a good assessment of their potential for harm (Boot, McGregor, and Hall 2000; Macleod and others 2004). The remainder of this chapter is concerned with disease control priorities for illicit opioid dependence, because dependent users account for most of the illicit opioids consumed and experience most of the harm such dependence causes (W. Hall, Degenhardt, and Lynskey 1999). NATURE, CAUSES, AND HEALTH CONSEQUENCES OF ILLICIT OPIOID USE Before considering interventions, we briefly summarize what is known about the antecedents, causes, and health consequences of illicit opioid use. Antecedents of Heroin Use Law enforcement efforts to reduce the availability of heroin aim to increase its price, deter illicit drug use, and promote social values that discourage heroin use (Fergusson, Horwood, and Lynskey 1998; Hawkins, Catalano, and Miller 1992; Newcomb and Bentler 1988). These gains may be at the cost of increasing harm among the minority who use opioids despite the prohibition—for example, by encouraging injecting use as the most efficient way to use an expensive drug and increasing needle sharing because clean injecting equipment is not freely available (Rhodes and others 2003; Strathdee and others 2003). Two aspects of the family environment are associated with increased rates of both licit and illicit drug use in young people in developed countries. The first is exposure to a disadvantaged home environment, with parental conflict and poor discipline and supervision; the second is exposure to parents’ and siblings’ use of alcohol and other drugs (Hawkins, Catalano, and Miller 1992). In developed countries, children who perform 907 ©2006 The International Bank for Reconstruction and Development / The World Bank 109 poorly in school because of impulsive or problem behavior and those who are early users of alcohol and other drugs are most likely to use illicit opioids (Fergusson, Horwood, and SwainCampbell 2002). Affiliation with drug-using peers is a risk factor for drug use that operates independently of individual and family risk factors (Fergusson, Horwood, and Lynskey 1998; Hawkins, Catalano, and Miller 1992). Health Consequences of Heroin Use The following sections describe the major health consequences of heroin use. They include dependence, increased mortality and morbidity attributable to drug overdoses, and bloodborne viruses. Heroin Dependence. In household surveys, 1 to 2 percent of adults in Australia, the United States, and Europe report using heroin at some time in their lives (Australian Institute of Health and Welfare 1999; EMCDDA 2002; SAMHSA 2002). The highest rates are typically among adults age 20 to 29. Selfreported heroin use in population surveys probably underestimates rates of use because heroin users are undersampled and those who are sampled underreport their use (W. Hall, Lynskey, and Degenhardt 1999). In developed countries, one in four of those who report heroin use become dependent on it (Anthony, Warner, and Kessler 1994). People who are heroin dependent continue to use heroin in the face of problems that they know (or believe) to be caused by its use. These problems include being arrested or imprisoned, having interpersonal and family problems, catching infectious diseases, and suffering from drug overdoses. Many heroin users who seek treatment have typically been daily heroin injectors, although in Europe (EMCDDA 2002), North America (Office of National Drug Control Policy 2001), and parts of Asia, illicit opioid users also smoke or “chase” the drug (inhale the fumes released when heroin is heated) (UNODC 2004). The American Psychiatric Association defines drug dependence as “a cluster of cognitive, behavioral, and physiologic symptoms indicating that the individual continues use of the substance despite significant substance-related problems” (American Psychiatric Association 1994, 176). In the fourth edition of the association’s Diagnostic and Statistical Manual of Mental Disorders (1994,), a diagnosis of substance dependence requires that three or more of the following occur together: At any time in the same 12-month period: 1. tolerance, as defined by either of the following: a. need for markedly increased amounts of the substance to achieve intoxication or desired effect b. markedly diminished effect with continued use of the same amount of the substance; 2. withdrawal, as manifested by either of the following: a. the characteristic withdrawal syndrome for the substance b. the same (or closely related) substance is taken to relieve or avoid withdrawal symptoms; 3. the substance is often taken in larger amounts or over a longer period than was intended; 4. there is a persistent desire or unsuccessful efforts to cut down or control substance use; 5. a great deal of time is spent in activities necessary to obtain the substance (e.g., visiting multiple doctors, driving long distances), use the substance (e.g., chain smoking), or recover from its effects; 6. important social, occupational, or recreational activities are given up or reduced because of substance use; 7. the substance use is continued despite knowledge of having a persistent or recurrent physical or psychological problem that is likely to have been caused or exacerbated by the substance. Indirect estimation methods suggest that in Australia, the United Kingdom, and the European Union fewer than 1 percent of adults age 15 to 54 are heroin dependent (EMCDDA 2002; W. Hall and others 2000). Research in the United States indicates that dependent heroin users who seek treatment or who come to the attention of the legal system may use heroin for decades (Goldstein and Herrera 1995; Hser, Anglin, and Powers 1993), with periods of use punctuated by abstinence (Bruneau and others 2004; Galai and others 2003), drug treatment, and imprisonment (Gerstein and Harwood 1990). When periods of abstinence are included, dependent heroin users use heroin daily for 40 to 60 percent of the 20 years that they typically are addicts (Ball, Shaffer, and Nurco 1983; Maddux and Desmond 1992). Illicit opioid use increased in Asia, Europe, and Oceania and, to a lesser extent, in Africa and South America in the 1990s, but it has stabilized or declined since 2000 (UNODC 2004). Most illicit opioid users (7.8 million) live in Asian countries that surround the major opium-producing countries, Afghanistan and Myanmar. Europe accounts for about 25 percent of illicit opioid use (4 million users or 0.8 percent of the adult population age 15 to 64). Two-thirds of users are in Eastern Europe, which reported large increases in illicit opioid use during the second half of the 1990s (Atlani and others 2000; Hamers and Downs 2003; Kelly and Amirkhanian 2003; Rhodes and others 1999; Uuskula and others 2002). Illicit opioid use stabilized in much of Asia between 2000 and 2002 (UNODC 2004) as a result of decreased opium production after the rapid expansion during the 1990s (Dorabjee and Samson 2000; Reid and Crofts 2000). After 2000, India and Pakistan reported stabilizing rates of illicit opioid use but increased injection of pharmaceutical opiates (Ahmed and others 2003; Dorabjee and Samson 2000; Strathdee and others 908 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 110 2003). China has reported a steady rate of growth in illicit opiate use in its southern and northern provinces (Beyrer 2003; Beyrer and others 2000; Yu and others 1998) and a 15-fold increase in the number of registered opioid addicts between 1990 and 2002, bringing the total to about 1 million (UNODC 2004). Oceania experienced a marked rise in heroin use in the late 1990s, largely driven by a dramatic increase in the availability of heroin in Australia (Darke, Topp, and others 2002; W. Hall, Degenhardt, and Lynskey 1999). In late 2000, an abrupt heroin shortage resulted in a large reduction in fatal and nonfatal overdoses (Day and others 2004; Degenhardt, Day, and Hall 2004). account for 25 to 33 percent of deaths of young adult males (EMCDDA 2002). Mortality, Morbidity, and Heroin Dependence. In developed countries, dependent heroin users have an increased risk of premature death from drug overdoses, violence, suicide, and alcohol-related causes (Darke and Ross 2002; Goldstein and Herrera 1995; Vlahov and others 2004). Heroin users treated before the HIV epidemic were 13 times more likely to die prematurely than their peers (Hulse and others 1999), with opioid overdose the most frequent cause of death (W. Hall, Degenhardt, and Lynskey 1999). In countries with a high prevalence of HIV infection, AIDS is a major cause of premature death among drug users (EMCDDA 2002; UNAIDS and WHO 2002). Fatal opioid overdose deaths increased in many developed countries during the 1990s before declining after 2000 (UNODC 2004). In parts of Asia, Eastern Europe, and the United States, the sharing of contaminated injecting equipment accounts for a substantial proportion of new HIV infections (EMCDDA 2002; UNAIDS and WHO 2002; UNODC 2004). Injecting opioid use has been a major driver of HIV epidemics in China (Yu and others 1998), Myanmar (Beyrer and others 2000), the Russian Federation and former Soviet republics (Hamers and Downs 2003), and Vietnam (Beyrer and others 2000; Hien and others 2001). The prevalence of infection with hepatitis B and C viruses among injecting drug users is greater than 60 percent in Australia (National Centre in HIV Epidemiology and Clinical Research 1998), Canada (Fischer and others 2004), China (Ruan and others 2004), the United States (Fuller and others 2004), and the European Union (EMCDDA 2002). Chronic infection occurs in 75 percent of infections, and 3 to 11 percent of chronic hepatitis C virus carriers develop liver cirrhosis within 20 years (Hepatitis C Virus Projections Working Group 1998). Heroin-related deaths primarily occur among young adults and account for a large number of life years lost in developed societies. In Australia in 1996, for example, such deaths accounted for 2.2 percent of life years lost, with each death accounting for 22 years of life lost (Mathers, Vos, and Stephenson 1999). In Scotland and Spain, opiate-related deaths Degenhardt, Hall, and others (2004) estimate the contribution of illicit opioid dependence to the global burden of disease using data on deaths caused by opioid and other drug overdoses, suicides and accidents, and HIV/AIDS. When estimates of morbidity attributable to illicit drug use were added in, illicit opioid use accounted for 0.7 percent of global disabilityadjusted life years (DALYs) in 2000 (WHO 2003). These estimates suggest that illicit opioid use is a significant global cause of premature mortality and disability among young adults. Even so, they probably underestimate the disease burden attributable to illicit opioids, because they omit differences across subregions in the quality of data on causes of mortality and estimates of mortality and morbidity attributable to hepatitis and violence (Degenhardt, Hall, and others 2004). Economic Costs of Illicit Opioid Use. In Canada, Xie and others (1996) calculate the costs of illicit drugs as 0.2 percent of gross domestic product (GDP). In Australia, Collins and Lapsley (1996) estimate the economic costs of illicit drug abuse at 2 percent of GDP. CONTRIBUTION OF OPIOID DEPENDENCE TO THE GLOBAL BURDEN OF DISEASE INTERVENTIONS FOR ILLICIT OPIOID DEPENDENCE Methods adopted to control the problems arising from illicit opioid dependence include source-country control; interdiction of supply into end-use countries; enforcement by the police force and the criminal justice system of legal prohibitions on the supply, possession, and use of opioids; treatment of those who are opioid dependent, both voluntarily and under legal coercion from the criminal justice system; school-based and mass media preventive educational programs; and regulatory policies restricting the prescription of opioids (Manski, Pepper, and Petrie 2001). Prevention of Heroin Use Countries use a variety of interventions in attempts to prevent the initiation of use of illicit drugs such as cannabis (Manski, Pepper, and Petrie 2001; Spooner and Hall 2002), in the belief that early initiation of cannabis use leads to an increased risk of using illicit opioids (Fergusson, Horwood, and Swain-Campbell 2002). These interventions include legal prohibitions on the manufacture, sale, and use of opioid drugs Illicit Opiate Abuse | 909 ©2006 The International Bank for Reconstruction and Development / The World Bank 111 for nonmedical purposes; enforcement of these sanctions by law enforcement officials by means of fines and imprisonment; and enforcement of restrictions on medically prescribed opioids to prevent their diversion (Manski, Pepper, and Petrie 2001). Preventive measures also include mass media and school-based educational campaigns about the health risks of opioid and other illicit drug use (Spooner and Hall 2002). It is unclear how effective these interventions are in preventing cannabis use and even less clear whether they reduce the initiation of opioids (Caulkins and others 1999; Manski, Pepper, and Petrie 2001). The most popular interventions against illicit opioid use in many developed societies have been the interdiction of drug supply and the enforcement of legal sanctions against the possession, use, and sale of opioid drugs (Manski, Pepper, and Petrie 2001). As a consequence, imprisonment is the most common intervention to which many illicit opioid users have been exposed (Gerstein and Harwood 1990). In Asia and Eastern Europe, high rates of imprisonment of drug users have been a factor in HIV transmission, because drug users engage in highrisk injecting while imprisoned (Beyrer and others 2000). Interventions to Reduce Heroin-Related Harm The most effective intervention to reduce bloodborne virus infection arising from illicit injecting of opioids and other drugs is the provision of clean injecting equipment to reduce users’ risks of contracting or transmitting bloodborne viruses. This intervention has been widely supported in most developed countries, but it has been incompletely adopted in developing countries that have problems with the concept of facilitating the injection of illicit drugs (UNAIDS and WHO 2002). Vaccinations are available against hepatitis B but not hepatitis C. These important interventions are covered in chapter 18. A number of strategies can potentially reduce deaths from opioid overdoses (Darke and Hall 2003; Sporer 2003). First, injecting drug users can be educated about the dangers of combining the use of opioids with alcohol and benzodiazepines (McGregor and others 2001), both of which heighten the risk of a fatal opioid overdose (Darke and Zador 1996; WarnerSmith and others 2001). Heroin users also need to be discouraged from injecting in the streets or alone, thereby denying themselves assistance in the event of an overdose. These interventions have yet to be evaluated. A second strategy is to encourage drug users who witness overdoses to seek medical assistance and to use simple resuscitation techniques until help arrives. A more controversial option is to distribute the opioid antagonist naloxone to highrisk heroin users (Darke and Hall 1997; Strang and others 1996). Neither of these interventions has been evaluated. A third strategy is to provide supervised injecting facilities in areas with high rates of injecting opioid use (Dolan and others 2000; Kimber and others 2003). Supervised injecting facilities have been introduced in Germany, the Netherlands, and Switzerland (Dolan and others 2000; Kimber and others 2003), but their effect on overdose deaths has not been rigorously evaluated to date. A supervised injecting facility was evaluated in Australia, but the evaluation was limited by the concurrent onset of a heroin shortage that resulted in a 40 percent decline in overdose deaths (Kaldor and others 2003). A fourth strategy is to increase methadone maintenance among older, high-risk opioid-dependent people, because individuals enrolled in methadone maintenance treatment (MMT) are substantially less likely to suffer from a fatal overdose (Caplehorn and others 1994; Gearing and Schweitzer 1974; Langendam and others 2001). Treatment Interventions for Dependent Opioid Users The range of treatment interventions includes voluntary programs such as detoxification, abstinence-oriented treatments, and oral Methadone maintenance treatment, as well as involuntary options imposed by criminal justice systems. Detoxification. Detoxification is supervised withdrawal from a drug of dependence that attempts to minimize withdrawal symptoms. It is not a treatment for heroin dependence; it provides a respite from opioid use and may be a prelude to abstinence-based treatment (Mattick and Hall 1996). Naltrexone is a longer-acting opiate antagonist than naloxone; it can be used to accelerate the opioid withdrawal process. Ultra-rapid opioid detoxification accelerates withdrawal by giving the patient naltrexone under general anesthetic. There is no evidence that accelerated withdrawal in itself reduces the high rate of relapse to heroin use in the absence of further treatment (W. Hall and Mattick 2000). Abstinence-Oriented Treatments. Abstinence-oriented treatments aim to achieve enduring abstinence from all opioid drugs by providing some type of intervention after withdrawal to reduce the high rate of relapse to opioids (Mattick and Hall 1996). The interventions may include social and psychological support only or such support supplemented by pharmacological methods. Residential treatment in therapeutic communities and outpatient drug counseling may entail encouraging patients to become involved in self-help groups such as Narcotics Anonymous. These approaches share a commitment to achieving abstinence from all opioids, using group and psychological interventions to help dependent heroin users remain abstinent. Therapeutic communities and drug counseling are usually provided through specialist addiction or mental health services. The former are residential, and the latter are provided on an outpatient basis. 910 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 112 No randomized controlled trials of therapeutic communities or outpatient drug counseling have been carried out. Observational studies in the United Kingdom (Gossop, Marsden, and Stewart 1998; Gossop and others 1997) and the United States (Hubbard and others 1989; Simpson and Sells 1982) have found that therapeutic communities and drug counseling were less successful than MMT in attracting and retaining dependent heroin users, but they substantially reduced heroin use and crime among those who remained in treatment for at least three months (Gerstein and Harwood 1990; Gossop, Marsden, and Stewart 1998; Gossop and others 1997). Some evidence indicated that therapeutic communities may be more effective if they are used in combination with legal coercion to ensure that heroin users are retained in treatment long enough to benefit from it (Gerstein and Harwood 1990). Recovering drug users run Narcotics Anonymous groups using an adaptation of the 12-step philosophy of Alcoholics Anonymous. Some individuals use these groups as their sole form of support for abstinence, whereas for others these groups complement therapeutic communities that are based on the same principles. Such groups are usually not open to people who are in opioid substitution treatment programs. The most extensive research on self-help has been in the treatment of alcohol dependence. Treated alcoholics who participate in Alcoholics Anonymous groups have higher rates of abstinence than those who do not (see, for example, Tonigan, Connors, and Miller 2003; Tonigan, Toscova, and Miller 1996). The good outcome in those who attend Alcoholics Anonymous meetings may reflect the self-selection of motivated participants into self-help groups. Recent studies that have attempted to control for this possibility using sophisticated statistical methods have produced mixed results, with some showing the persistence of an effect of self-help after correction (Tonigan, Connors, and Miller 2003) while others do not (Fortney and others 1998). Shepard and others (forthcoming) evaluate the effect of selfhelp participation on substance abuse 24 months after treatment for members of a mixed population of substance abusers treated at two treatment facilities in the United States, some of whom had problems with heroin. They find that participation in self-help groups was associated with longer abstinence from all drugs. Correction for self-selection did not eliminate the association in one treatment setting, but it made the results much more equivocal in the other. Oral Methadone Maintenance Treatment. This treatment substitutes a long-acting, orally administered opioid for the shorter-acting heroin, with the aim of stabilizing dependent heroin users so that they are amenable to rehabilitation (Marsh and others 1990; Ward, Hall, and Mattick 1998). When given in high or blockade doses, methadone blocks the euphoric effects of injected heroin, allowing the individual to take advantage of psychotherapeutic and rehabilitative services. Every one of the small number of randomized controlled trials of MMT compared with placebo or no treatment has produced positive results (W. Hall, Ward, and Mattick 1998; Mattick and others 2003). Large observational studies show that MMT decreases heroin use and criminal activity and reduces HIV transmission while patients remain in treatment (Gerstein and Harwood 1990; Simpson and Sells 1990; Ward, Hall, and Mattick 1998). MMT is the best-supported form of opioid maintenance treatment (Farre and others 2002; Marsch 1998; Mattick and others 2003). Buprenorphine is a mixed agonist-antagonist that also blocks the effects of heroin. When given in high doses, its effects can last for up to three days, while its antagonist effects substantially reduce overdose and abuse (Oliveto and Kosten 1997; Ward, Hall, and Mattick 1998). Meta-analyses have found that buprenorphine is effective in the treatment of heroin dependence (Mattick and others 2003) and is of equivalent efficacy to MMT when delivered in primary health care and specialist treatment settings in Australia (Gibson and others 2003). Bammer and others (2003) have proposed injectable heroin maintenance as a way of attracting into treatment those heroin users who are not interested in or have failed to respond to MMT. This method has recently been evaluated in the Netherlands (Central Committee on the Treatment of Heroin Addicts 2002) and Switzerland (Perneger and others 1998; Uchtenhagen, Gutzwiller, and Dobler-Mikola 1998). Perneger and others (1998) report a randomized controlled trial of injectable heroin maintenance in people who had failed at MMT. Stabilizing and safely maintaining heroin addicts on injectable heroin (self-administered on-site in a comprehensive health and social service) proved feasible for six months and substantially improved their health and social well-being. The Swiss trials showed that it was possible to maintain opioid addicts on injectable heroin for up to two years (Rehm and others 2001; Uchtenhagen, Gutzwiller, and Dobler-Mikola 1998). A recent randomized controlled trial in the Netherlands (Central Committee on the Treatment of Heroin Addicts 2002) confirms the findings of Perneger and others (1998). Criminal Justice Interventions for Dependent Illicit Opioid Users. The most common intervention for illicit opioid dependence in most developed societies is imprisonment (EMCDDA 2003; Gerstein and Harwood 1990). Imprisonment is not intended to be a health intervention. Nonetheless, it is an ineffective way of reducing opioid dependence, when judged by the high recidivism in longitudinal studies of dependent heroin users (see, for example, Hser, Anglin, and Powers 1993; Manski, Pepper, and Petrie 2001). Legally coerced treatment is treatment that is legally forced on those who have been charged with or convicted of an Illicit Opiate Abuse | 911 ©2006 The International Bank for Reconstruction and Development / The World Bank 113 offense to which their drug dependence has contributed (W. Hall 1997). It is most often provided as an alternative to imprisonment, under the threat of imprisonment if the person fails to comply with the treatment (W. Hall 1997; Manski, Pepper, and Petrie 2001; Spooner, Hall, and Mattick 2001). Its major justification is that it is an effective way of treating offenders’ drug dependence that reduces the likelihood of their offending again (Gerstein and Harwood 1990). A consensus view prepared for the World Health Organization (WHO) (Porter, Arif, and Curran 1986) was that compulsory treatment was legally and ethically justified only if the rights of the individuals were protected by due process and if the treatment provided was effective and humane. Research into the effectiveness of legally coerced treatment for opioid dependence has been limited to observational studies (W. Hall 1997; Manski, Pepper, and Petrie 2001; Wild, Roberts, and Cooper 2002). Anglin’s (1988) quasi-experimental studies of the California Civil Addict Program provide the strongest evidence of efficacy. These studies compared heroin-dependent offenders who entered the program between 1962 and 1964 with a group of similar offenders who went through the criminal justice system during the same period. They found that compulsory hospital treatment followed by close supervision in the community substantially reduced heroin use and crime. The effectiveness of less coercive forms of treatment has been supported by analyses of the effectiveness of communitybased treatment provided while on probation or parole (Hubbard and others 1989; Simpson and others 1986). These studies showed that individuals who entered community-based therapeutic communities and drug-free outpatient counseling under legal pressure did as well as those who did so voluntarily (Hubbard and others 1988; Simpson and Friend 1988). The recent creation of specialized drug courts in the United States to process those arrested for drug-related offenses awaits rigorous evaluation (Belenko 2002; Manski, Pepper, and Petrie 2001). Legally coerced MMT is also effective. The strongest evidence comes from a study in which drug offenders were randomly assigned to parole with and without community-based MMT (Dole and others 1969). This study showed a greater reduction in heroin use and lower rates of incarceration among those enrolled in MMT in the year following their release from prison. These findings are supported by observational studies that found no major differences in response to MMT between those who enrolled under legal coercion and those who did not (Anglin, Brecht, and Maddahain 1989; Brecht, Anglin, and Wang 1993; Hubbard and others 1988). Economic Evaluations of Interventions for Illicit Opioid Dependence The few published economic evaluations of treatment interventions for illicit opioid dependence indicate varying levels of cost-effectiveness. Detoxification. The National Evaluation of Pharmacotherapies for Opioid Dependence Project in Australia conducted a cost-effectiveness analysis of five interventions: • naltrexone-induced rapid opioid detoxification under anesthesia • naltrexone-induced rapid opioid detoxification under sedation • conventional inpatient detoxification • conventional outpatient detoxification • buprenorphine outpatient detoxification. A successful outcome was defined as achieving abstinence from heroin for one week (Mattick and others 2001). Rapid detoxification under sedation was the most costeffective method of detoxification (US$2,355 for one week of abstinence) and conventional outpatient detoxification the least cost-effective (US$12,031). Rapid detoxification under anesthesia achieved high rates of abstinence in the first week, but its expense reduced its cost-effectiveness (Mattick and others 2001). Doran and others (2003) compared the cost-effectiveness of detoxification from heroin using buprenorphine in a specialist Australian clinic and in a shared care setting. They conducted a randomized controlled trial with 115 heroin-dependent patients receiving a five-day treatment regime of buprenorphine. The specialist clinic was a community-based treatment agency in Sydney. Shared care involved treatment by a general practitioner, supplemented by weekend dispensing and some counseling at the specialist clinic. They estimate that buprenorphine detoxification in the shared care setting was US$17 more expensive per patient than the costs of treatment at the clinic (US$236 per patient). Drug-Free Treatment. The limited economic evaluations of drug-free treatment have used data from observational studies of treatment outcomes in samples of patients who have mixed substance abuse problems that include opioids. For example, Shepard, Larson, and Hoffmann (1999) calculate a range of estimated costs for achieving an abstinent year in 408 patients at two different treatment facilities in the United States. The cost-effectiveness depended on the severity of the problem and the intensiveness and cost of the intervention. For outpatients with the least severe drug problems, the cost of an abstinent year was US$7,000, whereas the same outcome in patients with more severe problems who received long-term residential treatment cost US$20,000. Shepard and others (forthcoming) use these data to estimate the cost-effectiveness of involvement in mutual selfhelp groups, such as Alcoholics Anonymous and Narcotics Anonymous, in sustaining abstinence for up to 24 months after treatment. They find a positive association between self-help 912 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 114 involvement and abstinence 12 and 24 months after treatment. Applying statistical methods to correct for the effects of self-selection into self-help, they find that in a Veterans Administration hospital, the effects of self-help on abstinence persisted after the statistical correction, but at the other site, the results depended on the method of analysis that was used. They estimate the cost of achieving an abstinent year by means of self-help in the year following treatment at US$13,000, all of that due to the costs that participants incurred in attending a group. Oral Opioid Maintenance Treatment. Goldschmidt’s (1976) economic evaluation of MMT found that it was as effective as a therapeutic community intervention and twice as costeffective. Cartwright’s (2000) review of the literature since 1976 identified a number of studies, all of which reported positive benefit-cost ratios for MMT. Gerstein, Harwood, and Suter’s (1994) California Drug and Alcohol Treatment Assessment study is the most comprehensive cost-benefit analysis carried out to date. The authors examine the effects of treatment—residential programs, outpatient programs, and methadone programs—on alcohol and drug use, criminal activity, health and health care utilization, and source of income. For each treatment modality, they found that the benefits during the first year of treatment significantly exceeded the cost of delivering the care. The benefit-cost ratio was 4.8 for residential treatment and 11.0 and 12.6 for outpatients and discharged methadone participants, respectively. Doran and others (2003) compared the cost-effectiveness of buprenorphine and methadone treatment for opioid dependence. In a randomized controlled trial, 405 subjects were randomly assigned to each treatment at one of three specialist outpatient drug treatment centers. The study found that treatment with methadone was less expensive and more effective than treatment with buprenorphine, but the difference in cost (US$143 per additional heroin-free day gained) had a wide range of uncertainty around it (⫺US$1,469 to US$1,284). The National Evaluation of Pharmacotherapies for Opioid Dependence Project also provided a cost-effectiveness analysis of methadone, buprenorphine, LAAM (levo-alpha-acetylmethadol), and naltrexone maintenance treatments (Mattick and others 2001). The daily costs of these maintenance treatments were similar for methadone and LAAM, but naltrexone was slightly more expensive. Buprenorphine maintenance treatment (BMT) was more expensive, but its cost-efficiency could have been improved to make its cost similar to that for the other treatments. MMT was the most cost-effective treatment for opioid dependence because it achieved one of the highest rates of retention in treatment among the four pharmacotherapies examined. Naltrexone treatment was the least cost-effective. The costs of injectable heroin maintenance in the Dutch study was between US$18,015 and US$23,243 per patient per year (Bammer and others 2003). Most of the costs arose from the supervision of heroin use and the security required to prevent the diversion of heroin to the black market. Injectable heroin maintenance needs to produce substantially greater benefits for each participant than MMT to make it as costeffective as MMT. Economic Modeling of the Cost-Effectiveness of Opioid Maintenance Treatment. Barnett (1999), using data on the efficacy of MMT in reducing mortality derived from Gronbladh, Ohlund, and Gunne’s (1990) Swedish study and U.S. cost data, estimated that MMT saved an additional year of life at a cost of US$5,900. Barnett, Zaric, and Brandeau (2001), using a similar approach, estimated that the use of buprenorphine by patients who would not use methadone would cost less than US$45,000 per quality-adjusted life year. Overall, however, they found that BMT was much less effective and more costly than MMT. Zaric, Barnett, and Brandeau (2000) assessed the economic benefits of using MMT to reduce HIV transmission in heroin users. They found that for heroin users living in a community with a high prevalence of HIV infection, expanding MMT use produced an additional year of qualityadjusted life at a cost of US$8,200. Comparing the Cost-Effectiveness of Different Interventions Comparative cost-effectiveness analyses of these interventions face major obstacles because the small number of published studies used different methods to cost interventions and different endpoints to assess the outcome of treatment. The following list, therefore, only ranks treatment interventions in the approximate order of their cost-effectiveness. We believe that estimates of their likely contribution to DALYs worldwide would be too speculative. • Detoxification. Buprenorphine and supervised naltrexoneaccelerated withdrawal delivered on an outpatient basis are the most efficient and effective ways to achieve withdrawal from opioids. • Self-help groups. These groups provide the simplest form of postwithdrawal support for enduring abstinence and are also a low-cost intervention, because patients bear most of the costs; however, they have a low rate of uptake, and their effectiveness is only modest. • Oral opioid agonist maintenance treatment. This form of treatment is the most widely used intervention for illicit opioid dependence in developed societies. It has a better uptake than other interventions, and it is moderately effective under the usual delivery conditions. • Drug-free residential treatment. This form of treatment has a relatively low rate of treatment uptake and is costly because Illicit Opiate Abuse | 913 ©2006 The International Bank for Reconstruction and Development / The World Bank 115 of its residential character and the need for intensive staffpatient interaction. It is effective for the minority of people who are retained in treatment long enough to benefit from it (usually three months). Retention in treatment may be improved if patients enter treatment under some form of legal coercion. • Naltrexone maintenance treatment. This form of treatment has not been rigorously evaluated. • Injectable opioid maintenance. This intervention is a more expensive variant of agonist maintenance treatment that has been used for patients with more severe cases of dependency but for whom retention and treatment outcomes have been good. Calculation of the Averted, Avertable, and Unavertable Burden Assuming that the disease burden from opioid dependence is potentially avertable, we used the following approach to estimate the avoidable burden of opioid dependence. We initially modeled the avertable burden using MMT and used this model for BMT. The first step was to establish the base case for opioid dependence using 2002 as the baseline year. We established the model of the base case for opioid dependence for regions and subregions according to WHO country classifications. We used population estimates for each region for those age 15 to 59, the age range in which heroin dependence is most prevalent. We incorporated Degenhardt, Hall, and others’ (2004, table 13.1) figures for the prevalence of opioid use by region, assuming that the prevalence was 30 percent higher among male users than female users. We obtained population-attributable fractions related to opioid dependence from the editors of this volume. We used nine relevant WHO categories to estimate the burden of disease attributable to opioid dependence—namely HIV/AIDS, drug-use disorders, road traffic accidents, poisonings, falls, fires, drownings, other unintentional injuries, and self-inflicted injuries. We calculated the mortality rate for opioid deaths by dividing the number of deaths by the estimated number of users. We took estimates of years of life lost (YLLs) and years lived with disability (YLDs), by gender, for each region from data obtained from the editors of this volume. We then used those estimates to calculate the DALYs for male users, female users, and all users (YLL ⫹ YLD ⫽ DALY). We discounted the YLLs, YLDs, and DALYs using a 3 percent discount rate. The second step was to estimate the avertable burden by treatment with methadone or buprenorphine. Using the population and prevalence data, we assumed, in the first instance, that 50 percent of those dependent on opioids entered treatment. In the sensitivity analysis, we varied this proportion from 25 to 75 percent coverage. On the basis of Caplehorn and others’ (1994) meta-analysis, we assumed that MMT reduced mortality by 25 percent. In the sensitivity analysis, we varied the reduction from 15 to 35 percent (using the confidence intervals around the estimated reduction). We assumed that the reduction in mortality associated with BMT was 20 percent, which we varied in the sensitivity analysis from 10 to 30 percent. Finally, we assumed that those who were alive and in treatment experienced a 25 percent reduction in disability, consistent with the Dutch disability weights. The third step was to estimate the burden for those not treated. For those users not in treatment, we calculated DALYs using the original mortality rates. The fourth step was to estimate the total avertable burden from treatment with methadone or buprenorphine by (a) adding the results of the second and third steps, the revised DALYs for those in treatment, and the residual for those not in treatment and (b) subtracting those figures from the base case estimates. The fifth step was to cost the interventions using data on MMT and BMT from Doran and others (2003). They estimated the cost of MMT at $A 1,415 and of BMT at $A 1,729 for six months of treatment.We converted these estimates into U.S. dollars and multiplied them by two to provide yearly estimates of treatment costs of US$1,732 for MMT and US$2,117 for BMT. We applied relative price weights for each region using the Western Pacific as the reference case (1.00). We calculated the relative price weights for each cost type using data provided by the World Bank. The prices are a reflection of the public health systems in each region, and as far as possible they reflect the opportunity cost of health care resources in these regions. Results. Our results are presented in table 48.1. We explored various combinations of coverage and reductions in mortality for MMT and BMT. For each intervention, as coverage and reductions in mortality increased, the number of DALYs averted increased. The wide discrepancies in DALYs averted within regions primarily reflect differences in population-attributable fractions for HIV/AIDS. Costs increased as a consequence of increased coverage for both interventions, whereas results for cost-effectiveness differ by both intervention and mortality. The cost-effectiveness analysis suggests that for MMT (with a coverage of 25, 50, or 75 percent and reductions in mortality of 35 percent) the cost in international dollars per DALY averted ranges from a low of $128 in Africa, with high child and adult mortality where the prevalence of illicit opioid dependence is low (0.01 percent), to a high of $3,726 in Eastern Europe, with low child and adult mortality where the prevalence of illicit opioid dependence is high (0.55 percent). Across all the regions, the average cost-effectiveness ratio for MMT (with 25, 50, and 75 percent coverage and 35 percent reduction in mortality) is estimated at $2,236 per DALY averted. Assessment. The results shown in table 48.1 provide a first approximation of the potential avertable burden in DALYs if 914 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 116 Table 48.1 Cost-Effectiveness Results Total effect (DALYs averted per 1 million population) Eastern Mediterranean The Americas Africa Southeast Asia Europe Western Pacific Coverage (%) Mortality (%) MMT 25 15 125 79 153 107 158 179 105 117 MMT 50 15 251 158 306 214 316 358 210 MMT 75 15 376 237 459 321 474 538 315 MMT 25 25 150 81 184 121 173 217 MMT 50 25 300 163 369 243 347 MMT 75 25 450 244 553 364 520 MMT 25 35 174 84 216 136 189 256 198 165 69 329 122 92 63 43 MMT 50 35 349 167 432 272 378 511 396 331 139 657 244 184 126 87 MMT 75 35 523 251 648 408 566 767 594 496 208 986 367 276 189 130 BMT 25 10 113 78 137 100 150 160 82 105 43 166 48 38 32 22 BMT 50 10 226 156 274 199 301 320 163 210 85 331 97 75 65 44 BMT 75 10 339 234 412 299 451 480 245 315 128 497 145 113 97 67 BMT 25 20 138 80 169 114 166 198 128 129 53 231 78 59 45 31 BMT 50 20 275 160 337 228 332 397 256 258 107 462 156 119 89 61 BMT 75 20 413 240 506 342 497 595 384 388 160 693 234 178 134 92 BMT 25 30 162 82 200 129 181 237 175 153 64 296 107 81 57 39 BMT 50 30 324 165 400 258 362 473 350 307 128 592 215 162 114 78 BMT 75 30 487 247 601 386 543 710 524 460 192 888 322 243 171 117 Treatment AFR-D AFR-E AMR-A AMR-B AMR-D EMR-A EMR-D EUR-A EUR-B EUR-C SEAR-B SEAR-D WPR-A WPR-B 48 198 63 234 96 397 126 352 144 595 190 151 141 59 264 93 435 303 283 117 527 652 454 424 176 791 48 39 26 97 77 53 145 116 79 70 51 35 185 140 102 70 278 211 152 105 Total costs (US$ per 1 million population) MMT 25 15, 25, 35 0.10 0.01 0.25 0.06 0.12 0.95 0.65 0.20 0.16 0.35 0.06 0.19 0.07 0.03 MMT 50 15, 25, 35 0.19 0.02 0.50 0.11 0.24 1.90 1.30 0.40 0.32 0.71 0.11 0.39 0.13 0.07 MMT 75 15, 25, 35 0.29 0.03 0.74 0.17 0.36 2.86 1.95 0.60 0.49 1.06 0.17 0.58 0.20 0.10 BMT 25 10, 20, 30 0.12 0.01 0.30 0.07 0.15 1.16 0.80 0.24 0.20 0.43 0.07 0.24 0.08 0.04 BMT 50 10, 20, 30 0.24 0.03 0.60 0.14 0.29 2.33 1.59 0.49 0.40 0.86 0.14 0.47 0.16 0.08 BMT 75 10, 20, 30 0.35 0.04 0.91 0.20 0.44 3.49 2.39 0.73 0.59 1.29 0.20 0.71 0.24 0.12 Cost-effectiveness (US$ per DALY averted) MMT 25, 50, 75 15 768 136 1,618 520 755 5,315 6,213 1,711 3,379 1,782 875 3,984 1,716 1,284 MMT 25, 50, 75 25 643 132 1,342 458 688 4,381 4,300 1,419 2,764 1,341 597 2,749 1,301 974 MMT 25, 50, 75 35 552 128 1,146 408 632 3,726 3,288 1,212 2,339 1,074 453 2,099 1,048 784 BMT 25, 50, 75 10 1,041 168 2,204 682 969 7,269 9,764 2,329 4,646 2,606 1,396 6,277 2,493 1,867 BMT 25, 50, 75 20 855 164 1,793 595 880 5,869 6,210 1,895 3,716 1,869 867 3,975 1,809 1,354 BMT 25, 50, 75 30 726 159 1,510 527 805 4,921 4,553 1,598 3,096 1,458 629 2,909 1,419 1,062 DALYs averted per US$1 million spent MMT 25, 50, 75 15 1,302 7,363 618 1,922 1,325 188 161 585 296 561 1,142 251 583 779 MMT 25, 50, 75 25 1,556 7,575 745 2,185 1,453 228 233 705 362 746 1,676 364 768 1,027 MMT 25, 50, 75 35 1,811 7,787 873 2,448 1,582 268 304 825 428 931 2,210 476 954 1,275 BMT 25, 50, 75 10 961 5,939 454 1,465 1,032 138 102 429 215 384 717 159 401 536 BMT 25, 50, 75 15 1,170 6,112 558 1,681 1,137 170 161 528 269 535 1,153 252 553 739 BMT 25, 50, 75 20 1,378 6,286 662 1,896 1,242 203 220 626 323 686 1,590 344 705 942 Illicit Opiate Abuse | 915 ©2006 The International Bank for Reconstruction and Development / The World Bank 117 MMT and BMT were applied to 50 percent of the opioiddependent population in each region. Because the methods and data used to estimate avertable DALYs are subject to certain limitations, those results should be considered preliminary. RELEVANCE TO DEVELOPING COUNTRIES Much of the epidemiological research on illicit opioid dependence, its disease burden, and its societal harm comes from Australasia, Europe, and the United States. The major exception is research on the role of injecting drug use in HIV transmission in developing countries (see, for example, Beyrer and others 2000; Yu and others 1998). In addition, research on the effectiveness and cost-effectiveness of interventions for illicit opioid dependence has been conducted primarily in developed countries (Ward, Hall, and Mattick 1998), with the exception of studies of the effectiveness of methadone treatment in Hong Kong, China (see, for instance, Newman and Whitehill 1979), and in Thailand (Vanichseni and others 1991), both of which showed comparable effectiveness to that found in developed countries (W. Hall, Ward, and Mattick 1998). Translating findings on interventions for opioid dependence in developed countries into disease control priorities for opioid dependence in developing countries presents three major challenges. First, countries differ in the scale of illicit opioid use and in the resulting disease burden. This variation reflects the effects of differences in the prevalence of injecting and noninjecting opioid users; the dependent opioid users’ access to treatment and health services for overdoses, bloodborne viruses, and other complications of drug use; the access to needle and syringe programs; the extent to which illicit opioid use is concentrated in socially disadvantaged minority groups; and the capacity of public health services to monitor and respond to emerging infectious disease and drug-use epidemics. The burden is likely to be greatest in settings where the primary route of administration is injecting and where public and personal health services are poorly developed, as appears to be the case in Asia and in Eastern Europe. Second, societal wealth and health care infrastructure affect the capacity of developing societies to treat illicit opioid dependence. A country’s capacity to provide opioid substitution treatment will be affected by the cost of oral opioid drugs, such as methadone, LAAM, and buprenorphine, and the existence of specialist drug treatment centers; trained medical, nursing, and pharmacy staff; and a drug regulatory system, which are required so as to deliver opioid substitution treatment safely and effectively. Few developing countries possess this infrastructure. However, examples exist of apparently successful drug substitution programs, using such tools as sublingual buprenorphine, that have been conducted with minimal resources in extremely poor settings (Crofts and others 1998). Third, in societies with a sizable illicit opioid dependence problem, cultural attitudes and beliefs will affect societal responses, especially attitudes toward illicit opioid use and dependence (Gerstein and Harwood 1990). A critical determinant of the social response will be the relative dominance of moral and medical understandings of drug dependence in general and opioid dependence in particular. A moral model of addiction sees addiction as largely a voluntary behavior, in which case it is seen as an excuse for bad behavior that allows drug users to continue to take drugs without assuming responsibility for their conduct (Szasz 1985). In this view, drug users who offend against the criminal code should be imprisoned (Szasz 1985). This model is the dominant one in many developed societies, which imprison drug users at high rates without any effect on the prevalence of drug abuse. Countries that adopt punitive policies toward drug users are reluctant to embrace harm reduction measures, such as needle and syringe programs and opioid maintenance treatment (Ainsworth, Beyrer, and Soucat 2003). A medical model of addiction, by contrast, recognizes that dependent opioid users require specific treatment if the sufferer is to become and remain abstinent (see, for example, Leshner 1997). These competing views will affect the societal acceptability of opioid maintenance and abstinence-oriented approaches to the treatment of opioid dependence (Cohen 2003). Those who have a moral view of addiction will tend to prefer drug-free and self-help approaches toward treatment. Supporters of medical models of addiction will favor some form of opioid substitution treatment and the provision of clean needles and syringes to reduce the transmission of bloodborne viruses by injecting opioid and other drug users. Stronger advocacy by international organizations and agencies is needed for the adoption of such harm reduction measures as needle and syringe programs and agonist substitution programs. RESEARCH AND DEVELOPMENT Two main areas are important for research and development. First, better estimates are needed of the prevalence of illicit opioid dependence and prospective studies of the morbidity and mortality that it causes in both developed and developing countries. These estimates are especially needed in countries where illicit opioid use is high because of their proximity to source countries. Second, we need evaluations of the effectiveness and cost-effectiveness of self-help, drug-free, and oral opioid substitution treatment in developing countries. A priority should be the identification of safe, innovative, and less expensive ways of effectively delivering culturally acceptable forms of opioid maintenance treatments in developing countries. This effort may require experimentation with a range of substitute opioids, such as buprenorphine, and cheaper options, such as codeine and opium tincture. 916 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 118 CONCLUSIONS: PROMISES AND PITFALLS Illicit opioid use, especially injecting use, contributes to premature mortality and morbidity in many developed and developing societies. Fatal overdoses and HIV/AIDS resulting from the sharing of dirty injecting equipment are major contributors to mortality and morbidity, and the economic costs of illicit opioid dependence are substantial. Illicit opioid dependence generates substantial externalities that are not included in burden-of-disease estimates, principally law enforcement costs incurred in handling drug dealing and property crime. The most popular interventions for illicit opioid dependence in many developed societies have been law enforcement efforts to interdict the drug supply and enforce legal sanctions against the use of opioid drugs. One consequence of this strategy has been that most illicit opioid users have been exposed to the least effective intervention: imprisonment for drug or property offenses. Prisons rarely take the opportunity to treat dependence using opioid maintenance or to reduce the harm caused by illicit opioid use by providing access to clean injecting equipment. In treatment settings, the most popular interventions have been detoxification (which is not a treatment but a prelude to treatment) and drug-free treatment (which is the least attractive and the least effective in retaining opioid-dependent people in treatment). Opioid agonist maintenance treatment has been ambivalently supported in many developed societies despite its being the treatment for which there is the best evidence of effectiveness, safety, and cost-effectiveness. The range of opioid agonists available for maintenance treatment is increasing. A number of developed countries have approved the use of BMT, which the limited data suggest may be approximately equivalent to MMT in efficacy and cost-effectiveness. Opioid antagonists have a niche role in the treatment of opioid dependence because of poor compliance and an increased risk of overdose on return to heroin use. Their efficacy may improve with the development of long-acting injectable forms of the drug. ANNEX 48.A: PREVALENCE OF USE, ADVERSE HEALTH EFFECTS OF AND INTERVENTIONS FOR CANNABIS, COCAINE, AMPHETAMINES, AND MDMA USE AND DEPENDENCE Cannabis Cannabis is the most widely used illicit drug globally, with about 150 million users, or 3.7 percent of the world’s population age 15 and older (UNODCCP 2003). Patterns of cannabis use have been most extensively studied in Australia, Canada, the United States, and Europe (W. Hall and Pacula 2003). Europe generally has lower rates of use than Australia, Canada, and the United States, with the highest rates in Denmark, France and the United Kingdom (EMCDDA 2002; W. Hall and Pacula 2003). The limited data from developing countries suggest that, with some exceptions (for example, Jamaica and South Africa), rates of cannabis use are lower in Africa, Asia, the Caribbean, and South America than they are in Europe and in English-speaking countries (W. Hall, Johnston, and Donnelly 1999). Surveys in the United States have found long waves of cannabis use among young people since 1975. Cannabis use increased during the 1970s to peak in 1979, before declining steadily between 1980 and 1991. Use rose sharply in 1992 and increased throughout the 1990s, before leveling off in the late 1990s (Johnston, O’Malley, and Bachman 1994a, 1994b). There was also a rise in cannabis use during the early 1990s in Australia, Canada, and some European countries (W. Hall and Pacula 2003). The natural history of cannabis use in the United States typically begins in the mid to late teens and reaches its maximum in the early 20s before declining in the mid to late 20s. Only a minority of young adults continue to use cannabis into their 30s (Bachman and others 1997; Chen and Kandel 1995). Getting married and having children substantially reduces rates of cannabis use (Bachman and others 1997). Cannibis use can have several adverse health effects, as discussed below. Acute Effects of Cannabis Use. The most frequent unpleasant effects of cannabis use are anxiety and panic reactions, which most often occur in users who are unfamiliar with the drug’s effects. Psychotic symptoms such as delusions and hallucinations may be experienced following very high doses. There are no cases of fatal cannabis poisoning in the medical literature, and the fatal dose in humans is likely to exceed what recreational users are able to ingest (W. Hall and Pacula 2003). Cannabis intoxication impairs a wide range of cognitive and behavioral functions that are involved in driving an automobile or operating machinery (Beardsley and Kelly 1999; Jaffe 1985). It has been difficult to determine whether these impairments increase the risk of being involved in motor vehicle accidents (Smiley 1999). Studies of the effect of cannabis on driving performance on the road have found only modest impairments, because cannabis-intoxicated drivers drive more slowly and take fewer risks than drivers intoxicated by alcohol (Smiley 1999). Cannabinoids are found in the blood of substantial proportions of persons killed in motor vehicle accidents (Bates and Blakely 1999; Chesher 1995; Walsh and Mann 1999), but these findings have been difficult to evaluate because they have not distinguished between past and recent cannabis use (Ramaekers and others 2004). More recent research using better indicators of recent cannabis use has found a dose-response Illicit Opiate Abuse | 917 ©2006 The International Bank for Reconstruction and Development / The World Bank 119 relationship between cannabis and risk of motor vehicle crashes (Ramaekers and others 2004). Cannabis used in combination with alcohol substantially increases risk of accidents (Bates and Blakely 1999; Ramaekers and others 2004). Health Effects of Chronic Cannabis Use. Cannabis smoke is a potential cause of cancer because it contains many of the same carcinogenic substances as cigarette smoke (Marselos and Karamanakos 1999). Cancers have been reported in the aerodigestive tracts of young adults who were daily cannabis smokers (W. Hall and MacPhee 2002), and a case-control study has found an association between cannabis smoking and head and neck cancer (Zhang and others 1999). A prospective cohort study of 64,000 adults did not find any increase in rates of head and neck or respiratory cancers (Sidney and others 1997). Further studies are needed to clarify the issue. Three studies of different types of cancer have reported an association with maternal cannabis use during pregnancy (W. Hall and MacPhee 2002). There have not been any increases in the rates of these cancers that parallel increases in rates of cannabis use (W. Hall and MacPhee 2002). High doses of cannabinoids impair cell-mediated and humoral immunity and reduce resistance to infection by bacteria and viruses in rodents (Klein 1999). Cannabis smoke impairs the functioning of alveolar macrophages, the first line of the body’s immune defense system in the lungs. The doses that produce these effects have been very high, and extrapolation to the doses used by humans is complicated by the fact that tolerance to these effects develops (Hollister 1992). There is as yet no epidemiological evidence that rates of infectious disease are higher among chronic heavy cannabis users. Several large prospective studies of HIV-positive homosexual men have not found that cannabis use makes it more likely that HIV-positive men develop AIDS (W. Hall and Pacula 2003). Chronic administration of tetrahydrocannabinol (THC) disrupts male and female reproductive systems in animals, reducing testosterone secretion and sperm production, motility, and viability in males and disrupting ovulation in females (Brown and Dobs 2002). It is uncertain whether cannabis use has these effects in humans because of the limited research on human males and females (Murphy 1999). The use of cannabis during pregnancy is associated with smaller birthweight (English and others 1997; Fergusson, Horwood, and Northstone 2002), but it does not appear to increase the risk of birth defects (W. Hall and Pacula 2003). In some studies, infants exposed to cannabis during pregnancy show behavioral and developmental effects during the first few months after birth; these effects are smaller than those seen after tobacco use during pregnancy (Fried and Smith 2001). The changes that cannabis smoking causes in heart rate and blood pressure are unlikely to harm healthy young adults, but they may harm patients with hypertension, cerebrovascular disease, and coronary atherosclerosis (Chesher and Hall 1999; Sidney 2002). One controlled study suggests that cannabis use can precipitate heart attacks in middle-aged cannabis users who have atherosclerosis in the heart, brain, and peripheral blood vessels (Mittleman and others 2001). Regular cannabis smoking impairs the functioning of the large airways and causes chronic bronchitis (Tashkin 1999; Taylor and others 2002). Given that tobacco and cannabis smoke contain similar carcinogenic substances, it is likely that chronic cannabis smoking increases the risks of respiratory cancer (Tashkin 1999). Psychological Effects of Chronic Cannabis Use. Psychological effects of chronic cannabis use can include a dependence syndrome, cognitive effects, and psychotic disorders. Dependence Syndrome A cannabis dependence syndrome occurs in heavy chronic users of cannabis (American Psychiatric Association 1994). Regular cannabis users develop tolerance to THC. Some experience withdrawal symptoms on cessation of use (Kouri and Pope 2000), and some report problems controlling their cannabis use (W. Hall and Pacula 2003). The risk of dependence is about 1 in 10 among those who ever use the drug, between 1 in 5 and 1 in 3 among those who use cannabis more than a few times, and about 1 in 2 among daily users (W. Hall and Pacula 2003). Cognitive Effects Long-term daily cannabis use does not severely impair cognitive function, but it may more subtly impair memory, attention, and the ability to integrate complex information (Solowij 1998; Solowij and others 2002). It remains uncertain whether these effects are due to the cumulative effect of regular cannabis use on cannabinoid receptors in the brain or whether they are residual effects of THC that will disappear after an extended period of abstinence (W. Hall and Pacula 2003). Psychotic Disorders There is now good evidence that chronic cannabis use may precipitate psychosis in vulnerable individuals (see, for example, Arseneault and others 2002; van Os and others 2002; Zammit and others 2002). It is less likely that cannabis use can cause psychosis de novo, because the incidence of schizophrenia has either remained stable or declined while cannabis use has increased among young adults (Degenhardt, Hall, and Lynskey 2003). Effects of Cannabis Use on Adolescents. Cannabis use has a number of effects on adolescents. Gateway Hypothesis Adolescents in developed societies typically use alcohol and tobacco before using cannabis, which in turn, they use before using hallucinogens, amphetamines, heroin, and cocaine (Kandel 2002). Generally, the earlier the age of first use and the greater the involvement with any drug 918 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 120 in the sequence, the more likely a young person is to use the next drug in the sequence (Kandel 2002). The role played by cannabis in this sequence remains controversial (W. Hall and Lynskey forthcoming; W. Hall and Pacula 2003). The simplest hypothesis is that cannabis use has a pharmacological effect that increases the risk of using drugs later in the sequence. Equally plausible hypotheses are that it is due to a combination of (a) early recruitment into cannabis use of nonconforming and deviant adolescents who are likely to use alcohol, tobacco, and illicit drugs; (b) a shared genetic vulnerability to dependence on alcohol, tobacco, and cannabis; and (c) socialization of cannabis users within an illicit drug–using subculture, which increases the opportunity, and encouragement to use other illicit drugs (W. Hall and Pacula 2003). Adolescent Psychosocial Outcomes Cannabis use is associated with early withdrawal from high school, early family formation, poor mental health, and involvement in drug-related crime. In the case of each of these outcomes, the strong associations in cross-sectional data are more modest when account is taken of the fact that cannabis users show characteristics before they use cannabis that predict these outcomes. For example, they have lower academic aspirations and poorer school performance than peers who do not use cannabis (Lynskey and Hall 2000; Macleod and others 2004). Nonetheless, the evidence increasingly suggests that regular cannabis use adds to the risk of these outcomes in adolescents already at risk (W. Hall and Pacula 2003). Interventions for Cannabis Dependence. Although many dependent cannabis users may succeed in quitting without professional help, some are unable to stop on their own and will need assistance to do so. There has not been a great deal of research on pharmacological treatments for cannabis dependence, although a recent study trialed divalproex sodium with promising results (Levin and others 2004). Limited research exists on the effectiveness of different types of psychosocial treatments for dependent cannabis use (Budney and others 2000; Copeland and others 2001; Stephens, Roffman, and Simpson 1994). These approaches have involved short-term cognitive behavioral treatments modeled on similar treatments for alcohol dependence, usually given in three to six sessions on an outpatient basis. In all of these studies, rates of abstinence at the end of treatment have been modest (20 to 40 percent), and subsequent high rates of relapse mean that rates of abstinence after 12 months have been very modest (Budney and Moore 2002). Nonetheless, treatment does substantially reduce cannabis use and problems. These outcomes are not very different from those observed in the treatment for alcohol and other forms of drug dependence (Budney and Moore 2002). Much more research is needed before sensible advice can be given about the best ways to achieve abstinence from cannabis. Cocaine After cannabis, cocaine is one of the most widely used illicit drugs in developed and developing societies. Some 14 million people were estimated to have used cocaine globally in 2003, with demand for treatment second only to heroin (UNODCCP 2003). The highest rates of reported cocaine use—and the best data on trends in cocaine use—come from the United States, the world’s largest cocaine market. Rates of cocaine use in the United States increased from the mid 1970s until 1985, when 5.7 million Americans age 12 and older reported using cocaine in the preceding month. Rates of cocaine use in the preceding month have declined steadily since 1985. In 2000, 11.2 percent of Americans over age 12 reported that they had used cocaine at some time in their lives, and 0.4 percent (800,000 people) reported weekly cocaine use (SAMHSA 2001). Among young U.S. adults age 18 to 25, lifetime prevalence was 14.9 percent in 2001, rising slightly to 15.4 percent in 2002 (SAMHSA 2003). In 2002, annual prevalence figures from student surveys were 15 percent lower than 1998 figures and 60 percent lower than 1985 figures (UNODCCP 2003). A more recent study of U.S. adults age 35 years found that 6 percent of men and 3 percent of women had used cocaine within the preceding 12 months (Merline and others 2004). The reported prevalence of cocaine use in other developed societies is much lower than that in the United States. In Europe, for example, rates of lifetime cocaine use range from 0.5 percent to 5 percent (EMCDDA 2003), compared with 12.3 percent among American adults in 2001 (SAMHSA 2001). Rates of cocaine use in Australia resemble those in Europe, with 4.3 percent of adults reporting lifetime use (Darke and others 2000). The prevalence of cocaine use is likely to be lower in developing societies, but the poor quality of the available data makes it difficult to be sure (UNDCP 1997). There probably has been an increase in cocaine use in some developing countries in recent years, but it is difficult to estimate the size of the increase (United Nations Commission on Narcotic Drugs 2000). The region with the highest rates of cocaine use among developing societies is likely to be Central and South America. The botanical source is indigenous to the region and has traditionally been used by local populations. Moreover, several nations in Central and South America have a history of production and export to global markets. Recent reports indicate that cocaine abuse is increasing in South America (UNODCCP 2003), and a recent household survey on drug abuse in São Paulo, Brazil, estimated cocaine prevalence at 2.1 percent (Galduroz and others 2003). Adverse Health Effects of Cocaine. Most cocaine use is infrequent; regular cocaine use (monthly or more frequently) can be a major public health problem. Regular cocaine users who Illicit Opiate Abuse | 919 ©2006 The International Bank for Reconstruction and Development / The World Bank 121 inject cocaine or smoke crack cocaine are especially likely to develop dependence and to experience problems related to their cocaine use (Platt 1997). In the United States, it has been estimated that one in six of those who ever use cocaine become dependent on the drug (Anthony, Warner, and Kessler 1994). High rates of cocaine dependence are found among people treated for alcohol and drug problems and among arrestees in the United States (Anglin and Perrochet 1998). In large doses, cocaine may be harmful in both cocainenaive and cocaine-tolerant individuals (Platt 1997; Vasica and Tennant 2002). The vasoconstrictor effects of cocaine in large doses place great strains on a number of the body’s physiological systems (McCann and Ricaurte 2000). Effects on the cardiovascular system can result in a range of difficulties, from chest pain to fatal cardiac arrests (Lange and Hillis 2001). Neurological problems include cerebral vascular accidents such as strokes or seizures. Other effects of cocaine can include gastrointestinal problems such as vomiting, colitis, and bowel infarction and respiratory problems such as asthma, respiratory collapse, pulmonary edema, and bronchitis. Hyperthermia may occur because of the increased metabolism, peripheral vasoconstriction, and inability of the thalamus to control body temperature (Crandall, Vongpatanasin, and Victor 2002). Obstetric complications can include irregularities in placental blood flow, premature labor, and low neonate birthweight (Majewska 1996; Platt 1997; Vasica and Tennant 2002). Adverse health effects from cocaine are potentially fatal and can occur among healthy users irrespective of cocaine dose and frequency of use (Lange and Hillis 2001; Vasica and Tennant 2002). Although the likelihood of health problems may increase with dosage and frequency of use, there is wide individual variation in reactions to cocaine and, therefore, no specific combination of conditions under which adverse health effects can be predicted. There is no antidote to cocaine overdose as there is for an overdose of heroin (Platt 1997). The impact of cocaine on mental health is also complex. Although cocaine can produce feelings of pleasure, it may also result in negative psychological symptoms such as anxiety, depression, paranoia, hallucinations, and agitation (American Psychiatric Association 1994). Regular cocaine users experience high rates of psychiatric disorders. In the United States, regular cocaine users report high rates of anxiety and affective disorders (Gawin and Ellinwood 1988; Platt 1997). The repeated use of large doses of cocaine can also produce a paranoid psychosis (Majewska 1996; Manschreck and others 1988; Platt 1997; Satel and Edell 1991). People who are acutely intoxicated by cocaine can become violent, especially those who develop a paranoid psychosis (Platt 1997). Animal studies suggest that cocaine use may be neurotoxic in large doses—that is, it can produce permanent changes in the brain and neurotransmitter systems (Majewska 1996; Platt 1997). It is unclear whether use is also neurotoxic in humans. Previous studies have documented a variety of neuropsychological effects of cocaine use, including deficits in memory and problem solving (Beatty and others 1995; Hoff and others 1996; O’Malley and others 1992). More recently, a twin study indicated that cocaine may lead to impaired attention and motor skills up to one year after the conclusion of heavy use (Toomey and others 2003). The method by which cocaine is administered can result in adverse health effects (Platt 1997). Snorting cocaine through the nose can lead to rhinitis, damage to the nasal septum, and loss of the sense of smell. Smoking cocaine can lead to respiratory problems, and injecting cocaine leads to the risks of infections and bloodborne viruses associated with all injecting drug use. Users who inject cocaine, either on its own or in combination with heroin (“speedballs”), inject much more frequently than other injecting drug users and, as a consequence, engage in more needle sharing, take more sexual risks, and have higher rates of HIV infection (Chaisson and others 1989; Schoenbaum and others 1989; van Beek, Dwyer, and Malcolm 2001). Associations between cocaine use and HIV risk-taking have been reported in Europe (Torrens and others 1991), Australia (Darke and others 1992), and the United States (Chaisson and others 1989). Recent Australian research has indicated that injecting cocaine users report more problems related to injecting drug use—such as vascular problems, abscesses, and infections—than other injecting drug users (Darke, Kaye, and Topp 2002). The link between cocaine use and HIV risk is not restricted to those who inject cocaine. Crack smoking has been linked to higher levels of needle risk, sexual risk taking, and HIV infection (Chaisson and others 1989; Chirgwin and others 1991; Desjalais and others 1992; Grella, Anglin, and Wugalter 1995). Two mechanisms probably underlie the relationship between cocaine use and HIV infection. First, the short half-life of cocaine promotes a much higher frequency of injecting by users than that seen in heroin injectors. Second, cocaine itself disinhibits and stimulates users, encouraging them to take greater risks with sexual activity and needle use (Darke and others 2000). Cocaine is associated with a risk of intentional injuries and injuries in general. A recent review reported that 28.7 percent of people with intentional injuries and 4.5 percent of injured drivers tested positive for cocaine (Macdonald and others 2003). Users are also at risk of death from an accidental overdose of cocaine. A recent study of accidental deaths from drug overdose in New York between 1990 and 1998 found that 70 percent of deaths were caused by cocaine, often in combination with opiates (Coffin and others 2003). The causes of cocaine-related deaths are usually related to cardiovascular complications (Vasica and Tennant 2002), but death may also be due to brain hemorrhage, stroke, and kidney failure (Brands, Sproule, and 920 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 122 Marshman 1998). Injection of cocaine is most likely to cause an overdose, followed by smoking it, with intranasal use involving the least risk (Pottieger and others 1992). Much less is known about nonfatal cocaine overdose. A study in Miami, Florida, found that 40 percent of users had overdosed on cocaine at least once (Pottieger and others 1992). More recently, a study in Brazil found that 20 percent of users had experienced an overdose, with 50 percent knowing someone who had died from an overdose (Mesquita and others 2001). A study in Sydney, Australia, found that 17 percent of injecting cocaine users and 6 percent of noninjecting cocaine users had ever overdosed, with 9 percent and 3 percent, respectively, overdosing in the preceding 12 months (Kaye and Darke 2003). Frequency of cocaine use, severity of dependence, and route of administration did not predict an overdose, supporting the view that cocaine overdose is an unpredictable event. Interventions for Cocaine Dependence. Efforts at intervention have included pharmacological treatments as well as psychotherapy and cognitive behavioral therapy. Pharmacological Interventions Despite much research effort there are no effective pharmacological treatments for cocaine dependence (Kreek 1997; McCance 1997; Mendelson and Mellon 1996; Nunes 1997; Silva de Lima and others 2002; van den Brink and van Ree 2003). Attempts have been made to develop longer-acting agonist drugs that act on the same molecular targets as cocaine without producing its euphoric effects (for example, methylphenidate) (Kreek 1997) or that block its rewarding and euphoric effects (McCance 1997). There has also been a search for drugs that indirectly change the effects that cocaine has on the brain by acting on other neurotransmitter systems, such as the serotonergic system (for example, fluoxetine) (McCance 1997). None of these approaches has produced an effective pharmacotherapy for cocaine dependence (Lima and others 2003; Platt 1997; Soares and others 2003). Development of pharmacological therapies for cocaine dependence and their evaluation is complicated by the multiple interactive processes that may have contributed—for example, coexisting substance abuse or mental health issues (Mendelson and Mellon 1996). Many of the approaches to the treatment of cocaine dependence have also been used in treating patients with alcoholism and other substance abuse disorders. A number of drugs have been used to treat cocaine based on their relevance to the symptoms of cocaine dependence (Silva de Lima and others 2002; van den Brink and van Ree 2003). The frequency of depressive symptoms has led to the exploration of the effectiveness of antidepressant drugs. Desipramine has been used with mixed effectiveness for cocaine detoxification and the maintenance of abstinence (Covi and others 1994; Gawin, Kleber, and Byck 1989), but it appears to be most effective when there is evidence of previous or consequent symptoms of depression. Other antidepressants have been used with mixed results: imipramine and trazodone have been found to have more adverse effects than desipramine, and fluoxetine has not been found to be effective (Mendelson and Mellon 1996). A recent systematic review found no current evidence to support the use of antidepressants in the treatment of cocaine dependence (Lima and others 2003). Dopamimetic drugs have also been used to treat cocaine dependence; such treatments are based on the action of cocaine to block reuptake of dopamine. Unfortunately, although some of these drugs are relatively effective, they also result in quite severe adverse effects (Mendelson and Mellon 1996). Current evidence does not support the clinical use of dopamine agonists for cocaine dependence (Soares and others 2003). Opioid antagonists (for example, naltrexone) or opioid mixed agonistantagonists (such as buprenorphine) have been explored, on the basis that cocaine dependence may be accompanied by dependence on opiates. Although there have been problems with compliance with naltrexone therapy (National Research Council Committee on Clinical Evaluation of Narcotic Antagonists 1978), buprenorphine has shown promising preclinical and clinical trial results (Kosten, Kleber, and Morgan 1989). Other promising directions include cannabinoid receptor antagonists and cortisol synthesis inhibitors (van den Brink and van Ree 2003) and vaccination against the effects of cocaine (Kantak 2003), but there is as yet no evidence on the effectiveness of any of these interventions. Acupuncture has also been used to treat cocaine dependence. Auricular acupuncture is frequently used, but the small number of trials that have been conducted have not provided sufficient evidence of effectiveness (van den Brink and van Ree 2003). Psychotherapy and Cognitive Behavioral Therapy The lack of evidence for pharmacological therapy means that treatment for cocaine dependence currently relies on cognitive behavior therapies combined with contingency management strategies. Unfortunately, psychosocial treatments for cocaine dependence are also of limited effectiveness. Treatments such as therapeutic communities, cognitive behavioral treatments, contingency management, and 12 step–based self-help approaches benefit cocaine-dependent people by reducing their rates of cocaine use and improving their health and well-being, but rates of relapse to cocaine use after treatment remain high (Platt 1997). Mendelson and Mellon (1996) conclude that there are no specific cognitive or behavioral interventions that are uniquely effective in treating cocaine dependence. However, some success has been demonstrated with incentive-based programs in which rewards are provided for urine samples that are free of cocaine, although there is doubt about whether results are sustained (Roozen and others 2004). Such programs are generally more Illicit Opiate Abuse | 921 ©2006 The International Bank for Reconstruction and Development / The World Bank 123 effective when the patient’s family and friends are involved (Higgins and others 1994). Petry and others (2004) suggested that contingency management was effective in reducing cocaine use in a community-based treatment setting. They found that the benefits of treatment depended on the magnitude of reward, with those earning up to US$240 obtaining better results than those earning up to US$80. They suggested that this form of intervention may work best for people with more severe dependence on cocaine. A multicenter investigation examining the efficacy of four psychosocial treatments for cocaine-dependent patients concluded that individual drug counseling in combination with group drug counseling showed the most promise for effective treatment of cocaine dependence over two forms of traditional psychotherapy (Crits-Christoph and others 1999). Community reinforcement involving an intensive, biopsychosocial, multifaceted approach to lifestyle change has shown positive effects over four to six weeks and has the advantage of being tailored to individual goals (Roozen and others 2004). The few studies of the long-term effects of treatment have not shown particularly encouraging results. A one-year follow-up of the U.S. Drug Abuse Treatment Outcome Studies reported that reductions in the use of cocaine in the year following treatment were associated with longer duration of treatment, particularly six months or more in long-term residential or outpatient treatments (Hubbard, Craddock, and Anderson 2003). A five-year national follow-up study of 45 U.S. treatment programs found that only 33 percent of the sample had highly favorable outcomes (Flynn and others 2003). Amphetamines According to WHO, amphetamines and methamphetamines are the most widely abused illicit drugs after cannabis, with an estimated 35 million users worldwide (Rawson, Anglin, and Ling 2002). In Australia, the lifetime prevalence of amphetamine use is between 6 and 8 percent in the general population, making amphetamines the most commonly used illicit drug after cannabis during that period (Makkai and McAllister 1998). In 1998, the lifetime prevalence of amphetamine use was highest (25 percent) among male users age 20 to 29. The use of amphetamines is generally less frequent than that of opioids (Darke and Hall 1995; Darke, Kaye, and Ross 1999; W. Hall, Bell, and Carless 1993; Hando, Topp, and Hall 1997; Vincent and others 1998). This pattern is no doubt due to the physical and psychological toll taken by regular amphetamine use. Although such use is less frequent overall, however, there is widespread bingeing on amphetamines, with frequent use over several consecutive days, which may be followed by benzodiazepine use to “come down.” Polydrug use is particularly common among amphetamine users, who show a marked preference for stimulant drugs such as hallucinogens and cocaine (Darke and Hall 1995; Hando and Hall 1994; Vincent and others 1998). Globally, Europe is the main center of amphetamine production, particularly Belgium, the Netherlands, and Poland, with production increasing in Eastern Europe (UNODCCP 2003). Half of all Western European countries reported an increase in amphetamine abuse in 2000, but in 2001 the figure fell to 33 percent (UNODCCP 2003). Lifetime use of amphetamines is reported to be between 0.5 percent and 6 percent among European Union countries, with the exception of the United Kingdom, where the figure is 11 percent. Denmark and Norway also have relatively higher rates of use (EMCDDA 2003). Adverse Health Effects of Amphetamine Use. Amphetamine users who inject the drug are at high risk of bloodborne infections through needle sharing. Amphetamine users are as likely as opioid users to share injection equipment (Darke, Ross, Cohen, and others 1995; Darke, Ross, and Hall 1995; W. Hall, Bell, and Carless 1993; Hando and Hall 1994; Kaye and Darke 2000; Loxley and Marsh 1991). In addition, the youth of amphetamine users places them at risk of sexual transmission of diseases such as HIV and hepatitis B virus (although not hepatitis C). Primary amphetamine users have been demonstrated to be a sexually active group, and small proportions engage in paid sex to support their drug use (Darke, Ross, Cohen, and others 1995; Hando and Hall 1994). Among gay and bisexual men, amphetamines may be used to enhance sexual encounters, which may lead to unprotected anal intercourse and increased risk of HIV infection (Urbina and Jones 2004). High-dose amphetamine use, especially by injection, can result in a schizophreniform paranoid psychosis, associated with loosening of associations, delusions, and hallucinations (Gawin and Ellinwood 1988; Jaffe 1985). The psychosis could be reproduced by the injection of large doses in addicts (Bell 1973) and by the repeated administration of large doses to normal volunteers (Angrist and others 1974). High proportions of regular amphetamine injectors describe symptoms of anxiety, panic attacks, paranoia, and depression. The emergence of such symptoms is associated with injecting the drugs, greater frequency of use, and dependence on amphetamines (W. Hall and others 1996; McKetin and Mattick 1997, 1998). Recent evidence also suggests that women may experience more emotional effects of amphetamine intoxication than men and higher rates of anorexia nervosa than women without amphetamine disorders (Holdcraft and Iacono 2004). In sufficiently high doses, amphetamines can be lethal (Derlet and others 1989). However, the risk is low compared with the high risks of overdose associated with central nervous system depressants such as heroin. Typically, amphetaminerelated deaths are associated with the effects of amphetamines on the cardiovascular system—for example, cardiac failure and cerebral vascular accidents (Mattick and Darke 1995). 922 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 124 There is evidence that amphetamines are neurotoxic (Robinson and Becker 1986). Evidence from animal studies indicates that heavy amphetamine use results in dopaminergic depletion (Ellison 1992; Fields and others 1991). The few studies of the neuropsychological effects of amphetamine abuse report findings similar to those found with cocaine abuse. Deficits in memory and attention have been attributed to amphetamine use (McKetin and Mattick 1997, 1998). More recently, a twin study indicated that amphetamine abuse might lead to impaired attention and motor skills up to one year after the conclusion of heavy use (Toomey and others 2003). Interventions for Amphetamine Dependence. Treatment for methamphetamine abuse has been a relatively recent development and has generally been based on previous treatments for cocaine abuse (Huber and others 1997). Cretzmeyer and others (2003) reviewed treatments for methamphetamine abuse, noting that there has been little research on the effectiveness of drug treatment, probably because many amphetamine users use multiple drugs. The combination of methamphetamine use with use of marijuana or other sedating drugs indicates that effective treatments need to address the use of multiple drugs. A Cochrane Review concluded that evidence for success in treatment of amphetamine dependence is very limited, with no pharmacological treatment demonstrated to be effective (Srisurapanont, Jarusuraisin, and Kittirattanapaiboon 2003). An early study explored the use of aversion therapy in a multimodal treatment program using educational groups, individual counseling, occasional family counseling, and aftercare planning. The intervention paired an aversive stimulus (either chemical or electrical) with the act of using methamphetamines. Cocaine use was also treated in this way. After 12 months, 53 percent of patients were abstinent and the researchers noted that their results were promising, despite a number of limitations to the study (Frawley and Smith 1992). An intervention combining imipramine, a tricyclic antidepressant, with intensive group counseling has been evaluated with cocaine and methamphetamine abusers. Patients received either a low or higher dose (as needed) of imipramine, as well as intensive group counseling and access to medical and psychiatric care. Those who received the higher dose stayed in treatment longer, but the results did not support the use of imipramine for methamphetamine abuse (Galloway and others 1994). The Matrix Program for methamphetamine and cocaine abusers has also been evaluated. The Matrix Program uses a cognitive behavioral approach with an emphasis on relapse prevention (Huber and others 1997). The study evaluated the effectiveness of three conditions: Matrix treatment alone, Matrix treatment plus desipramine, and Matrix treatment plus placebo (Shoptaw and others 1994). The researchers concluded that those who received more Matrix treatment had better abstinence rates than those who had less treatment but that desipramine had no effect on treatment outcome. J. Hall and others (1999) conducted an evaluation of the effectiveness of the Iowa Case Management Project. The project was designed to supplement interventions provided by a drug abuse treatment agency and is a comprehensive social work intervention, including outreach activities and provision of limited emergency funds. The results of the evaluation showed that comprehensive case management was effective in improving employment status among amphetamine users subsequent to treatment. There was an almost significant lower incidence of depression among those who received the program compared with controls. Drug use decreased significantly for clients in both control and program conditions. More recently, an Australian study evaluated the effectiveness of brief cognitive-behavioral interventions among regular users of amphetamines (Baker, Boggs, and Lewin 2001). The researchers found a clinically significant reduction in daily amphetamine use among the intervention groups compared with controls and concluded that further studies of brief cognitive-behavioral interventions are feasible and warranted. Although some promising interventions have been identified to assist methamphetamine abusers, no single treatment option has yet been established as better than any other in a randomized controlled trial (Cretzmeyer and others 2003). Methylenedioxymethamphetamine Methylenedioxymethamphetamine is more widely known as ecstasy or MDMA. In Australia, the lifetime prevalence of MDMA use increased from 1 percent of the population in 1988 to 4.6 percent (about one in 20 persons ) in 1998, with 2.3 percent reporting MDMA use in the preceding 12 months (Topp and others 1998). In 2001, 6.1 percent of Australians age 14 years or older reported lifetime use of MDMA, with 2.9 percent reporting use within the preceding year (Degenhardt, Barker, and Topp 2004). Rates of use are generally higher among males than females (3.1 percent versus 1.5 percent). MDMA use in the preceding 12 months is most common among those age 20 to 29 (5 percent of females and 12 percent of males) (Topp and others 1998). The availability of MDMA has also increased, as indicated by the proportion of the population who have been offered MDMA (from 4 percent in 1988 to 7 percent in 1991) (Makkai and McAllister 1998), with 14 percent of those age 14 to 29 reporting that they had been offered MDMA in the preceding year. Research suggests that the pattern of MDMA use changed during the 1990s (Topp and others 1998). Users of MDMA are commencing use at a younger age, and they appear to be using larger doses more frequently. The incidence of bingeing on MDMA appears to have increased, as does the prevalence of the parenteral use of this drug. The increase in the use of MDMA Illicit Opiate Abuse | 923 ©2006 The International Bank for Reconstruction and Development / The World Bank 125 by injection has been noted among surveys of MDMA users and of injecting drug users generally. An examination of trends in the United States suggested that, although the use of MDMA has increased over time, its prevalence is significantly less than that of other drugs of abuse (Yacoubian 2003b). A study of 14,520 U.S. college students indicated 6 percent lifetime use of MDMA, 3 percent within the preceding 12 months, and 1 percent within the preceding 30 days. Those who had used MDMA in the preceding 12 months were more likely to be white and a member of a fraternity or sorority and to have used a range of other drugs (Yacoubian 2003a). Rates of use are much higher in surveys of club attendees. A recent U.S. survey found 86 percent reporting lifetime use, 51 percent 30-day use, and 30 percent use within the preceding 2 days (Yacoubian and others 2003). Abuse of MDMA had showed signs of decreasing in Western Europe but has recently shown signs of increase (UNODCCP 2003). Although MDMA use appears to be still diffusing, in 2003 only four countries (Ireland, the Netherlands, Spain, and the United Kingdom) reported a rate of more than 3 percent use among young adults in the preceding 12 months (EMCDDA 2003). In the United States, use declined in 2002 for the first time, but it increased in other regions, particularly the Caribbean, parts of South America, Oceania, Southeast Asia, the Near East, and southern Africa (UNODCCP 2003). Lifetime experience of MDMA is reported to range from 0.5 percent to 5 percent in European Union countries, with use more common in the Netherlands (EMCDDA 2003). Population survey findings from New Zealand reported an increase in the preceding-year use of MDMA from 1.5 percent in 1998 to 3.4 percent in 2001. The increase was particularly evident among young men age 20 to 24 (from 4.3 percent to 12.5 percent) (Wilkins and others 2003). Adverse Health Effects of MDMA. Early studies of MDMA use in Australia and the United States documented relatively few problems associated with the drug’s use (Beck 1990; Beck and Rosenbaum 1994; Downing 1986; Solowij, Hall, and Lee 1992). A survey of 100 MDMA users (Solowij, Hall, and Lee 1992) found that the most common adverse effects were the side effects of acute use, such as appetite loss, dry mouth, palpitations, and bruxism (teeth grinding). Among the few heavy users in the study, only two reported feeling dependent on the drug. With a change in the pattern of MDMA use in Australia, there has been an increase in the MDMA-related harms reported (Topp and others 1998). Some of the acute physical and psychological adverse effects that MDMA users have attributed to the use of this drug include energy loss, irritability, muscular aches, insomnia, and depression. More chronic adverse effects were also reported, including weight loss, depression, energy loss, insomnia, anxiety, and teeth problems. A recent U.K. study of 430 regular users of MDMA reported that 83 percent of participants reported low mood and 80 percent experienced impaired concentration. Long-term effects of MDMA included the development of tolerance to MDMA (59 percent), impaired ability to concentrate (38 percent), and depression (37 percent) (Verheyden and others 2003). Physical symptoms that were perceived as being due to MDMA use alone (Topp and others 1998) included an inability to urinate, blurred vision, vomiting, numbness or tingling, loss of sexual urge, and hot and cold flushes. As with amphetamines, the use of MDMA to facilitate sexual encounters may lead to risky sexual behavior and risk of sexually transmitted infections such as HIV. Studies of gay and bisexual men have found an association between MDMA use and high-risk sexual behavior (Urbina and Jones 2004). MDMA has been implicated in a growing number of deaths, both in Australia and in other countries (Henry, Jeffreys, and Dawling 1992; Solowij 1993; White, Bochner, and Irvine 1997). Although the reasons for extreme reactions have yet to be clearly determined, deaths have most often been attributed to hyperthermia when MDMA was used at dance venues. A combination of sustained exertion, high ambient temperatures, and inadequate fluid replacement appears to compound the effect of MDMA on thermoregulatory mechanisms, causing a rapid and fatal rise in body temperature (Topp and others 1998). Some deaths have been attributed to excessive water consumption, which causes cerebral edema (Cook 1996; Matthai and others 1996). REFERENCES Ahmed, M. A., T. Zafar, H. Brahmbhatt, G. Imam, S. ul Hassan, J. C. Bareta, and S. A. Strathdee. 2003. “HIV/AIDS Risk Behaviors and Correlates of Injection Drug Use among Drug Users in Pakistan.” Journal of Urban Health 80 (2): 321–29. Ainsworth, M., C. Beyrer, and A. Soucat. 2003. “AIDS and Public Policy: The Lessons and Challenges of ‘Success’ in Thailand.” Health Policy 64 (1): 13–37. American Psychiatric Association. 1994. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association. Anglin, M. D. 1988. “The Efficacy of Civil Commitment in Treating Narcotic Drug Addiction.” In Compulsory Treatment of Drug Abuse: Research and Clinical Practice, ed. C. G. Leukefeld and F. M. Tims, 8–34. Rockville, MD: National Institute on Drug Abuse. Anglin, M. D., M. L. Brecht, and E. Maddahain. 1989. “Pre-treatment Characteristics and Treatment Performance of Legally Coerced versus Voluntary Methadone Maintenance Admissions.” Criminology 27 (3): 537–57. Anglin, M. D., and B. Perrochet. 1998. “Drug Use and Crime: A Historical Review of Research Conducted by the UCLA Drug Abuse Research Center.” Substance Use and Misuse 33 (9): 1871–914. Angrist, B., G. Sathananthan, S. Wilk, and S. Gershon. 1974. “Amphetamine Psychosis: Behavioural and Biochemical Aspects.” Journal of Psychiatric Research 11: 13–23. Anthony, J. C., L. Warner, and R. Kessler. 1994. “Comparative Epidemiology of Dependence on Tobacco, Alcohol, Controlled 924 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 126 Substances, and Inhalants: Basic Findings from the National Comorbidity Survey.” Experimental and Clinical Psychopharmacology 2 (3): 244–68. Arseneault, L., M. Cannon, R. Poulton, R. Murray, A. Caspi, and T. E. Moffitt. 2002. “Cannabis Use in Adolescence and Risk for Adult Psychosis: Longitudinal Prospective Study.” British Medical Journal 325 (7374): 1212–13. Atlani, L., M. Carael, J. B. Brunet, T. Frasca, and N. Chaika. 2000. “Social Change and HIV in the Former USSR: The Making of a New Epidemic.” Social Science and Medicine 50 (11): 1547–56. Australian Institute of Health and Welfare. 1999. “1998 National Drug Strategy Household Survey: First Results.” Drug Statistics Series 1, Australian Institute of Health and Welfare, Canberra. Bachman, J. G., K. N. Wadsworth, P. M. O’Malley, L. D. Johnston, and J. Schulenberg. 1997. Smoking, Drinking, and Drug Use in Young Adulthood: The Impacts of New Freedoms and New Responsibilities. Mahwah, NJ: Lawrence Erlbaum. Baker, A., T. G. Boggs, and T. J. Lewin. 2001. “Randomized Controlled Trial of Brief Cognitive-Behavioral Interventions among Regular Users of Amphetamine.” Addiction 96: 1279–87. Ball, J. C., J. W. Shaffer, and D. N. Nurco. 1983. “The Day-to-Day Criminality of Heroin Addicts in Baltimore: A Study in the Continuity of Offence Rates.” Drug and Alcohol Dependence 12 (2): 119–42. Bammer, G., W. van den Brink, P. Gschwend, V. Hendriks, and J. Rehm. 2003. “What Can the Swiss and Dutch Trials Tell Us about the Potential Risks Associated with Heroin Prescribing?” Drug and Alcohol Review 22 (3): 363–71. Barnett, P. G. 1999. “The Cost-Effectiveness of Methadone Maintenance as a Health Care Intervention.” Addiction 94 (4): 479–88. Barnett, P. G., G. S. Zaric, and M. L. Brandeau. 2001. “The CostEffectiveness of Buprenorphine Maintenance Therapy for Opiate Addiction in the United States.” Addiction 96 (9): 1267–78. Bates, M. N., and T. A. Blakely. 1999. “Role of Cannabis in Motor Vehicle Crashes.” Epidemiologic Reviews 21: 222–32. Beardsley, P., and T. Kelly. 1999. “Acute Effects of Cannabis on Human Behavior and Central Nervous System Functions.” In The Health Effects of Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart, 127–265. Toronto, ON: Centre for Addiction and Mental Health. Beatty, W. W., V. M. Katzung, V. J. Moreland, and S. J. Nixon. 1995. “Neuropsychological Performance of Recently Abstinent Alcoholics and Cocaine Abusers.” Drug and Alcohol Dependence 37: 247–53. Beck, J. 1990. “The Public Health Implications of MDMA Use.” In Ecstasy: The Clinical, Pharmacological, and Neurotoxicological Effects of the Drug MDMA, ed. S. J. Peroutka. Boston: Kluwer. Beck, J., and M. Rosenbaum. 1994. Pursuit of Ecstasy: The MDMA Experience. Albany: State University of New York Press. Belenko, S. 2002. “The Challenges of Conducting Research in Drug Treatment Court Settings.” Substance Use and Misuse 37 (12–13): 1635–64. Bell, D. S. 1973. “The Experimental Reproduction of Amphetamine Psychosis.” Archives of General Psychiatry 29 (1): 35–40. Beyrer, C. 2003. “Hidden Epidemic of Sexually Transmitted Diseases in China: Crisis and Opportunity.” Journal of the American Medical Association 289 (10): 1303–5. Beyrer, C., M. H. Razak, K. Lisam, J. Chen, W. Lui, and X. F. Yu. 2000. “Overland Heroin Trafficking Routes and HIV-1 Spread in South and South-East Asia.” AIDS 14 (1): 75–83. Boot, B., I. McGregor, and W. D. Hall. 2000. “MDMA (Ecstasy) Neurotoxicity: Assessing and Communicating the Risks.” Lancet 355 (9217): 1818–21. Brands, B., B. Sproule, and J. Marshman. 1998. Drugs and Drug Abuse. 3rd ed. Toronto, ON: Addiction Research Foundation. Brecht, M. L., M. D. Anglin, and J. C. Wang. 1993. “Treatment Effectiveness for Legally Coerced versus Voluntary Methadone Maintenance Clients.” American Journal of Drug and Alcohol Abuse 19 (1): 89–106. Brown, T. T., and A. S. Dobs. 2002. “Endocrine Effects of Marijuana.” Journal of Clinical Pharmacology 42 (Suppl. 11): 90S–96S. Bruneau, J., S. B. Brogly, M. W. Tyndall, F. Lamothe, and E. L. Franco. 2004. “Intensity of Drug Injection as a Determinant of Sustained Injection Cessation among Chronic Drug Users: The Interface with Social Factors and Service Utilization.” Addiction 99 (6): 727–37. Budney, A. J., S. T. Higgins, K. J. Radonovich, and P. L. Novy. 2000. “Adding Voucher-Based Incentives to Coping Skills and Motivational Enhancement Improves Outcomes during Treatment for Marijuana Dependence.” Journal of Consulting and Clinical Psychology 68 (6): 1051–61. Budney, A. J., and B. A. Moore. 2002. “Development and Consequences of Cannabis Dependence.” Journal of Clinical Pharmacology 42 (Suppl. 11): 28S–33S. Caplehorn, J. R., S. Dalton, M. C. Cluff, and A. M. Petrenas. 1994. “Retention in Methadone Maintenance and Heroin Addicts’ Risk of Death.” Addiction 89 (2): 203–9. Cartwright, W. S. 2000. “Cost-Benefit Analysis of Drug Treatment Services: Review of the Literature.” Journal of Mental Health Policy and Economics 3 (1): 11–26. Caulkins, J. P., C. P. Rydell, S. M. S. Everingham, J. R. Chiesa, and S. Bushway. 1999. An Ounce of Prevention, a Pound of Uncertainty: The Cost-Effectiveness of School-Based Drug Prevention Programs. Santa Monica, CA: Rand. Central Committee on the Treatment of Heroin Addicts. 2002. Medical Co-prescription of Heroin: Two Randomized Controlled Trials. Utrecht, Netherlands: Central Committee on the Treatment of Heroin Addicts. Chaisson, R. E., P. Bacchetti, D. Osmond, B. Brodie, M. A. Sande, and A. R. Moss. 1989. “Cocaine Use and HIV Infection in Intravenous Drug Users in San Francisco.” Journal of the American Medical Association 261 (4): 561–65. Chen, K., and D. B. Kandel. 1995. “The Natural History of Drug Use from Adolescence to the Mid-Thirties in a General Population Sample.” American Journal of Public Health 85 (1): 41–47. Chesher, G. 1995. “Cannabis and Road Safety: An Outline of Research Studies to Examine the Effects of Cannabis on Driving Skills and Actual Driving Performance.” In The Effects of Drugs (Other Than Alcohol) on Road Safety, ed. Parliament of Victoria Road Safety Committee, 67–96. Melbourne, Australia: Road Safety Committee. Chesher, G., and W. D. Hall. 1999. “Effects of Cannabis on the Cardiovascular and Gastrointestinal Systems.” In The Health Effects of Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart, 435–58. Toronto, ON: Centre for Addiction and Mental Health. Chirgwin, K., J. A. DeHovitz, S. Dillon, and W. M. McCormack. 1991.“HIV Infection, Genital Ulcer Disease, and Crack Cocaine Use among Patients Attending a Clinic for Sexually Transmitted Diseases.” American Journal of Public Health 81 (12): 1576–69. Coffin, P. O., S. Galea, J. Ahern, A. C. Leon, D. Vlahov, and K. Tardiff. 2003. “Opiate, Cocaine and Alcohol Combinations in Accidental Drug Overdose Deaths in New York City, 1990–1998.” Addiction 98: 739–47. Cohen, J. 2003. “Asia: The Next Frontier for HIV/AIDS.” Science 301 (5640): 1650–63. Collins, D., and H. Lapsley. 1996. The Social Costs of Drug Abuse in Australia in 1988 and 1992. Canberra: Australian Government Publishing Service. Cook, T. M. 1996. “Cerebral Oedema after MDMA (‘Ecstasy’) and Unrestricted Water Intake.” British Medical Journal 313: 689. Copeland, J., W. Swift, R. Roffman, and R. Stephens. 2001. “A Randomized Controlled Trial of Brief Cognitive-Behavioral Interventions for Illicit Opiate Abuse | 925 ©2006 The International Bank for Reconstruction and Development / The World Bank 127 Cannabis Use Disorder.” Journal of Substance Abuse Treatment 21 (2): 55–64. Covi, L., I. D. Montoya, J. Hess, and N. Kreiter. 1994. “Double-Blind Comparison of Desipramine and Placebo for Treatment of Cocaine Dependence.” Clinical Pharmacology and Therapeutics 55: 132. Crandall, C. G., W. Vongpatanasin, and R. G. Victor. 2002. “Mechanism of Cocaine-Induced Hyperthermia in Humans.” Annals of Internal Medicine 136: 785–91. Cretzmeyer, M., M. V. Sarrazin, D. L. Huber, R. I. Block, and J. A. Hall. 2003. “Treatment of Methamphetamine Abuse: Research Findings and Clinical Directions.” Journal of Substance Abuse Treatment 24 (3): 267–77. Crits-Christoph, P., L. Siqueland, J. Blaine, A. Frank, L. Luborsky, L. S. Onken, and others. 1999. “Psychosocial Treatments for Cocaine Dependence: National Institute on Drug Abuse Collaborative Cocaine Treatment Study.” Archives of General Psychiatry 56 (6): 493–502. Crofts, N., G. Costigan, P. Narayanan, J. Gray, J. Dorabjee, B. Langkham, and others. 1998. “Harm Reduction in Asia: A Successful Response to Hidden Epidemics—The Asian Harm Reduction Network.” AIDS 12 (Suppl. B): S109–15. Darke, S., A. Baker, J. Dixon, A. Wodak, and N. Heather. 1992. “Drug Use and HIV Risk-Taking Behaviour among Clients in Methadone Maintenance Treatment.” Drug and Alcohol Dependence 29: 263–68. Darke, S., and W. D. Hall. 1995. “Levels and Correlates of Polydrug Use among Heroin Users and Regular Amphetamine Users.” Drug and Alcohol Dependence 39: 231–35. ———. 1997. “The Distribution of Naloxone to Heroin Users.” Addiction 92 (9): 1195–99. ———. 2003. “Heroin Overdose: Research and Evidence-Based Intervention.” Journal of Urban Health 80 (2): 189–200. Darke, S., S. Kaye, and J. Ross. 1999. “Transitions between the Injection of Heroin and Amphetamines.” Addiction 94: 1803–11. Darke, S., S. Kaye, and L. Topp. 2002. “Cocaine Use in New South Wales, Australia, 1996–2000: 5-Year Monitoring of Trends in Price, Purity, Availability, and Use from the Illicit Drug Reporting System.” Drug and Alcohol Dependence 6: 81–88. Darke, S., and J. Ross. 2002. “Suicide among Heroin Users: Rates, Risk Factors, and Methods.” Addiction 97 (11): 1383–94. Darke, S., J. Ross, J. Cohen, J. Hando, and W. D. Hall. 1995. “Injecting and Sexual Risk-Taking Behavior among Regular Amphetamine Users.” AIDS Care 7: 17–24. Darke, S., J. Ross, and W. D. Hall. 1995. “Benzodiazepine Use among Injecting Heroin Users.” Medical Journal of Australia 162: 645–47. Darke, S., J. Ross, J. Hando, W. D. Hall, and L. Degenhardt. 2000. Illicit Drug Use in Australia: Epidemiology, Use Patterns, and Associated Harm. National Drug Strategy Monograph 43. Canberra: Department of Health and Aged Care. Darke, S., I. Topp, H. Kaye, and W. Hall. 2002. “Heroin Use in New South Wales, Australia, 1996–2000: Five-Year Monitoring of Trends in Price, Purity, Availability, and Use from the Illicit Drug Reporting System (IDRS).” Addiction 97 (2): 179–86. Darke, S., and D. Zador. 1996. “Fatal Heroin ‘Overdose’: A Review.” Addiction 91 (12): 1765–72. Day, C., L. Degenhardt, S. Gilmour, and W. D. Hall. 2004. “Effects of Reduction in Heroin Supply on Injecting Drug Use: Analysis of Data from Needle and Syringe Programmes.” British Medical Journal 329 (7463): 428–29. Degenhardt, L., B. Barker, and L. Topp. 2004. “Patterns of Ecstasy Use in Australia: Findings from a National Household Survey.” Addiction 99 (2): 187–95. Degenhardt, L., C. Day, and W. D. Hall, eds. 2004. The Causes, Course, and Consequences of the Heroin Shortage in Australia. Adelaide, Australia: National Drug Law Enforcement Research Fund. Degenhardt, L., W. D. Hall, and M. Lynskey. 2003. “Testing Hypotheses about the Relationship between Cannabis Use and Psychosis.” Drug and Alcohol Dependence 71 (1): 37–48. Degenhardt, L., W. D. Hall, M. Warner-Smith, and M. Lynskey. 2004.“Illicit Drug Use.” In Comparative Risk Assessment, vol. 1, ed. M. Ezzati, A. Lopez, and C. Murray, 1109–76. Geneva: World Health Organization. Derlet, R. W., P. Rice, B. Z. Horowitz, and R. V. Lord. 1989. “Amphetamine Toxicity: Experience with 127 Cases.” Journal of Emergency Medicine 7 (2): 157–61. Desjalais, D. C., J. Wenston, S. R. Friedman, J. L. Sotheran, R. Maslansky, and M. Marmor. 1992. “Crack Cocaine Use in a Cohort of Methadone Maintenance Patients.” Journal of Substance Abuse Treatment 9: 319–25. Dolan, K., J. Kimber, C. Fry, J. Fitzgerald, D. MacDonald, and F. Trautmann. 2000. “Drug Consumption Facilities in Europe and the Establishment of Supervised Injecting Centres in Australia.” Drug and Alcohol Review 19 (3): 337–46. Dole, V. P., J. W. Robinson, J. Oracca, E. Towns, P. Searcy, and E. Caine. 1969. “Methadone Treatment of Randomly Selected Addicts.” New England Journal of Medicine 280 (25): 1372–75. Dorabjee, J., and L. Samson. 2000. “A Multi-Centre Rapid Assessment of Injecting Drug Use in India.” International Journal of Drug Policy 11 (1–2): 99–112. Doran, C. M., M. Shanahan, R. P. Mattick, R. Ali, J. White, and J. Bell. 2003. “Buprenorphine versus Methadone Maintenance: A Cost-Effectiveness Analysis.” Drug and Alcohol Dependence 71 (3): 295–302. Downing, J. 1986. “The Psychological and Physiological Effects of MDMA on Normal Volunteers.” Journal of Psychoactive Drugs 18: 335–40. Ellison, G. 1992. “Continuous Amphetamine and Cocaine Have Similar Neurotoxic Effects in Lateral Habenular and Fasciculus Retroflexus.” Brain Research 598: 353–56. EMCDDA (European Monitoring Centre for Drugs and Drug Addiction. 2002. Annual Report on the State of the Drugs Problem in the European Union, 2001. Lisbon: EMCDDA. ———. 2003. Annual Report 2003: The State of the Drugs Problem in the European Union and Norway. Lisbon: EMCDDA. English, D., G. Hulse, E. Milne, C. Holman, and C. Bower. 1997. “Maternal Cannabis Use and Birth Weight: A Meta-Analysis.” Addiction 92: 1553–60. Farre, M.,A. Mas, M. Torrens,V. Moreno, and J. Cami. 2002.“Retention Rate and Illicit Opioid Use during Methadone Maintenance Interventions: A Meta-Analysis.” Drug and Alcohol Dependence 65 (3): 283–90. Fergusson, D. M., L. J. Horwood, and M. Lynskey. 1998. “Child and Adolescent Psychiatric Disorders.” In Mental Health in New Zealand from a Public Health Perspective, eds. P. Ellis and S. Collings, 136–63. Wellington: Ministry of Health. Fergusson, D. M., L. J. Horwood, and K. Northstone. 2002. “Maternal Use of Cannabis and Pregnancy Outcome.” British Journal of Obstetrics and Gynaecology 109 (1): 21–27. Fergusson, D. M., L. J. Horwood, and N. Swain-Campbell. 2002. “Cannabis Use and Psychosocial Adjustment in Adolescence and Young Adulthood.” Addiction 97 (9): 1123–35. Fields, J. Z., L. Wichlinski, G. E. Drucker, K. Engh, and J. H. Gordon. 1991. “Long-Lasting Dopamine Receptor Up-Regulation in AmphetamineTreated Rats Following Amphetamine Neurotoxicity.” Pharmacology, Biochemistry, and Behavior 40 (4): 881–86. Fischer, B., E. Haydon, J. Rehm, M. Krajden, and J. Reimer. 2004. “Injection Drug Use and the Hepatitis C Virus: Considerations for a Targeted Treatment Approach—The Case Study of Canada.” Journal of Urban Health 81 (3): 428–47. Flynn, P. M., G. W. Joe, K. M. Broome, D. D. Simpson, and B. S. Brown. 2003. “Looking Back on Cocaine Dependence: Reasons for Recovery.” American Journal on Addictions 12 (5): 398–411. 926 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 128 Fortney, J., B. Booth, M. Zhang, J. Humphrey, and E. Wiseman. 1998. “Controlling for Selection Bias in the Evaluation of Alcoholics Anonymous as Aftercare Treatment.” Journal of Studies on Alcohol 59 (6): 690–707. Frawley, P., and J. L. Smith. 1992. “One-Year Follow-up after Multimodal Inpatient Treatment for Cocaine and Methamphetamine Dependencies.” Journal of Substance Abuse Treatment 9: 271–86. Fried, P. A., and A. R. Smith. 2001. “A Literature Review of the Consequences of Prenatal Marihuana Exposure: An Emerging Theme of a Deficiency in Aspects of Executive Function.” Neurotoxicology and Teratology 23 (1): 1–11. Fuller, C. M., D. C. Ompad, S. Galea, Y. Wu, B. Koblin, and D. Vlahov. 2004. “Hepatitis C Incidence: A Comparison between Injection and Noninjection Drug Users in New York City.” Journal of Urban Health 81 (1): 20–24. Galai, N., M. Safaeian, D. Vlahov, A. Bolotin, and D. D. Celentano. 2003. “Longitudinal Patterns of Drug Injection Behavior in the Alive Study Cohort, 1988–2000: Description and Determinants.” American Journal of Epidemiology 158 (7): 695–704. Galduroz, J. C., A. R. Noto, S. A. Nappo, and E. L. Carlini. 2003. “First Household Survey on Drug Abuse in São Paulo, Brazil, 1999: Principal Findings.” São Paulo Medical Journal 121 (6): 231–37. Galloway, G., J. A. Newmeyer, T. Knapp, S. Stalcup, and D. Smith. 1994. “Imipramine for the Treatment of Cocaine and Methamphetamine Dependence.” Journal of Addictive Diseases 13 (4): 201–16. Gawin, F. H., and E. H. Ellinwood Jr. 1988. “Cocaine and Other Stimulants. Actions, Abuse, and Treatment.” New England Journal of Medicine 318 (18): 1173–82. Gawin, F. H., H. D. Kleber, and R. Byck. 1989. “Desipramine Facilitation of Initial Cocaine Abstinence.” Archives of General Psychiatry 46: 117–21. Gearing, F. R., and M. D. Schweitzer. 1974. “An Epidemiologic Evaluation of Long-Term Methadone Maintenance Treatment for Heroin Addiction.” American Journal of Epidemiology 100 (2): 101–12. Gerstein, D., and H. Harwood. 1990. Treating Drug Problems. Vol. 1 of A Study of Effectiveness and Financing of Public and Private Drug Treatment Systems. Washington, DC: National Academy Press. Gerstein, D., H. Harwood, and N. Suter. 1994. Evaluating Recovery Services: The California Drug and Alcohol Treatment Assessment. Sacramento: State of California Health and Welfare Agency, Department of Alcohol and Drug Programs. Gibson, A. E., C. M. Doran, J. R. Bell, A. Ryan, and N. Lintzeris. 2003. “A Comparison of Buprenorphine Treatment in Clinic and Primary Care Settings: A Randomised Trial.” Medical Journal of Australia 179 (1): 38–42. Hall, J., M. Vaughan, T. Vaughn, R. I. Block, D. L. Huber, and A. Schut. 1999. “Iowa Case Management for Rural Drug Abuse: Preliminary Results.” Case Management Journal 1: 232–43. Hall, W. D. 1997. “The Role of Legal Coercion in the Treatment of Offenders with Alcohol and Heroin Problems.” Australian and New Zealand Journal of Criminology 30 (2): 103–20. Hall, W. D., J. Bell, and J. Carless. 1993. “Crime and Drug Use among Applicants for Methadone Maintenance.” Drug and Alcohol Dependence 31: 123–29. Hall, W. D., L. J. Degenhardt, and M. T. Lynskey. 1999. “Opioid Overdose Mortality in Australia, 1964–1997: Birth-Cohort Trends.” Medical Journal of Australia 171 (1): 34–37. Hall, W. D., J. Hando, S. Darke, and J. Ross. 1996. “Psychological Morbidity and Route of Administration among Amphetamine Users in Sydney, Australia.” Addiction 91 (1): 81–87. Hall, W. D., L. Johnston, and N. Donnelly. 1999. “Epidemiology of Cannabis Use and Its Consequences.” In The Health Effects of Cannabis, ed. H. Kalant, W. Corrigal, W. D. Hall, and R. Smart, 69–125. Toronto, ON: Centre for Addiction and Mental Health. Hall, W. D., and M. Lynskey. Forthcoming. “Testing Hypotheses about the Relationship between Cannabis Use and the Use of Other Illicit Drugs.” Drug and Alcohol Review. Hall, W. D., M. Lynskey, and L. Degenhardt. 1999. Heroin Use in Australia: Its Impact on Public Health and Public Order. NDARC Monograph 42. Sydney, Australia: National Drug and Alcohol Research Centre. Hall, W. D., and D. MacPhee. 2002. “Cannabis Use and Cancer.” Addiction 97 (3): 243–47. Hall, W. D., and R. P. Mattick. 2000. “Is Ultra-Rapid Opioid Detoxification a Viable Option in the Treatment of Opioid Dependence?” CNS Drugs 14 (4): 251–55. Hall, W. D., and R. L. Pacula. 2003. Cannabis Use and Dependence: Public Health and Public Policy. Melbourne, Ausralia: Cambridge University Press. Hall, W. D., J. E. Ross, M. T. Lynskey, M. G. Law, and L. J. Degenhardt. 2000. “How Many Dependent Heroin Users Are There in Australia?” Medical Journal of Australia 173 (10): 528–31. Hall, W. D., J. Ward, and R. Mattick. 1998. “The Effectiveness of Methadone Maintenance Treatment 1: Heroin Use and Crime.” In Methadone Maintenance Treatment and Other Opioid Replacement Therapies, ed. J. Ward, R. Mattick, and W. D. Hall, 17–57. Amsterdam: Harwood Academic. Hamers, F. F., and A. M. Downs. 2003. “HIV in Central and Eastern Europe.” Lancet 361 (9362): 1035–44. Goldschmidt, P. G. 1976. “A Cost-Effectiveness Model for Evaluating Health Care Programs: Application to Drug Abuse Treatment.” Inquiry 13 (1): 29–47. Hando, J., and W. D. Hall. 1994. “HIV Risk-Taking Behavior among Amphetamine Users in Sydney, Australia.” Addiction 89 (1): 79–85. Goldstein, A., and J. Herrera. 1995. “Heroin Addicts and Methadone Treatment in Albuquerque: A 22-Year Follow-up.” Drug and Alcohol Dependence 40 (2): 139–50. Hando, J., L. Topp, and W. D. Hall. 1997. “Amphetamine-Related Harms and Treatment Preferences of Regular Amphetamine Users in Sydney, Australia.” Drug and Alcohol Dependence 46 (1-2): 105–13. Gossop, M., J. Marsden, and D. Stewart. 1998. NTORS at One Year: Changes in Substance Use, Health, and Criminal Behaviour One Year after Intake. London: Department of Health. Hawkins, J. D., R. F. Catalano, and J. Y. Miller. 1992. “Risk and Protective Factors for Alcohol and Other Drug Problems in Adolescence and Early Adulthood: Implications for Substance Abuse Prevention.” Psychological Bulletin 112 (1): 64–105. Gossop, M., J. Marsden, D. Stewart, C. Edwards, P. Lehmann, A. Wilson, and G. Segar. 1997. “The National Treatment Outcome Research Study in the United Kingdom: Six-Month Follow-up Outcomes.” Psychology of Addictive Behaviors 11 (4): 324–37. Grella, C. E., M. D. Anglin, and S. E. Wugalter. 1995.“Cocaine and Crack Use and HIV Risk Behaviors among High-Risk Methadone Maintenance Clients.” Drug and Alcohol Dependence 37 (1): 15–21. Gronbladh, L., L. Ohlund, and L. Gunne. 1990. “Mortality in Heroin Addiction: Impact of Methadone Treatment.” Acta Psychiatrica Scandinavica 82 (3): 223–27. Henry, J. A., K. L. Jeffreys, and S. Dawling. 1992. “Toxicity and Deaths from 3,4-Methylenedioxymethamphetamine (‘Ecstasy’).” Lancet 340: 384–87. Hepatitis C Virus Projections Working Group. 1998. Estimates and Projections of the Hepatitis C Virus Epidemic in Australia. Sydney, Australia: National Centre in HIV Epidemiology and Clinical Research. Hien, N. T., L. T. Giang, P. N. Binh, W. Deville, E. J. van Ameijden, and I. Wolffers. 2001. “Risk Factors of HIV Infection and Needle Sharing among Injecting Drug Users in Ho Chi Minh City, Vietnam.” Journal of Substance Abuse 13 (1–2): 45–58. Illicit Opiate Abuse | 927 ©2006 The International Bank for Reconstruction and Development / The World Bank 129 Higgins, S. T., A. J. Budney, W. K. Bickel, F. E. Foerg, R. Donham, and G. J. Badger. 1994. “Incentives Improve Outcome in Outpatient Behavioral Treatment of Cocaine Dependence.” Archives of General Psychiatry 51 (7): 568–76. Hoff, A. L., H. Riordan, L. Morris, V. Cestaro, M. Wieneke, R. Alpert, and G. J. Wang. 1996. “Effects of Crack Cocaine on Neurocognitive Function.” Psychiatry Research 60: 167–76. Holdcraft, L. C., and W. G. Iacono. 2004. “Cross-Generational Effects on Gender Differences in Psychoactive Drug Abuse and Dependence.” Drug and Alcohol Dependence 74 (2): 147–58. Hollister, L. 1992. “Marijuana and Immunity.” Journal of Psychoactive Drugs 24: 159–64. Hser, Y. I., D. Anglin, and K. Powers. 1993. “A 24-Year Follow-up of California Narcotics Addicts.” Archives of General Psychiatry 50 (7): 577–84. Hubbard, R. L., J. J. Collins, J. V. Rachal, and E. R. Cavanaugh. 1988. “The Criminal Justice Client in Drug Abuse Treatment.” In Compulsory Treatment of Drug Abuse: Research and Clinical Practice, ed. C. G. Leukefeld and F. M. Tims, 57–80. Rockville, MD: National Institute on Drug Abuse. Hubbard, R. L., S. G. Craddock, and J. Anderson. 2003. “Overview of 5-Year Followup Outcomes in the Drug Abuse Treatment Outcome Studies (Datos).” Journal of Substance Abuse Treatment 25 (3): 125–34. Hubbard, R. L., M. Marsden, J. V. Rachal, H. Harwood, E. Cavanaugh, and H. Ginzburg. 1989. Drug Abuse Treatment: A National Study of Effectiveness. Chapel Hill: University of North Carolina Press. Huber, A., W. Ling, S. Shoptaw, V. Gulati, P. Brethen, and R. Rawson. 1997. “Integrating Treatments for Methamphetamine Abuse: A Psychosocial Perspective.” Journal of Addictive Diseases 16 (4): 41–50. Hulse, G. K., D. R. English, E. Milne, and C. D. Holman. 1999. “The Quantification of Mortality Resulting from the Regular Use of Illicit Opiates.” Addiction 94 (2): 221–29. Jaffe, J. 1985. “Drug Addiction and Drug Abuse.” In The Pharmacological Basis of Therapeutics, eds. A. Gilman, L. Goodman and F. Murad, 532–81. New York: Macmillan. Johnston, L. D., P. M. O’Malley, and J. G. Bachman. 1994a. National Survey Results on Drug Use from the Monitoring the Future Study, 1975–1993: College Students and Young Adults. Rockville, MD: National Institute on Drug Abuse. ———. 1994b. National Survey Results on Drug Use from the Monitoring the Future Study, 1975–1993: Secondary School Students. Rockville, MD: National Institute on Drug Abuse,. Kaldor, J., H. Lapsley, R. P. Mattick, D. Weatherburn, and A. Wilson. 2003. Final Report on the Evaluation of the Sydney Medically Supervised Injecting Centre. Sydney, Australia: Medically Supervised Injecting Centre Evaluation Committee. Kandel, D. B., ed. 2002. Stages and Pathways of Drug Involvement: Examining the Gateway Hypothesis. New York: Cambridge University Press. Kantak, K. M. 2003. “Vaccines against Drugs of Abuse: A Viable Treatment Option?” Drugs 63 (4): 341–52. Kaye, S., and S. Darke. 2000. “A Comparison of the Harms Associated with the Injection of Heroin and Amphetamines.” Drug and Alcohol Dependence 58 (1–2): 189–95. ———. 2003. “Non-Fatal Cocaine Overdose and Other Adverse Events among Injecting and Non-Injecting Cocaine Users.” NDARC Technical Report 170, National Drug and Alcohol Research Centre, University of New South Wales, Sydney, Australia. Kelly, J. A., and Y. A. Amirkhanian. 2003. “The Newest Epidemic: A Review of HIV/AIDS in Central and Eastern Europe.” International Journal of STD and AIDS 14 (6): 361–71. Kimber, J., K. Dolan, I. van Beek, D. Hedrich, and H. Zurhold. 2003. “Drug Consumption Facilities: An Update since 2000.” Drug and Alcohol Review 22 (2): 227–33. Klein, T. 1999. “Cannabis and Immunity.” In The Health Effects of Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart, 347–73. Toronto, ON: Centre for Addiction and Mental Health. Kosten, T. R., H. D. Kleber, and C. Morgan. 1989. “Role of Opioid Antagonists in Treating Intravenous Cocaine Abuse.” Life Science 44: 887–92. Kouri, E. M., and H. G. Pope. 2000. “Abstinence Symptoms during Withdrawal from Chronic Marijuana Use.” Experimental and Clinical Psychopharmacology 8 (4): 483–92. Kreek, M. J. 1997. “Opiate and Cocaine Addictions: Challenge for Pharmacotherapies.” Pharmacology, Biochemistry, and Behavior 57 (3): 551–69. Lange, R. A., and L. D. Hillis. 2001. “Cardiovascular Complications of Cocaine Use.” New England Journal of Medicine 345: 351–58. Langendam, M. W., G. H. van Brussel, R. A. Coutinho, and E. J. van Ameijden. 2001. “The Impact of Harm-Reduction-Based Methadone Treatment on Mortality among Heroin Users.” American Journal of Public Health 91 (5): 774–80. Leshner, A. I. 1997. “Addiction Is a Brain Disease, and It Matters.” Science 278 (5335): 45–47. Levin, F. R., D. McDowell, S. M. Evans, E. Nunes, E. Akerele, S. Donovan, and S. K. Vosburg. 2004. “Pharmacotherapy for Marijuana Dependence: A Double-Blind, Placebo-Controlled Pilot Study of Divalproex Sodium.” American Journal on Addictions 13 (1): 21–32. Lima, M. S., A. A. Reisser, B. G. Soares, and M. Farrell. 2003. “Antidepressants for Cocaine Dependence.” Cochrane Database of Systematic Reviews (2): CD002950. [PMID: 12804445]. Loxley, W., and A. Marsh. 1991. “Nodding and Speeding: Age and Injecting Drug Use in Perth.” National Centre for Research into the Prevention of Drug Abuse, Curtin University of Technology, Perth, Australia. Lynskey, M., and W. D. Hall. 2000. “The Effects of Adolescent Cannabis Use on Educational Attainment: A Review.” Addiction 96 (3): 433–43. Macdonald, S., K. Anglin-Bodrug, R. E. Mann, P. Erickson, A. Hathaway, M. Chipman, and M. Rylett. 2003.“Injury Risk Associated with Cannabis and Cocaine Use.” Drug and Alcohol Dependence 72 (2): 99–115. Macleod, J., R. Oakes, A. Copello, I. Crome, M. Egger, M. Hickman, and others. 2004. “Psychological and Social Sequelae of Cannabis and Other Drug Use by Young People: A Systematic Review of Longitudinal, General Population Studies.” Lancet 363 (9421): 1579–88. Maddux, J. F., and D. P. Desmond. 1992. “Methadone Maintenance and Recovery from Opioid Dependence.” American Journal of Drug and Alcohol Abuse 18 (1): 63–74. Majewska, M. D., ed. 1996. Neurotoxicity and Neuropathology Associated with Cocaine Abuse. NIDA Research Monograph 163. Rockville, MD: U.S. Department of Health and Human Services. Makkai, T., and I. McAllister. 1998. Patterns of Drug Use in Australia, 1985–95. Canberra: Australian Government Publishing Service. Manschreck, T. C., J. A. Laughery, C. C. Weisstein, D. Allen, B. Humblestone, M. Neville, and others. 1988. “Characteristics of Freebase Cocaine Psychosis.” Yale Journal of Biology and Medicine 61 (2): 115–22. Manski, C. F., J. V. Pepper, and C. V. Petrie, eds. 2001. Informing America’s Policy on Illegal Drugs: What We Don’t Know Keeps Hurting Us. Washington, DC: National Academy Press. Marsch, L. A. 1998. “The Efficacy of Methadone Maintenance Interventions in Reducing Illicit Opiate Use, HIV Risk Behavior, and Criminality: A Meta-Analysis.” Addiction 93 (4): 515–32. Marselos, M., and P. Karamanakos. 1999. “Mutagenicity, Developmental Toxicity and Carcinogeneity of Cannabis.” Addiction Biology 4 (1): 5–12. 928 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 130 Marsh, K., G. Joe, D. Simpson, and W. Lehman. 1990. “Treatment History.” In Opioid Addiction and Treatment: A 12-Year Follow-Up, eds. D. Simpson and S. Sells, 137–56. Malabar, FL: Krieger. Mathers, C., T. Vos, and C. Stephenson. 1999. The Burden of Disease and Injury in Australia. Canberra: Australian Institute of Health and Welfare. Matthai, S. M., J. A. Sills, D. C. Davidson, and D. Alexandrou. 1996. “Cerebral Oedema after Ingestion of MDMA (‘Ecstasy’) and Unrestricted Intake of Water.” British Medical Journal 312: 1359. Mattick, R. P., and S. Darke. 1995. “Drug Replacement Treatments: Is Amphetamine Substitution a Horse of a Different Color?” Drug and Alcohol Review 14: 389–94. Mattick, R. P., E. Digiusto, C. M. Moran, S. O’Brien, M. Shanahan, J. Kimber, N. Henderson, C. Breen, J. Shearer, J. Gates, A. Shakeshaft, NEPOD Trial Investigators. 2001. “National Evaluation of Pharmacotherapies for Opioid Dependence (NEPOD).” Canberra: Commonwealth Department of Health and Ageing. Mattick, R. P., and W. D. Hall. 1996. “Are Detoxification Programmes Effective?” Lancet 347 (8994): 97–100. Mattick, R. P., J. Kimber, C. Breen, and M. Davoli. 2003. “Buprenorphine Maintenance versus Placebo or Methadone Maintenance for Opioid Dependence.” Cochrane Database Systematic Review (2) CD002207 [PMID: 12804429]. McCance, E. F. 1997. “Overview of Potential Treatment Medications for Cocaine Dependence.” In Medication Development for the Treatment of Cocaine Dependence: Issues in Clinical Efficacy Trials, NIDA Research Monograph 175, 36–72. Rockville, MD: U.S. Department of Health and Human Services. McCann, U. D., and G. A. Ricaurte. 2000. “Drug Abuse and Dependence: Hazards and Consequences of Heroin, Cocaine, and Amphetamines.” Current Opinion in Psychiatry 13: 321–25. McGregor, C., R. Ali, P. Christie, and S. Darke. 2001. “Overdose among Heroin Users: Evaluation of an Intervention in South Australia.” Addiction Research 9 (5): 481–501. Newcomb, M. D., and P. M. Bentler. 1988. Consequences of Adolescent Drug Use: Impact on the Lives of Young Adults. Thousand Oaks, CA: Sage. Newman, R. G., and W. B. Whitehill. 1979. “Double-Blind Comparison of Methadone and Placebo Maintenance Treatments of Narcotic Addicts in Hong Kong.” Lancet 2 (8141): 485–88. Nunes, E. V. 1997. Methodologic Recommendations for Cocaine Abuse Clinical Trials: A Clinician-Researcher’s Perspective. In Medication Development for the Treatment of Cocaine Dependence: Issues in Clinical Efficacy Trials, NIDA Research Monograph 175, 73–95. Rockville, MD: U.S. Department of Health and Human Services. Office of National Drug Control Policy. 2001. Pulse Check: Trends in Drug Abuse, November 2001. Washington, DC: Executive Office of the President, Office of National Drug Control Policy. Oliveto, A., and T. Kosten. 1997. “Buprenorphine.” In New Treatments for Opioid Dependence, eds. S. Stine and T. Kosten, 25–67. New York: Guilford Press. O’Malley, S., M. Adamse, R. K. Heaton, and F. H. Gawin. 1992. “Neuropsychological Impairment in Chronic Cocaine Abusers.” American Journal of Drug and Alcohol Abuse 18: 131–44. Perneger, T. V., F. Giner, M. del Rio, and A. Mino. 1998. “Randomised Trial of Heroin Maintenance Programme for Addicts Who Fail in Conventional Drug Treatments.” British Medical Journal 317 (7150): 13–8. Petry, N. M., J. Tedford, M. Austin, C. Nich, K. M. Carroll, and B. J. Rounsaville. 2004. “Prize Reinforcement Contingency Management for Treating Cocaine Users: How Low Can We Go, and with Whom?” Addiction 99 (3): 349–60. Platt, J. J. 1997. Cocaine Addiction: Theory, Research and Treatment. Cambridge, MA: Harvard University Press. Porter, L., A. Arif, and W. Curran. 1986. The Law and the Treatment of Drug and Alcohol Dependent Persons: A Comparative Study of Existing Legislation. Geneva: World Health Organization. McKetin, R., and R. P. Mattick. 1997. “Attention and Memory in Illicit Amphetamine Users.” Drug and Alcohol Dependence 48 (3): 235–42. Pottieger, A. E., P. A. Tressell, J. A. Inciardi, and T. A. Rosales. 1992. “Cocaine Use Patterns and Overdose.” Journal of Psychoactive Drugs 24: 399–410. ———. 1998. “Attention and Memory in Illicit Amphetamine Users: Comparison with Non-Drug-Using Controls.” Drug and Alcohol Dependence 50 (2): 181–4. Ramaekers, J. G., G. Berghaus, M. van Laar, and O. H. Drummer. 2004. “Dose Related Risk of Motor Vehicle Crashes after Cannabis Use.” Drug and Alcohol Dependence 73 (2): 109–19. Mendelson, J. H., and N. K. Mellon. 1996. “Management of Cocaine Abuse and Dependence.” New England Journal of Medicine 334 (15): 965–72. Rawson, R., M. Anglin, and W. Ling. 2002. “Will the Methamphetamine Problem Go Away?” Journal of Addictive Diseases 21: 5–19. Merline, A. C., P. M. O’Malley, J. E. Schulenberg, J. G. Bachman, and L. D. Johnston. 2004. “Substance Use among Adults 35 Years of Age: Prevalence, Adulthood Predictors, and Impact of Adolescent Substance Use.” American Journal of Public Health 94 (1): 96–102. Rehm, J., P. Gschwend, T. Steffen, F. Gutzwiller, A. Dobler-Mikola, and A. Uchtenhagen. 2001. “Feasibility, Safety, and Efficacy of Injectable Heroin Prescription for Refractory Opioid Addicts: A Follow-Up Study.” Lancet 358 (9291): 1417–23. Mesquita, F., A. Kral, A. Reingold, I. Haddad, M. Sanches, G. Turienzo, and others. 2001. “Overdoses among Cocaine Users in Brazil.” Addiction 96: 1809–13. Reid, G., and N. Crofts. 2000. “Rapid Assessment of Drug Use and HIV Vulnerability in South-East and East Asia.” International Journal of Drug Policy 11 (1–2): 113–24. Mittleman, M. A., R. A. Lewis, M. Maclure, J. B. Sherwood, and J. E. Muller. 2001. “Triggering Myocardial Infarction by Marijuana.” Circulation 103: 2805–9. Rhodes, T., A. Ball, G. V. Stimson, Y. Kobyshcha, C. Fitch, V. Pokrovsky, and others. 1999. “HIV Infection Associated with Drug Injecting in the Newly Independent States, Eastern Europe: The Social and Economic Context of Epidemics.” Addiction 94 (9): 1323–36. Murphy, L. 1999. “Cannabis Effects on Endocrine and Reproductive Function.” In The Health Effects of Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart, 375–400. Toronto, ON: Centre for Addiction and Mental Health. National Centre in HIV Epidemiology and Clinical Research. 1998. HIV/AIDS and Related Diseases in Australia: Annual Surveillance Report 1998. Sydney, Australia: National Centre in HIV Epidemiology and Clinical Research. National Research Council Committee on Clinical Evaluation of Narcotic Antagonists. 1978. “Clinical Evaluation of Naltrexone Treatment of Opiate-Dependent Individuals.” Archives of General Psychiatry 35: 355–40. Rhodes, T., L. Mikhailova, A. Sarang, C. M. Lowndes, A. Rylkov, M. Khutorskoy, and A. Renton. 2003. “Situational Factors Influencing Drug Injecting, Risk Reduction, and Syringe Exchange in Togliatti City, Russian Federation: A Qualitative Study of Micro Risk Environment.” Social Science and Medicine 57 (1): 39–54. Robinson, T. E., and J. B. Becker. 1986. “Enduring Changes in Brain and Behavior Produced by Chronic Amphetamine Administration: A Review and Evaluation of Animal Models of Amphetamine Psychosis.” Brain Research 396 (2): 157–98. Roozen, H. G., J. J. Boulogne, M. W. van Tulder, W. van den Brink, C. A. De Jong, and A. J. Kerkhof. 2004. “A Systematic Review of the Effectiveness Illicit Opiate Abuse | 929 ©2006 The International Bank for Reconstruction and Development / The World Bank 131 of the Community Reinforcement Approach in Alcohol, Cocaine, and Opioid Addiction.” Drug and Alcohol Dependence 74 (1): 1–13. Ruan, Y. H., K. X. Hong, S. Z. Liu, Y. X. He, F. Zhou, G. M. Qin, and others. 2004. “Community-Based Survey of HCV and HIV Coinfection in Injection Drug Abusers in Sichuan Province of China.” World Journal of Gastroenterology 10 (11): 1589–93. SAMHSA (Substance Abuse and Mental Health Services Adminstration). 2001. Summary of Findings from the 2000 National Household Survey on Drug Abuse. Rockville, MD: Office of Applied Statistics, SAMHSA. ———. 2002. Results from the 2001 Household Survey on Drug Abuse. Vol. 1 of Summary of National Findings. Rockville, MD: Office of Applied Statistics, SAMHSA. ———. 2003. Overview of Findings from the 2002 National Survey on Drug Use and Health Office. Rockville, MD: Office of Applied Statistics, SAMHSA. Satel, S. L., and W. S. Edell. 1991.“Cocaine-Induced Paranoia and Psychosis Proneness.” American Journal of Psychiatry 148 (12): 1708–11. Schoenbaum, E. E., D. Hartel, P. A. Selwyn, R. S. Klein, K. Davenny, M. Rogers, and others. 1989. “Risk Factors for Human Immunodeficiency Virus Infection in Intravenous Drug Users.” New England Journal of Medicine 321 (13): 874–49. Shepard, D., M. J. Larson, and N. G. Hoffmann. 1999. “Cost-Effectiveness of Substance Abuse Services: Implications for Public Policy.” Psychiatric Clinics of North America 22 (2): 385–400. Shepard, D., G. Strickler, J. McKay, D. Bury-Maynard, H. Yeom, C. Love, and others. Forthcoming. “Cost-Effectiveness of Self-Help for Controlling Substance Use: Controlling for Self-Selection.” Shoptaw, S., R. Rawson, M. McCann, and J. Obert. 1994. “The Matrix Model of Outpatient Stimulant Abuse Treatment: Evidence of Efficacy.” In Experimental Therapeutics in Addiction Medicine, ed. S. Magura and S. Rosenblum, 129–41. Binghamton, NY: Haworth Press. Sidney, S. 2002. “Cardiovascular Consequences of Marijuana Use.” Journal of Clinical Pharmacology 42 (11 Suppl.): 64S–70S. Sidney, S., J. E. Beck, I. S. Tekawa, C. P. Quesenberry, and G. D. Friedman. 1997. “Marijuana Use and Mortality.” American Journal of Public Health 87 (4): 585–90. Silva de Lima, M., B. Garcia de Oliveira Soares, A. Alves Pereira Reisser, and M. Farrell. 2002. “Pharmacological Treatment of Cocaine Dependence: A Systematic Review.” Addiction 97: 931–49. Simpson, D. S., and H. J. Friend. 1988. “Legal Status and Long-Term Outcomes for Addicts in the DARP Followup Project.” In Compulsory Treatment of Drug Abuse: Research and Clinical Practice, ed. C. G. Leukefeld and F. M. Tims, 81–96. Rockville, MD: National Institute on Drug Abuse. Simpson, D. S., G. W. Joe, W. E. K. Lehman, and S. B. Sells. 1986. “Addiction Careers: Etiology, Treatment, and 12-Year Follow-up Outcomes.” Journal of Drug Issues 16 (1): 107–21. Simpson, D. S., and S. Sells. 1982. “Effectiveness of Treatment for Drug Abuse: An Overview of the DARP Research Program.” Advances in Alcohol and Substance Abuse 2 (1): 7–29. ———, eds. 1990. Opioid Addiction and Treatment: A 12-Year Follow-up. Malabar, FL: Krieger. Smiley, A. 1999. “Marijuana: On Road and Driving Simulator Studies.” In The Health Effects of Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart, 171–91. Toronto, ON: Centre for Addiction and Mental Health. Soares, B. G., M. S. Lima, A. A. Reisser, and M. Farrell. 2003. “Dopamine Agonists for Cocaine Dependence.” Cochrane Database of Systematic Reviews (2): CD003352 [PMID: 12804461]. Solowij, N. 1993. “Ecstasy (3,4-Methalenedioxymethamphetamine).” Current Opinion in Psychiatry 6: 411–15. ———. 1998. Cannabis and Cognitive Functioning. Cambridge, U.K.: Cambridge University Press. Solowij, N., W. D. Hall, and N. Lee. 1992. “Recreational MDMA Use in Sydney: A Profile of ‘Ecstasy’ Users and Their Experiences with the Drug.” British Journal of Addiction 87: 1161–72. Solowij, N., R. S. Stephens, R. A. Roffman, T. Babor, R. Kadden, M. Miller, and others. 2002. “Cognitive Functioning of Long-Term Heavy Cannabis Users Seeking Treatment.” Journal of the American Medical Association 287 (9): 1123–31. Spooner, C. and W. D. Hall. 2002. “Public Policy and the Prevention of Substance Use Disorders.”Current Opinion in Psychiatry 15 (3): 235–39. Spooner, C., W. D. Hall, and R. P. Mattick. 2001.“An Overview of Diversion Strategies for Australian Drug-Related Offenders.” Drug and Alcohol Review 20 (3): 281–94. Sporer, K. A. 2003. “Strategies for Preventing Heroin Overdose.” British Medical Journal 326 (7386): 442–44. Srisurapanont, M., N. Jarusuraisin, and P. Kittirattanapaiboon. 2003. “Treatment for Amphetamine Dependence and Abuse.” Cochrane Database of Systematic Reviews (4): CD003022 [PMID: 11687171]. Stephens, R. S., R. A. Roffman, and E. E. Simpson. 1994. “Treating Adult Marijuana Dependence—A Test of the Relapse Prevention Model.” Journal of Consulting and Clinical Psychology 62 (1): 92–99. Strang, J., S. Darke, W. D. Hall, M. Farrell, and R. Ali. 1996. “Heroin Overdose: The Case for Take-Home Naloxone: Home-Based Supplies of Naloxone Would Save Lives.” British Medical Journal 312 (7044): 1435–36. Strathdee, S. A., T. Zafar, H. Brahmbhatt, A. Baksh, and S. ul Hassan. 2003. “Rise in Needle Sharing among Injection Drug Users in Pakistan during the Afghanistan War.” Drug and Alcohol Dependence 71 (1): 17–24. Szasz, T. 1985. Ceremonial Chemistry: The Ritual Persecution of Drugs, Addicts, and Pushers. Holmes Beach, FL: Learning Publications. Tashkin, D. P. 1999. “Effects of Cannabis on the Respiratory System.” In The Health Effects of Cannabis, ed. H. Kalant, W. Corrigall, W. D. Hall, and R. Smart, 311–45. Toronto, ON: Centre for Addiction and Mental Health. Taylor, D. R., D. M. Fergusson, B. J. Milne, L. J. Horwood, T. E. Moffitt, M. R. Sears, and R. Poulton. 2002. “A Longitudinal Study of the Effects of Tobacco and Cannabis Exposure on Lung Function in Young Adults.” Addiction 97 (8): 1055–61. Tonigan, J., R. Toscova, and W. Miller. 1996.“Meta-Analysis of the Literature on Alcoholics Anonymous: Sample and Study Characteristics Moderate Findings.” Journal of Studies on Alcohol 57 (1): 65–72. Tonigan, J. S., G. J. Connors, and W. R. Miller. 2003. “Participation and Involvement in Alcoholics Anonymous.” In Treatment Matching in Alcoholism, ed. T. F. Babor and F. K. Del Boca, 184–204. Cambridge, UK: Cambridge University Press. Toomey, R., M. J. Lyons, S. A. Eisen, H. Xian, S. Chantarujikapong, L. J. Seidman, and others. 2003. “A Twin Study of the Neuropsychological Consequences of Stimulant Abuse.” Archives of General Psychiatry 60 (3): 303–10. Topp, L., J. Hando, L. Degenhardt, P. Dillon, A. Roche, and N. Solowij. 1998. Ecstasy Use in Australia. NDARC Monograph 39. Sydney, Australia: National Drug and Alcohol Research Centre, University of New South Wales. Torrens, M., L. San, J. M. Peri, and J. M. Olle. 1991. “Cocaine Abuse among Heroin Addicts in Spain.” Drug and Alcohol Dependence 27 (1): 29–34. Uchtenhagen, A., F. Gutzwiller, and A. Dobler-Mikola. 1998. Medical Prescription of Narcotics Research Programme: Final Report of the Principal Investigators. Zurich, Switzerland: Institut für Sozial- und Präventivmedizin der Universität Zurich. UNAIDS (Joint United Nations Programme on HIV/AIDS) and WHO (World Health Organization). 2002. AIDS Epidemic Update December 930 | Disease Control Priorities in Developing Countries | Wayne Hall, Chris Doran, Louisa Degenhardt, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 132 2002. Geneva: Joint United Nations Programme on HIV/AIDS and World Health Organization. UNDCP (United Nations Drug Control Programme). 1997. World Drug Report. Oxford, U.K.: Oxford University Press. United Nations Commission on Narcotic Drugs. 2000. “World Situation with Regard to Drug Abuse, in Particular among Children and Youth.” Vienna, United Nations Commission on Narcotic Drugs. UNODC (United Nations Office on Drugs and Crime). 2004. World Drug Report. Vienna: UNODC. UNODCCP (United Nations Office for Drug Control and Crime Prevention). 2002. Global Illicit Drug Trends, 2002. New York: UNODCCP. ———. 2003. Global Illicit Drug Trends, 2003. New York: UNODCCP. Urbina, A., and K. Jones. 2004. “Crystal Methamphetamine, Its Analogues, and HIV Infection: Medical and Psychiatric Aspects of a New Epidemic.” Clinical Infectious Diseases 38 (6): 890–4. Uuskula, A., A. Kalikova, K. Zilmer, L. Tammai, and J. DeHovitz. 2002. “The Role of Injection Drug Use in the Emergence of Human Immunodeficiency Virus Infection in Estonia.” International Journal of Infectious Diseases 6 (1): 23–27. van Beek, I., R. Dwyer, and A. Malcolm. 2001.“Cocaine Injecting: The Sharp End of Drug-Related Harm!” Drug and Alcohol Review 20: 333–42. van den Brink,W., and J. M. van Ree. 2003.“Pharmacological Treatments for Heroin and Cocaine Addiction.” European Neuropsychopharmacology 13 (6): 476–87. Vanichseni, S., B. Wongsuwan, K. Choopanya, and K. Wongpanich. 1991. “A Controlled Trial of Methadone Maintenance in a Population of Intravenous Drug Users in Bangkok: Implications for Prevention of HIV.” International Journal of the Addictions 26 (12): 1313–20. van Os, J., M. Bak, M. Hanssen, R. V. Bijl, R. de Graaf, and H. Verdoux. 2002. “Cannabis Use and Psychosis: A Longitudinal Population-Based Study.” American Journal of Epidemiology 156 (4): 319–27. Vasica, G., and C. C. Tennant. 2002. “Cocaine Use and Cardiovascular Complications.” Medical Journal of Australia 177 (5): 260–62. Verheyden, S. L., J. A. Henry, and H. V. Curran. 2003. “Acute, Sub-Acute, and Long-Term Subjective Consequences of ‘Ecstasy’ (MDMA) Consumption in 430 Regular Users.” Human Psychopharmacology 18 (7): 507–17. Vincent, N., J. Shoobridge, A. Ask, S. Allsop, and R. Ali. 1998. “Physical and Mental Health Problems in Amphetamine Users from Metropolitan Adelaide, Australia.” Drug and Alcohol Review 17: 187–95. Vlahov, D., C. L. Wang, N. Galai, J. Bareta, S. H. Mehta, S. A. Strathdee, and K. E. Nelson. 2004. “Mortality Risk among New Onset Injection Drug Users.” Addiction 99 (8): 946–54. Walsh, G. W., and R. E. Mann. 1999. “On the High Road: Driving under the Influence of Cannabis in Ontario.” Canadian Journal of Public Health-Revue Canadienne De Santé Publique 90 (4): 260–63. Ward, J., W. D. Hall, and R. P. Mattick. 1998. Methadone Maintenance Treatment and Other Opioid Replacement Therapies. Amsterdam: Harwood Academic. Warner-Smith, M., S. Darke, M. Lynskey, and W. D. Hall. 2001. “Heroin Overdose: Causes and Consequences.” Addiction 96 (8): 1113–25. White, J. M., F. Bochner, and R. J. Irvine. 1997. “The Agony of ‘Ecstasy’: How Can We Avoid More ‘Ecstasy’-Related Deaths?” Medical Journal of Australia 166: 117–18. WHO (World Health Organization). 2003. The World Health Report 2003: Shaping the Future. Geneva: WHO. WHO (World Health Organization) Programme on Substance Abuse. 1997. Cannabis: A Health Perspective and Research Agenda. Geneva: WHO, Division of Mental Health and Prevention of Substance Abuse. Wild, T. C., A. B. Roberts, and E. L. Cooper. 2002. “Compulsory Substance Abuse Treatment: An Overview of Recent Findings and Issues.” European Addiction Research 8 (2): 84–93. Wilkins, C., K. Bhatta, M. Pledger, and S. Casswell. 2003. “Ecstasy Use in New Zealand: Findings from the 1998 and 2001 National Drug Surveys.” New Zealand Medical Journal 116 (1171): U383. Xie, X., J. Rehm, E. Single, and L. Robson. 1996. The Economic Costs of Alcohol, Tobacco, and Illicit Drug Abuse in Ontario: 1992. Toronto, ON: Addiction Research Foundation. Yacoubian, G. S. Jr. 2003a. “Correlates of Ecstasy Use among Students Surveyed through the 1997 College Alcohol Study.” Journal of Drug Education 33 (1): 61–69. ———. 2003b. “Tracking Ecstasy Trends in the United States with Data from Three National Drug Surveillance Systems.” Journal of Drug Education 33 (3): 245–58. Yacoubian, G. S. Jr., C. Boyle, C. A. Harding, and E. A. Loftus. 2003. “It’s a Rave New World: Estimating the Prevalence and Perceived Harm of Ecstasy and Other Drug Use among Club Rave Attendees.” Journal of Drug Education 33(2): 187–96. Yu, X. F., J. Chen, Y. Shao, C. Beyrer, and S. Lai. 1998. “Two Subtypes of HIV-1 among Injection-Drug Users in Southern China.” Lancet 351 (9111): 1250. Zammit, S., P. Allebeck, S. Andreasson, I. Lundberg, and G. Lewis. 2002. “Self Reported Cannabis Use as a Risk Factor for Schizophrenia in Swedish Conscripts of 1969: Historical Cohort Study.” British Medical Journal 325 (7374): 1199–201. Zaric, G. S., P. G. Barnett, and M. L. Brandeau. 2000. “HIV Transmission and the Cost-Effectiveness of Methadone Maintenance.” American Journal of Public Health 90 (7): 1100–11. Zhang, Z. F., H. Morgenstern, M. R. Spitz, D. P. Tashkin, G. P. Yu, J. R. Marshall, and others. 1999. “Marijuana Use and Increased Risk of Squamous Cell Carcinoma of the Head and Neck.” Cancer Epidemiology Biomarkers and Prevention 8 (12): 1071–78. Illicit Opiate Abuse | 931 ©2006 The International Bank for Reconstruction and Development / The World Bank 133 ©2006 The International Bank for Reconstruction and Development / The World Bank 134 Part Three Strengthening Health Systems • Strengthening Public Health Services • Strengthening Personal Health Services • Capacity Strengthening and Management Reform ©2006 The International Bank for Reconstruction and Development / The World Bank 135 ©2006 The International Bank for Reconstruction and Development / The World Bank 136 Chapter 53 Public Health Surveillance: A Tool for Targeting and Monitoring Interventions Peter Nsubuga, Mark E. White, Stephen B. Thacker, Mark A. Anderson, Stephen B. Blount, Claire V. Broome, Tom M. Chiller, Victoria Espitia, Rubina Imtiaz, Dan Sosin, Donna F. Stroup, Robert V. Tauxe, Maya Vijayaraghavan, and Murray Trostle What gets measured gets done. —Anonymous Public health surveillance is the ongoing systematic collection, analysis, and interpretation of data, closely integrated with the timely dissemination of these data to those responsible for preventing and controlling disease and injury (Thacker and Berkelman 1988). Public health surveillance is a tool to estimate the health status and behavior of the populations served by ministries of health, ministries of finance, and donors. Because surveillance can directly measure what is going on in the population, it is useful both for measuring the need for interventions and for directly measuring the effects of interventions. The purpose of surveillance is to empower decision makers to lead and manage more effectively by providing timely, useful evidence. Increasingly, top managers in ministries of health and finance in developing countries and donor agencies are recognizing that data from effective surveillance systems are useful for targeting resources and evaluating programs. The HIV and severe acute respiratory syndrome (SARS) epidemics underscored the critical role of surveillance in protecting individual nations and the global community. For example, in 2005, China rapidly began to expand its surveillance and response capacity through its Field Epidemiology Training Program (FETP); Brazil and Argentina chose to use World Bank loans to develop surveillance capacity; and the U.S. Agency for International Development (USAID) redesigned its surveillance strategy to focus on the use of data to improve public health interventions (USAID 2005). Additionally, the guidelines for implementing the 2004 draft revised International Health Regulations require World Health Organization (WHO) member states to have key persons and core capacities in surveillance (http:// www.who.int/csr/ihr/howtheywork/faq/en/#draft). Just as decision makers require competent, motivated economists to provide quality technical analyses, they also need competent staff members to provide scientifically valid surveillance information and communicate the results as information for action. Competent epidemiologists and surveillance staff members are not a luxury in developing countries; they are a necessity for rational planning, implementation, and intervention (Narasimhan and others 2004). DEFINITIONS AND BASIC CONCEPTS In this chapter, we use the following definitions: • Indicator: a measurable factor that allows decision makers to estimate objectively the size of a health problem and monitor the processes, the products, or the effects of an intervention on the population (for example, the number of new cases of diarrhea, the proportion of children fully 997 ©2006 The International Bank for Reconstruction and Development / The World Bank 137 • • • • • • • immunized in a district, or the percentage of high school students who report that they smoke at least one cigarette a day). Active surveillance: a system employing staff members to regularly contact heath care providers or the population to seek information about health conditions. Active surveillance provides the most accurate and timely information, but it is also expensive. Passive surveillance: a system by which a health jurisdiction receives reports submitted from hospitals, clinics, public health units, or other sources. Passive surveillance is a relatively inexpensive strategy to cover large areas, and it provides critical information for monitoring a community’s health. However, because passive surveillance depends on people in different institutions to provide data, data quality and timeliness are difficult to control. Routine health information system: a passive system in which regular reports about diseases and programs are completed by public health staff members, hospitals, and clinics. Health information and management system: a passive system by which routine reports about financial, logistic, and other processes involved in the administration of the public health and clinical systems can be used for surveillance. Categorical surveillance: an active or passive system that focuses on one or more diseases or behaviors of interest to an intervention program. These systems are useful for program managers. However, they may be inefficient at the district or local level, at which staff may need to fill out multiple forms on the same patient (that is, the HIV program, the tuberculosis program, the sexually transmitted infections program, and the Routine Health Information System). At higher levels, allocating the few competent surveillance experts to one program may leave other programs underserved, and reconciling the results of different systems to establish the nation’s official estimates may be difficult. Integrated surveillance: a combination of active and passive systems using a single infrastructure that gathers information about multiple diseases or behaviors of interest to several intervention programs (for example, a health facility–based system may gather information on multiple infectious diseases and injuries). Managers of diseasespecific programs may be evaluated on the results of the integrated system and should be stakeholders. Even when an integrated system is functioning well, program managers may continue to maintain categorical systems to collect additional disease-specific data and control the quality of the information on which they are evaluated. This practice may lead to duplication and inefficiency. Syndromic surveillance: an active or passive system that uses case definitions that are based entirely on clinical features without any clinical or laboratory diagnosis (for example, collecting the number of cases of diarrhea rather than cases of cholera, or “rash illness” rather than measles). Because syndromic surveillance is inexpensive and is faster than systems that require laboratory confirmation, it is often the first kind of surveillance begun in a developing country. However, because of the lack of specificity (for example, a “rash illness” could be anything from the relatively minor rubella to devastating hemorrhagic fevers), reports require more investigation from higher levels. Also an increase in one disease causing a syndrome may mask an epidemic of another (for example, rotavirus diarrhea decreases at the same time cholera increases). In the specialized area of surveillance for biologic terrorism, syndromic surveillance refers to active surveillance of syndromes that may be caused by potential agents used by biologic terrorists and sometimes refers to alternative measures such as increases in the use of over-the-counter drugs or increases in calls to emergency departments. • Behavioral risk factor surveillance system (BRFSS): an active system of repeated surveys that measure behaviors that are known to cause disease or injury (for example, tobacco or alcohol use, unprotected sex, or lack of physical exercise). Because the aim of many intervention program strategies is to prevent disease by preventing unhealthy behavior, these surveys provide a direct measure of their effect in the population, often long before the anticipated health effects are expected. These surveys are useful for providing timely measures of program effectiveness for both communicable and noncommunicable disease interventions. OBJECTIVES OF SURVEILLANCE SYSTEMS Public health surveillance provides the scientific and factual database essential to informed decision making and appropriate public health action. The key objective of surveillance is to provide information to guide interventions. The public health objectives and actions needed to make successful interventions determine the design and implementation of surveillance systems. For example, if the objective is to prevent the spread of epidemics of acute infectious diseases, such as SARS, managers need to intervene quickly to stop the spread of disease. Therefore, they need a surveillance system that provides rapid early warning information from clinics and laboratories. In contrast, chronic diseases and health-related behaviors change slowly. Managers typically monitor the effect of programs to change risky behaviors such as tobacco smoking or chronic diseases once a year or even less often. A surveillance system to measure the population effects of a tuberculosis control program might provide information only every one to five years— for example, through a series of demographic and health surveys. The principle is that different public health objectives and the actions required to reach them require different information 998 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 138 Box 53.1 Different Objectives, Different Data, Different Systems Objective Action Data System Detect epidemics Monitor intervention programs Monitor impact of policy change Monitor health system Epidemic response Program monitoring Health policy Resource allocation Early warning information Program indicators Health indicators Administrative data Active surveillance Health information Health information Health information and management Source: Nsubuga, Eseko, and others 2002. systems. The type of action that can be taken, when or how often that action needs to be taken, what information is needed to take or monitor the action, and when or how frequently the information is needed should determine the type of surveillance or health information system (box 53.1). Table 53.1 Utility of Surveillance Data Immediate detection of Epidemics Newly emerging health problems Changes in health practices Changes in antibiotic resistance Changes in distribution of population at risk for disease PRINCIPLES AND USES OF SURVEILLANCE Periodic dissemination for Foege, Hogan, and Newton (1976) state that “the reason for collecting, analyzing, and disseminating information on a disease is to control that disease. Collection and analysis should not be allowed to consume resources if action does not follow.” The fundamental principle of public health surveillance is that the surveillance should be designed and implemented to provide valid (true) information to decision makers in a timely manner at the lowest possible cost. Because managers are unlikely to need to make interventions to address small differences between areas, sacrificing precision makes sense to improve timeliness and save resources that can be used for public health interventions. The utility of surveillance data can be viewed as immediate, annual, and archival, on the basis of the public health actions that can be taken (table 53.1; Thacker and Stroup 1998b). Estimating magnitude of a health problem, including cost Assessing control activities Setting research priorities Determining risk factors for disease Facilitating planning Monitoring risk factors Monitoring changes in health practices Documenting distribution and spread of disease and injury Stored information for Describing natural history of diseases Facilitating epidemiologic and laboratory research Validating use of preliminary data Setting research priorities Documenting distribution and spread of disease and injury ESTABLISHING AND MAINTAINING A SURVEILLANCE SYSTEM What is worth doing is worth doing right. Managers who decide to use public health surveillance as a management tool must recognize that they will need to commit political support and human and financial resources. As with every health system, competent, motivated health workers need to be found or trained and provided with career paths and supervision. After a manager decides to create a surveillance system, there are six steps to establishing the system. Because the system must adapt constantly to changes in the population and the physical Source: Adapted from Thacker and Stroup 1998b, 65. and social environment, these steps are linked continuously (figure 53.1; Thacker and Stroup 1998a). ANALYSIS AND DISSEMINATION OF SURVEILLANCE DATA Surveillance information is analyzed by time, place, and person. Knowledgeable technical personnel should review data regularly to ensure their validity and to identify information Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 999 ©2006 The International Bank for Reconstruction and Development / The World Bank 139 Establish goals Develop case definitions Select appropriate personnel Acquire tools and clearances for collection, analysis, and dissemination Implement surveillance system Evaluate surveillance activities Source: Adapted from Thacker and Stroup 1998a, 119. Figure 53.1 Elements in Establishing and Maintaining a Surveillance System of use to top managers. Simple tables and graphs are most useful for summarizing and presenting data. Timely dissemination of data to those who make policy and implement intervention programs is critical to the usefulness of surveillance data. The rapidly evolving field of public health informatics, which deals with collection, classification, storage, retrieval, analysis, and presentation of large amounts of health data, offers the potential for truly integrated public health surveillance based on data standardization, a communications infrastructure, and policies on data access and sharing. Surveillance will benefit by incorporating a systematic approach to standards for data content. For example, the U.S. Centers for Disease Control and Prevention (CDC) has used standardsbased systems to support automatic electronic reporting of diagnostic laboratory results of notifiable diseases, thereby both increasing the number of cases reported and receiving results more rapidly (Effler and others 1999). Use of data standards facilitates comparability of surveillance information over time (for example, measurement of effect of program interventions), across different surveillance approaches (for example, facility-based reporting compared with sample surveys), and across countries and regions. To be credible, a standard should be developed through an open, participatory process, by an internationally recognized accredited standardsdevelopment organization that is also capable of long-term maintenance and evolution of the standard. Public health data needs extend into multiple areas beyond clinical medicine (for example, environmental toxins, unintentional injury, and food safety). One international standard computer program used in many countries’ information systems is Epi Info, an epidemiology surveillance and biostatistics program widely used around the world for the analysis of surveillance data (http://www.cdc. gov/epiinfo). CDC created, maintains, and distributes Epi Info at no cost to users. SURVEILLANCE AS A COMPONENT OF NATIONAL PUBLIC HEALTH SYSTEMS WHO and the World Bank cite public health surveillance as an essential function of a public health system (World Bank 2001). When linked to policy and program units, surveillance information improves the efficiency and effectiveness of health services by targeting interventions and documenting their effect on the population. A critical challenge in the health sector in developing countries is to ensure quality and effectiveness of surveillance and public health response in an environment of decentralization. National-level program and surveillance system managers may lose control of the quality and timeliness of locally collected data. This situation can be avoided by training local decision makers in how to use information to meet their needs and negotiating with them over the core information collected by each district local unit. National-level managers or donors can also improve information quality by sponsoring national surveillance scientific and quality assurance networks, linking funding to provision of adequate data, and performing periodic surveys to confirm the results of local reporting. If the responsibility for implementing programs is devolved to local managers, then national-level managers need only a few summary indicators, rather than the detailed information they may be used to. District or local managers tend to prefer integrated systems to minimize filling out redundant forms. Donors need surveillance data to target and evaluate their investments. If they perceive weakness in the national system, they may create parallel nongovernmental surveillance systems to gather data directly to meet their needs. These systems invariably pay workers more than government jobs do, so the most competent people in the government system may leave to work for the parallel system. Although this system meets donors’ short-term needs, it invariably weakens government systems. Parallel systems may weaken the very ministries that they are meant to help and may not be sustainable after external funding ends. Therefore, parallel systems are inherently inequitable and should be used only as a last resort. 1000 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 140 Public health Response Data collection Data generation hemisphere Data use hemisphere Interpretation Analysis Source: Authors. Figure 53.2 Surveillance and Response Conceptual Framework SURVEILLANCE AS A TOOL TO IMPROVE PUBLIC HEALTH Managers need focused, timely, scientifically sound information that provides evidence to make decisions on interventions for improving the health of the population in their jurisdiction. Simply collecting data and presenting them are not enough. Using Surveillance Information for Evidence-Based Decisions A major gap in promoting effective surveillance lies between the production of data and the ability to convert those data into usable information and then initiate the appropriate public health action. Surveillance and response can be described in terms of a data generation hemisphere and a data use hemisphere (USAID 2005). The data generation hemisphere is the traditional view of surveillance, whereas the data use hemisphere is the public health response that begins with interpretation of the data from the surveillance system (figure 53.2). Substantial attention and the accompanying resources in surveillance have been devoted to the prompt and complete production of surveillance data. Although developing countries experience weaknesses in both hemispheres, more attention is needed to creating and strengthening the local capacity within developing countries to identify and manage effective responses to disease outbreaks and public health conditions of national and international concern. In some disease-specific programs, this capacity has to be imported through short-term expatriate assistance. Even when local capacity is developed, it is often specific to the disease program, making transfer of skills to other areas problematic. The failure to develop this indigenous capacity has limited the ability of developing countries to build national surveillance systems that respond to both international public health threats and local health concerns. This capability is essential to the sustained development of countries. Role of Field Epidemiologists in Providing Evidence Developed countries have constructed their public health and disease control strategies by using the principles of field epidemiology. Developing countries need to build and sustain human capacity in field epidemiology. Strengthened field epidemiologic capacity can serve a country in the following specific areas: • providing a response to acute problems • providing the scientific basis for program and policy decisions • implementing disease surveillance systems • supporting national health planning • making resource allocation decisions • allocating the human capacity base for national health priorities. Specific competencies that should be developed include, but are not limited to, the ability to accomplish the following: • design, implement, and evaluate surveillance for a health event • identify and assess an actual public health problem • design and conduct a scientific investigation • analyze and interpret data from an investigation • recommend logical and practical public health actions after the analysis and interpretation of data Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1001 ©2006 The International Bank for Reconstruction and Development / The World Bank 141 • be proficient in all aspects of diseases of public health importance (for example, HIV and AIDS, sexually transmitted diseases, malaria, tuberculosis, and zoonoses). These competencies need to be tailored to the various levels of the health care system. Since 1975, CDC and WHO have collaborated with more than 30 countries to strengthen health systems and address training needs for disease detection and response in a countryspecific, flexible, and sustainable manner. More than half of the world’s population now lives in a country where surveillance, investigation, and response are carried out by staff members and trainees of FETPs and allied programs, which include the Epidemic Intelligence Service in the United States, the European Program for Intervention Epidemiology Training, and Public Health Schools without Walls (PHSWOWs). These programs generally function within central ministries of health and may not be visible outside the public health system. It can be argued that these programs provide most of the surveillance of and response to emerging infections in the world, in addition to training most of the public health workers who manage surveillance systems at the top level. FETPs are two-year courses designed to provide a ministry with a motivated, professional group of field epidemiologists with the expertise to respond to managers’ needs for evidence, perform surveillance, respond to outbreaks, and train and supervise technical personnel at other levels (White and others 2001). Other models have evolved. Guatemala’s marriage of its FETP (part of a larger, Central American FETP) with the Data for Decision Making program (Pappaioanou and others 2003) exemplifies this successful local adaptation. Data for Decision Making recruits health workers from the community and subdistrict levels to receive training in surveillance and outbreak investigation in the context of their daily work. This training is delivered as a series of linked workshops with practical fieldbased projects, providing short-term service at the local levels. The most promising graduates of the course are selected for further training in an FETP. India, with its decentralized system, complex cultural and population dynamics, and wide variance in the sophistication of public health institutions, provides another model for strengthening national surveillance. The World Bank initiated the Integrated Disease Surveillance Project, which develops the capacity of local and midlevel surveillance workers in India. Additionally, FETP graduates are recruited as the project’s surveillance officers at the state level to coordinate the surveillance activities of the hundreds of local health workers throughout the states. SELECTED SURVEILLANCE STRATEGIES Surveillance systems need to be designed and implemented to meet top management’s needs for focused, reliable, timely evidence gathered efficiently and presented effectively. Because these needs differ, depending on management’s needs, a number of different strategies have been developed. Here are some of the most useful. Sentinel Surveillance In a sentinel surveillance system,a prearranged sample of reporting sources agrees to report all cases of defined conditions, which might indicate trends in the entire target population (Birkhead and Maylahn 2000).When properly implemented, these systems offer an effective method of using limited resources and enable prompt and flexible monitoring and investigation of suspected public health problems. Examples of sentinel surveillance are networks of private practitioners reporting cases of influenza or a laboratory-based sentinel system reporting cases of certain bacterial infections among children. Sentinel surveillance is excellent for detecting large public health problems,but it may be insensitive to rare events, such as the early emergence of a new disease, because these infections may emerge anywhere in the population. Periodic Population-Based Surveys Population-based surveys can be used for surveillance if they are repeated on a regular basis (Thacker and Berkelman 1988). Examples of population-based surveys in surveillance include the BRFSS in the United States, HIV-prevalence surveys, household surveys, and the demographic and health surveys that many developing countries conduct every five years (http://www.orcmacro.com). Population-based surveys require careful attention to the methodology, particularly the use of standard protocols, supervision of interviewers, comparable sampling strategy, and standard questionnaires. These surveys require a clear definition of the target population to which the results can be generalized, and they need careful attention to the sample size, based on efficiency and the epidemiologic characteristics of the health condition under surveillance (for example, rare conditions require substantial samples). Supervising interviewers and maintaining high response rates are critical to avoid bias. Because the surveys are repetitive, population changes (caused, for example, by mortality or mobility) might bias results. Laboratory-Based Surveillance The methods used for infectious disease surveillance form a spectrum that evolves with the economic development of a country. Foodborne disease (FBD) surveillance, for example, is divided into four distinct levels of surveillance. Each level is more complex and has greater capacity for controlling and detecting disease, but it also depends on more resources and infrastructure (figure 53.3). 1002 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 142 Improving capacity and infrastructure No formal surveillancea Syndromic surveillanceb Lab-based surveillancec Integrated food-chain surveillanced Outbreak detection can occur at all levels, but ability increases along the spectrum Source: Authors. a. Without a formal surveillance, only large or unusual outbreaks can be detected. b. Syndromic surveillance is based on groups of signs or symptoms indicative of a common diagnosis, such as acute gastroenteritis. c. Laboratory-based surveillance relies on laboratory-confirmed pathogens, such as Salmonella or Shigella. d. Integrated food-chain surveillance uses data from across the food chain. Figure 53.3 Spectrum of Case-Based Foodborne Disease Surveillance For FBD, surveillance for clinical syndromes is the most common method of surveillance in the developing world. Surveillance of FBD outbreaks that are investigated by public health authorities is often a useful means of monitoring both the safety of the food supply and the activities of the public health system. Although both surveillance for clinical syndromes and outbreak surveillance will remain important, the future in FBD is in laboratory-based surveillance. If microbiologic diagnosis is sought routinely for a sample of patients with acute gastroenteritis, then surveillance based on those diagnoses is possible. For enteric bacterial pathogens such as Salmonella or Shigella, determining the serotype of the strains in central reference laboratories allows more rapid and complete identification of epidemics, which may otherwise lead to preventable death and disability. Laboratory-based surveillance systems require resources, facilities, and training. A central public health reference laboratory is essential for quality assurance and quality control and support. Such a laboratory-based system might begin with systematic referral of a sample of strains isolated at a sample of sentinel clinics, plus those strains that are part of outbreaks. A systematic sampling scheme provides better data than a more haphazard attempt at universal reporting. Regular sharing of information between the public health microbiology laboratory and epidemiologists is critical for the information to be used successfully. The utility of serotyping as an international language for Salmonella subtypes has led to its widespread adoption. In a recent survey, 61 countries reported that they used Salmonella serotyping for public health surveillance (Herikstad, Motarjemi, and Tauxe 2002). A collaborative WHO program called Global Salm-Surv promotes the use of Salmonella serotyping internationally among countries that wish to upgrade their national capacity for FBD surveillance (http://www.who.int/ salmsurv/en). Molecular subtyping is now expanding the power of laboratory-based surveillance to detect outbreaks in the background of sporadic cases by distinguishing the molecular “fingerprint” of an outbreak strain. CDC maintains PulseNet, an Internet-based network of all U.S. public health laboratories that uses a standardized genotyping method called pulsedfield-gel-electrophoresis (PFGE) as the basis for a national database of PFGE subtypes (Swaminathan and others 2001). Standardized subtyping protocols have now been developed for seven pathogens, and next-generation, gene-based technologies are under development for the future. Similar networks are developing around the world, with PulseNet Europe and PulseNet Canada already active and discussions rapidly advancing for PulseNet Asia Pacific and PulseNet Latin America. As with Salmonella serotyping itself, the global use of standard genotyping will facilitate the detection of multicontinent clusters. Integrated Disease Surveillance and Response The Integrated Disease Surveillance and Response (IDSR) strategy, first developed in Africa, links epidemiologic and laboratory data in communicable disease surveillance systems at all levels of the health system, with emphasis on integrating surveillance with response (WHO 1993, 1998). Districts were identified as a focus for strengthening efforts in collecting timely data, analyzing the collected data, and using the generated information for public health responses. The IDSR strategy is based on core activities, including case-patient detection, registration, and confirmation; reporting, analysis, use, and feedback of data; and epidemic preparedness and response (for example, outbreak investigations, contact tracing, and public health interventions). Support functions include coordination, supervision or performance evaluation, training, and resource provision for infrastructure, including communication (Nsubuga, Eseko, and others 2002). Key steps in implementing the IDSR strategy include sensitizing key health authorities and stakeholders; conducting situational analysis; preparing a strategic IDSR plan; identifying and training a motivated, competent workforce; developing national IDSR technical guidelines; implementing the plan; and monitoring and evaluating implementation to improve performance (WHO 2000b). Assessment of the existing national surveillance and response activities provides baseline data to measure progress; to identify and build consensus on the national priority communicable diseases; to identify surveillance gaps of the selected priority diseases; to document Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1003 ©2006 The International Bank for Reconstruction and Development / The World Bank 143 Table 53.2 Steps in the Development of the Philippine National Epidemic Sentinel Surveillance System Steps Data side Human capacity side 1. Identify the health problems thought to cause burden disease. Consult top managers, donors, international agencies, and experts. 2. Determine who will make interventions. Involve users in design. 3. Determine information users’ need to make interventions. Involve users in design. 4. Decide how often decision makers need reports. Involve users in design. 5. Identify who collects, tabulates, and analyzes reports and who disseminates information. Identify manager and staff to analyze, report, and enter data. 6. Design report. Involve users and staff in design. 7. Make table shells. Involve staff in design. 8. Design questionnaire. Involve staff in design. Pilot questionnaire. Involve staff in implementation and evaluation. 10. 9. Pilot data flow and analysis. Involve staff in implementation and evaluation. 11. Pilot system. Train staff in system and involve them in evaluation. 12. Run system. Involve staff in ongoing training and quality assurance monitoring. 13. Evaluate system: Was information used? Are data and analysis of good quality? Involve staff and users in design of external evaluation and in review of evaluator’s report. Source: Adapted from White and McDonnell 2000, 311. the strengths, weaknesses, and opportunities of the existing systems; and to make appropriate recommendations. The WHO Regional Office for Africa, collaborating with its partners, has prepared tools and guidelines for implementation of IDSR at the country level. Indicators to monitor the performance of the surveillance and response systems have been prepared and field tested and are now in use in Africa. Example: The Philippine National Epidemic Surveillance System In the late 1980s, the Philippine Department of Health (PDOH), relying on its integrated management information system, detected less than one outbreak per year in a population of more than 60 million people. In 1989, the PDOH designed the National Epidemic Sentinel Surveillance System, a hospitalbased sentinel surveillance system that encompasses both the flow of data and the personnel requirements needed to make the surveillance system work effectively (table 53.2). After the pilot study demonstrated promising results, the PDOH created personnel positions and a supervisory structure for sentinel physicians, nurses, and clerks in regional epidemiology and surveillance units (RESUs) integrated into the public health system. In 1995 alone, the system detected and formally investigated about 80 outbreaks, including 25 bacteriologically confirmed outbreaks of typhoid and 5 of cholera. As the Philippines developed HIV serological and behavioral risk surveillance, the RESU staff conducted surveys in their communities. By integrating surveillance functions that were based on the skills of the workforce, PDOH was able to avoid the duplications, inefficiencies, and sustainability problems of multiple vertical systems (White and McDonnell 2000). Informal Networks as Critical Elements of Surveillance Systems WHO and other agencies frequently receive telephone calls or informal reports about urgent health events. WHO publishes an informal list of these“rumors,”which allows public health workers to respond to health risks promptly rather than waiting for formal reports (http://www.who.int/csr/don/en/). The graduates of FETPs, PHSWOWs, and similar programs that provide competency-based on-the-job training in ministries of health make up one of the most important informal networks. FETPs and allied programs both train epidemiologists and provide service to their ministries of health. For example, a student in the Brazilian FETP was assigned to review routine data on patients with leishmaniasis. She noted that some patients had symptoms of heavy metal poisoning, and further study indicated that a drug being used to treat leishmaniasis was contaminated with heavy metals. The drug was reformulated, and the problem was resolved. When this study was presented at a regional meeting of the Training Programs in Public Health Interventions Network (a network of FETPs and allied training programs), other countries banned the drug until it was reformulated (CDC 2004a). Large categorical surveillance systems are expensive, and staff members might become complacent, especially if the 1004 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 144 disease under surveillance is rare. For example, the polio surveillance system for acute flaccid paralysis in the Western Hemisphere detected no cases in July 2000. A trainee from the FETP of the Dominican Republic, while investigating a case of suspected poisoning in a child, documented the first outbreak of circulating vaccine-type poliovirus in the Western Hemisphere since 1991. There were 13 confirmed cases in the Dominican Republic and 8 cases in Haiti. Her investigation led to national immunization days in both countries, which raised immunization levels and stopped the outbreak (Kew and others 2002). THE ROLE OF SURVEILLANCE IN MAJOR OUTBREAKS It seems incredible that a disease as devastating as AIDS could have spread silently to many countries over many years before it was detected and before effective control measures were implemented in the 1980s. In recent years, surveillance and response systems at all levels have been more effective at identifying and preventing spread of infectious diseases. Example: Surveillance and Global Response to SARS An epidemic of severe pneumonia of unknown etiology was detected in Guangdong province, China, in November 2002, and control measures were instituted on the basis of the way the disease spread from person to person. In February and March 2003, the disease spread to Hong Kong (China) and then to Vietnam, Singapore, Canada, and elsewhere (WHO 2003b). This new disease was named severe acute respiratory syndrome, and a preliminary case definition was established on the basis of initial epidemiologic investigations. A novel coronavirus (SARS-CoV) was identified as the causative agent in March, and mapping of the full genome was completed in April. This global pandemic ended in July 2003, as transmission was interrupted in Taiwan (China), after more than 8,000 patients in 26 countries and five continents were affected and 774 deaths were confirmed (Peiris and others 2003). WHO spearheaded the global effort to control this pandemic, working with national and subnational health workers. In China, the FETP, which was initiated in October 2001 in the China Center for Disease Control, mobilized all 20 of its trainees, and they contributed substantially to the surveillance, investigation, and control of the SARS outbreak, working with local health officials (CDC 2003a). In Canada, which had the most cases of SARS outside Asia, 8 of the 10 FETP residents were involved in the SARS outbreak. They instituted surveillance, conducted epidemiologic investigations, designed prevention and control guidelines, responded to inquiries from the media and the public, and planned and implemented epidemiologic studies (http://www.phac-aspc.gc.ca/cfep-pcet/ outbreaks_e.html). The success of this global effort to control the first new epidemic disease of the 21st century depended on a combination of open collaboration among scientists and politicians of many countries and the rapid and accurate communication of surveillance data within and among countries. Rapid global spread was recognized, and a global surveillance network was established on the basis of an agreed-upon case definition that was sufficiently specific to ensure effective reporting. Public health surveillance is critical to recognizing new cases of SARS and differentiating this disease from other causes of severe respiratory illness, especially influenza (Heymann and Rodier 2004). Ongoing research into sources in the environment as well as clinical, laboratory, and epidemiologic concerns will improve surveillance for this critical public health problem. Notably, this highly contagious disease—for which there is neither a vaccine nor a cure—was controlled by competent, dedicated health workers with access to excellent communications. SARS presented a greater challenge than smallpox, for which long incubation periods and vaccine facilitate control (Mack 2005). Although it is reassuring that national, regional, and global systems were effective in controlling SARS, there is no reason to rest on our laurels. The only certainty is that there will be more new challenges, very possibly including further outbreaks of SARS. Example: Avian Influenza in Thailand The disastrous pandemic (worldwide epidemic) of influenza in 1918 is thought to have originated from epidemics in birds, as were the influenza pandemics of 1957 and 1968 (Ungchusak and others 2005). In early 2004, large epidemics of avian influenza were recognized in birds in eight Asian countries; by November, the disease had spread from birds to 44 humans, 73 percent of whom died (Ungchusak and others 2005). This contagion sparked fears that the highly lethal avian virus might be adapting to spread from person to person, which could cause extensive health and economic damage around the world. In Thailand, avian influenza was investigated by FETP graduates and others in the Thai Ministry of Health in partnership with CDC. By applying field epidemiologic techniques supported by laboratory studies, they detected that the virus was being spread from human to human in a family. It is likely that person-to-person transmission may have occurred in other countries, where field epidemiology was not used. The Thai example is important for achieving the following: (a) raising global awareness of the potential of a global catastrophe early enough that plans can be made to avert or decrease harm and (b) demonstrating that, as with SARS, the disease could be controlled with proven field epidemiologic methods supplemented by good communications, without vaccines, Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1005 ©2006 The International Bank for Reconstruction and Development / The World Bank 145 drugs, or a high-technology laboratory or surveillance system (Mack 2005). Example: Ebola in Uganda, the Role of the PHSWOW On October 8, 2000, a second-year student in the Ugandan PHSWOW returned to Gulu district in northern Uganda for his field project. He found a hospital jammed with patients with high fevers, diarrhea, and bleeding. He diagnosed viral hemorrhagic fever. He called the Ministry of Health in Kampala, where that weekend a graduate of the PHSWOW was in charge of taking calls about epidemics. She agreed with his diagnosis and arranged for samples to be rushed to the National Institute for Virology in South Africa, the nearest WHO reference center for viral hemorrhagic fevers. When the minister of health arrived at his office the next day, the graduate briefed him. Recognizing the gravity of the situation, the minister sent the graduate to head the public health team surveillance and control team in Gulu, and the student headed the clinical team that established infection control in hospitals and treated patients. Laboratory tests quickly confirmed that the illness was Ebola hemorrhagic fever, which usually kills more than 50 percent of those infected (Heymann 2004). Public health surveillance was difficult for several reasons. Because the disease was severe and rapidly fatal, rural villagers feared that they might be stigmatized if the government knew about cases in their area. Some sought out traditional healers; others fled as soon as they realized they had been exposed, which prompted outbreaks in two other districts. Gulu was a politically unstable area, and some villages were difficult to reach because of rebel or bandit activity. The Ugandan government mobilized its military to help with case finding and invited WHO, CDC, and other international teams to assist. Patients with Ebola infection require intense nursing and medical attention to control bleeding, diarrhea, and fevers. Some patients bleed easily, and all their secretions can be highly infectious. Hospitals in Gulu were desperately short of supplies to control the spread of infection from so many patients simultaneously. In spite of this situation, Ugandan health workers selflessly cared for the sick. By January 23, 2001, a total of 425 cases had occurred, the largest Ebola outbreak recorded. Only 53 percent of the patients had died, a proportion far less than the 88 percent reported in the 1976 Ebola outbreak in the Democratic Republic of Congo (formerly Zaire) and other previous epidemics (WHO Report of an International Commission 1978). Sadly, 22 health care workers were infected. Because the team from the Ugandan Ministry of Health set up active surveillance nationwide, the other two outbreaks, started when infected Gulu residents fled to distant villages, were quickly detected and controlled. International observers commented, “National notification and surveillance efforts led to the rapid identification of these foci and to effective containment” (CDC 2001). The Ugandan Ministry of Health invested in developing competent, motivated health workers through the PHSWOW, an active partnership with Makerere University, the Rockefeller Foundation, CDC, and WHO. Both students and graduates contributed to the ministry’s ability to rapidly identify and control this dangerous epidemic. Because the minister had timely evidence, he was able to notify other countries quickly and to bring in international teams before the disease spread further. Partially because of the lessons learned from this epidemic, Uganda has become one of the leading countries in implementing the IDSR program. SURVEILLANCE FOR SPECIFIC CONDITIONS Surveillance systems are important tools for targeting, monitoring, and evaluating many health risks and interventions. Because managers need a wide variety of information for specific interventions, systems have been developed and tested to meet those needs. Environmental Public Health Surveillance Surveillance for environmental public health practice requires the collection, analysis, and dissemination of data on hazards, exposures, and health outcomes (figure 53.4; Thacker and others 1996). Health outcomes of relevance include death, disease, injury, and disability. However, relating those outcomes to specific environmental hazards and exposures is critical to Agent is a hazard Agent is present in environment Hazard surveillance Route of exposure exists Host is exposed to agent Agent reaches target tissue Exposure surveillance Agent produces adverse effect Adverse effect becomes clinically apparent Outcome surveillance Source: Thacker and others 1996. Figure 53.4 The Process of Adverse Effects and the Corresponding Surveillance 1006 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 146 environmental public health surveillance. Hazards include toxic chemical agents, physical agents, biomechanical stressors, and biologic agents that are located in air, water, soil, food, and other environmental media. Exposure surveillance is the monitoring of members of the population for the presence of an environmental agent, its metabolites, or its clinically inapparent (for example, subclinical or preclinical) effects. Four challenges complicate environmental public health surveillance. First, the ability to link specific environmental causes to adverse outcomes is limited by our poor understanding of disease processes, long lead times, inadequate measures of exposure, and multiple potential causes of disease. Second, data collected for other purposes rarely include sufficient information to meet a case definition for a condition caused by an environmental agent. Third, public alarm is often out of proportion to the hazard of concern, and sentiment will often influence public policy disproportionately to scientific information. Fourth, biologic markers will become increasingly critical elements of environmental exposure surveillance. Obtaining data on exposure, which can include estimates derived from hazard data through sophisticated modeling or direct measurements of individual exposure obtained from use of personal monitors (for example, passive air samplers), is generally impractical in developing countries. Childhood blood lead levels are the only biomonitoring data that are collected routinely in several countries, either in national surveys or from screening programs for children at high risk. Health outcome surveillance as applied to environmental public health is similar to traditional surveillance efforts. In the United States, the focus is on surveillance for birth defects; developmental disabilities (for example, cerebral palsy, autism, and mental retardation); asthma and other chronic respiratory diseases (for example, bronchitis and emphysema); cancer; and neurological diseases (for example, Parkinson’s disease, multiple sclerosis, and Alzheimer’s disease) (McGeehin, Qualters, and Niskar 2004). Other nations have different sets of priority conditions for surveillance. Disease registries, vital statistics data, annual health surveys, and administrative data systems (for example, hospital discharge data) are sources that have been used for monitoring health conditions. The challenges mentioned previously have constrained our ability in all nations, regardless of level of development, to establish and maintain effective and comprehensive environmental public health surveillance systems. As we invest in understanding the enlarging threats in the global environment, we must overcome these challenges and establish improved surveillance systems. The health of the global community depends on this investment. Injury Surveillance Injuries are a major public health problem and are among the 10 leading causes of death worldwide, killing an estimated 5 million persons each year and causing high rates of disability. People from all economic groups are at risk for injuries, but death rates caused by injury tend to be higher in developing countries (Peden, McGee, and Sharma 2002). Injury surveillance includes monitoring the incidence, causes, and circumstances of fatal and nonfatal injuries. Injuries are classified by the intention of the act into two groups: unintentional injuries and violence-related injuries. WHO (Holder, Peden, and Krug 2001) and the Pan American Health Organization (Concha-Eastman and Villveces 2001) have developed guidelines for establishing injury surveillance systems in developing countries. If the range of fatal and nonfatal injuries, as well as the risk factors that can lead to injury, are to be fully captured, surveillance systems need to be established in multiple settings. Fatal injuries can be captured by using forensic or death certificate data. A far greater number of injuries are nonfatal and can be tracked through hospital- or primary care–based systems. Systematic information on nonfatal injuries, including prevalence, incidence, and related risk behaviors can also be obtained through ongoing population-based surveys. Critical points should be addressed when planning an injury surveillance system in a developing country. First, data sources need to be clarified. In some developing countries, routine data on injuries are not always captured in health information systems. It is therefore necessary to consider other sources of data—for example, law enforcement agencies, coroners, or medical examiners. Next, the events and variables in an injury surveillance system should be defined according to the objectives of the system. Criteria such as the intentionality (violencerelated injuries versus unintentional injuries); the outcome (fatal injuries versus nonfatal injuries); and the nature of violence-related injuries (physical, sexual, psychological, deprivation, or neglect) should be considered when establishing the system. Finally, case definitions and coding procedures should be defined before implementing the system. For example, the Nicaraguan Ministry of Health, in collaboration with CDC and the Pan American Health Organization, began developing and implementing an injury surveillance system in 2001 (Clavel-Arcas, Chacon, and Concha-Eastman 2004). The system, based on the medical facility emergency department (ED), collects data on injuries in keeping with the Injury Surveillance Guidelines established by WHO (Holder, Peden, and Krug 2001). Under the system, a reportable case is defined as a patient who died from or was treated for an injury in the ED. Cases include patients with unintentional and violence-related injuries. ED staff members identify cases and collect data in five hospitals in Nicaragua. Information used to complete the instrument is collected directly from the patients or their representatives. An ED admission clerk collects basic demographic data on the patient’s arrival. ED medical staff members (physicians and nurses) collect the remaining information Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1007 ©2006 The International Bank for Reconstruction and Development / The World Bank 147 (for example, location, mechanism of injury, nature, severity, and circumstances surrounding the injury) during triage and assessment. The hospital epidemiologist collects data collection forms daily from the ED, reviews the quality of data, and requests data from the ED staff if the forms are incomplete. The statistician reviews data daily. The country project coordinator also monitors the quality of the data periodically. Using Epi Info 2002 programs developed specifically for this project, the project coordinators analyze trends and identify potential risk factors (Noe and others 2004). The information is used to produce monthly reports for dissemination. Information is reported at both the regional and the country levels. Injury prevention programs in Nicaragua use surveillance data to assess the need for new policies or programs and to evaluate the effectiveness of existing policies and programs. For example, the municipality of León is using the information from the hospital to monitor the increase in suicide attempts among youths abusing pesticides and to evaluate an intersectoral campaign to promote life that includes primary through tertiary prevention strategies. Surveillance for Biologic Terrorism Surveillance for biologic terrorism is conducted primarily for outbreak detection and management. Surveillance must support early detection of an incident of biologic terrorism and its characterization in the same manner as for the detection and control of naturally occurring outbreaks of infectious diseases. Early detection of outbreaks can be achieved by the following (Buehler and others 2004): • timely and complete receipt, review, and investigation of disease case reports, including the prompt recognition and reporting to or consultation with health departments by physicians, health care facilities, and laboratories • improvement of the ability to recognize patterns indicative of a possible outbreak early in its course (for example, by using analytic tools that improve the predictive value of data at an early stage of an outbreak or by lowering the threshold for investigating possible outbreaks) • receipt of new types of data (such as purchases of health care products, absences from work or school, symptoms presented to a health care provider, or orders for laboratory tests) that can signify an outbreak earlier in its course. Environmental detection systems for microbial pathogens and toxins of concern for biologic terrorism might also be categorized as new types of data early in the course of an outbreak, before infection (Meehan and others 2004). The primary surveillance tools for event detection and management are the traditional disease-reporting systems for notifiable diseases discussed elsewhere in this chapter. These core surveillance tools should be robust before new data types can be considered for supplementing public health surveillance. Syndromic surveillance is an investigational approach by which health department staff members, assisted by automated data acquisition and generation of statistical signals (computerized algorithms), monitor disease indicators continually to detect outbreaks of disease earlier and more completely than might otherwise be possible with traditional reportable disease methods (Buehler and others 2004). CDC’s list of biologic terrorism agents and diseases can be found at http://www.bt.cdc.gov and an updated list of references dealing with syndromic surveillance is at http://www.cdc. gov/epo/dphsi/syndromic/. Complex Emergency Surveillance The key elements in planning a disaster surveillance system are establishing objectives, developing case definitions, determining data sources, developing simple data collection instruments, field testing the methods, developing and testing the analysis strategy, developing a dissemination plan for the report or results, and assessing the usefulness of the system. The surveillance needs are different in the preimpact, impact, and postimpact phases (Binder and Sanderson 1987). The role of surveillance in disaster situations has included the following broad framework of activities: • predisaster activities (for example, hazard mapping, provision of guidelines, and training for medical and rescue teams) • continuous monitoring and surveillance for priority health problems in affected populations (for example, in the posttsunami surveillance in Tamil Nadu, India, a one-page instrument was used for 10 priority health conditions for daily active surveillance in displaced populations at camps) • prospective surveillance of affected populations focusing on the natural history of exposure and health effects and longterm effects of stress disorders among survivors. Surveillance in Refugee Populations Support of relief efforts following national and global disasters has been a relatively new application of epidemiologic practice for the public health professionals. Nevertheless, since the initial CDC involvement with the United Nations in a large-scale relief effort concerning approximately 20 million displaced people affected by the 1967–70 civil war in Nigeria, CDC staff members have participated in several assessments of the health needs, damage, and nutrition in refugee populations resulting from man-made and natural disasters. The more notable and extended actions were conducted in the 1979–82 Khmer 1008 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 148 Thailand-Cambodia refugee-relief action, followed by longterm public health surveillance of Somalian refugees (1980–83), periodic but comprehensive health and nutritional assessments of Afghan refugees in Pakistan (1980–2002), and growth and nutritional assessments of internally displaced populations—especially children—in the Democratic People’s Republic of Korea (1990s) and southern Sudan. Although these relief efforts occurred many years and many thousands of miles apart, they shared several important characteristics: • Large numbers of people were in fixed camps or on the move searching for food and shelter. These needs were usually addressed by external aid agencies and many times caused local environmental degradation (fuel, temporary housing, water pollution, and so on). • Refugees, after the initial phase, competed with indigenous populations for scarce jobs, leading to social strife and stress. Refugees were also exploited and suffered violence— additional factors leading to stress and social maladjustment. • No administrative structure to provide and coordinate assistance of the necessary magnitude existed before the crisis, and thus, it had to be created after the fact. • Assistance was complicated by the uncertainty associated with military activity, crime, and hostile governments. • Data that were relatively simple to gather and analyze provided health workers and administrators information needed to plan and monitor assistance and its impact. • Close collaboration with other local and international relief organizations (such as the United Nations High Commissioner for Refugees, the International Red Cross, the United Nations Children’s Fund, WHO, and USAID) was essential to instituting and sustaining a meaningful surveillance system for refugees that led to interventions. The major goal of these activities is to identify and eliminate preventable causes of morbidity and mortality. Planning requires effective use of existing knowledge about characteristic or predictable demographic patterns, easily applied health indicators, and avoidable errors of omission or commission. As in disasters, the principles of surveillance (data collection, data analysis, response to data, and assessment of response) and other public health techniques should be an integral part of relief efforts. Retrospective evaluation of these efforts has also proved useful (CDC 1983). Chronic Disease Surveillance Systems Development and evaluation of policies for health improvement require a reliable assessment of the burden of disease and injury, an inventory of the disposition of resources for health, assessment of the policy environment, and information on the cost effectiveness of interventions and strategies. In all these areas, consideration of noncommunicable (mostly chronic) conditions becomes critical. In 1999, noncommunicable diseases were estimated to cause approximately 60 percent of the deaths in the world and 43 percent of the global burden of disease (WHO 2000a). WHO forecasts that by 2020 the burden of disease from noncommunicable diseases for developing and newly industrialized countries will have increased more than 60 percent (Murray and Lopez 1996). Some developing countries have found it difficult to acquire and analyze accurate mortality statistics regularly, let alone morbidity and quality-of-life information. Ensuring development, implementation, and widespread use of noncommunicable disease data for better decisions on resource allocation is critical to improving the quality of lives and promoting a more equitable future for health within and between countries. Hypertension, elevated blood cholesterol, tobacco use, excessive alcohol consumption, obesity, and the multiple diseases linked to these risk factors are a global public health problem. In one study, smoking, high blood pressure, and high cholesterol alone explained approximately two-thirds to threefourths of heart attacks and strokes (Vartiainen and others 1995). Until recently, surveillance for risk factors was an activity commonly associated with developed countries (Holtzman 2003). However, recently WHO has increased attention to noncommunicable disease surveillance by developing tools and working to achieve data comparability between countries (WHO 2003c). Data on key health behaviors, obesity, hypertension, lipids, and diabetes are collected inconsistently in developing countries, especially in Africa. Data on tobacco use are available through the Global Youth Tobacco Survey (http://www.cdc.gov/tobacco/global). Incidence data (the number and proportion of new cases in a population) are limited in developing countries. However, India’s National Cancer Registry program may serve as a notable exception (http://icmr.nic.in/ncrp/cancer_regoverview.htm). In 1981, the Indian Council of Medical Research, recognizing that there was a lack of information on follow-up of cancer patients to assess quality of care, instituted a cancer registry network. The network provides data on the magnitude and patterns of cancer in eight areas of India to enable studies of the histologic features correlating with prognosis and association studies (for example, whether a history of vasectomy is associated with cancer of the prostate). Another important example relates to the widespread use of folic acid in China and the resultant reduction in incidence of birth defects (Kelly and others 1996; Wald 2004). Surveillance data have been critical in establishing the importance of obesity as a public health priority in the United States. Data for individual states provided by CDC’s BRFSS have enabled individual health departments to document their obesity epidemic (Sturm 2003). These data provide a measure Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1009 ©2006 The International Bank for Reconstruction and Development / The World Bank 149 of the effectiveness of interventions to meet the control objectives. The BRFSS is a practical tool for developing and middle-income countries, as Jordan demonstrated when it implemented a BRFSS in 2002; the first survey documented substantial levels of obesity, especially among women, combined with low levels of physical activity (CDC 2003b). Table 53.3 Health Expenditures by National Income Level of Countries, 2001 Income groupa Government expenditures on health as a percentage of gross domestic product Total expenditures on health as a percentage of gross domestic product 6.30 10.74 High income ECONOMICS OF PUBLIC HEALTH SURVEILLANCE SYSTEMS The outbreak of SARS in 2003 demonstrates the far-reaching economic impact of not having an effective global public health surveillance system in place, with an estimated reduction in real gross domestic product of more than US$1.0 billion in Canada (Darby 2003) and estimated income losses in the range of US$12.3 billion to US$28.4 billion for East and Southeast Asia as a whole (Fan 2003). Public health surveillance is considered a global public good (Zacher 1999), particularly when it is used for eradication of such diseases as poliomyelitis. As eradication campaigns decrease the number of cases, maintaining systems to find the last few cases becomes more expensive. Often, the majority of the costs for these systems fall on hard-pressed developing countries. This factor raises questions of fairness and equity. For example, as poliomyelitis becomes rare, it ceases to be a significant risk to national populations, whereas other diseases, such as malaria and diarrhea, typically are major causes of morbidity and mortality. In such countries, it seems most fair and efficient for the global community to finance eradication campaigns, leaving national systems free to address the diseases that most affect their populations. The negative impact of globally mandated eradication surveillance systems can be mediated or reversed by leveraging on the eradication program’s infrastructure to gather surveillance data for diseases of concern to local governments (Nsubuga, McDonnell, and others 2002). A similar case can be made for influenza early warning systems in countries that gather information that will be used to create vaccines that will benefit other populations but not their own. Equity demands that the countries that benefit from such systems finance them. Public health surveillance systems serve an essential function in preventing and controlling disease spread within and across national borders. Although the private sector benefits, it lacks the incentive to invest in public health surveillance systems, and sovereign states depend on the contribution of others (WHO 2002); this situation has important implications for the financing of public health surveillance systems. Even within national borders, the difficulty of quantifying the benefits of surveillance systems for individual communities leads to neglect by local authorities, providing the economic rationale for funding by the national government. Developing countries Upper-middle income 3.68 6.41 Lower-middle income 2.58 5.63 Low income 1.22 4.78 Source: World Bank 2004. a. All World Bank member economies with populations of more than 30,000 are classified into income groups, divided according to 2003 gross national income per capita, calculated using the World Bank Atlas method. The groups are low income, US$765 or less; lower-middle income, US$766 to US$3,035; upper-middle income, US$3,036 to US$9,385; and high income, US$9,386 or more. are reportedly the weak link in the global surveillance framework, although they bear the greatest burden of disease, emerging and reemerging old pathogens, and drug-resistant pathogens (U.S. GAO 2001). The greatest need for surveillance systems is in these countries, but most lack both the resources and the political will to build human capacity and finance the systems (table 53.3). Resource constraints and intense pressure to provide care and treatment services lead public health authorities in the poorest countries to spend resources on surveillance (U.S. GAO 2001). Because the costs and benefits derive from surveillance systems spilling across national borders, donors should assist with capacity building in countries that have been unable to invest the human and material resources required. An interesting and unresolved feature of these global public goods—the solution to their adequate provision and supply rests at local, national, and sometimes regional levels—has prompted the international health community to advocate for capacity building in developing countries rather than for consolidation of the fragmented systems at the global level (WHO 2002). Standard tools of economic evaluation (Meltzer 2001) have been used to compare the benefits and costs of several public health interventions. The public good characteristics of surveillance systems, with benefits that are not easy to quantify, make the use of such tools difficult to implement in practice. However, economic evaluation of laboratory surveillance systems to detect specific disease-causing organisms have been undertaken in the developed world by comparing benefits and costs now and in the future (Elbasha, Fitzsimmons, and Meltzer 2000). These evaluations have not been done in developing countries and are needed. At best, an analysis of the benefits and costs of existing or proposed surveillance systems is feasible. This analysis requires an estimate of the cost of illness and answers the 1010 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 150 question of how many cases of a particular disease need to be prevented by the surveillance system to be exactly equal to the expenditure on the system. Given expenditures on specific health interventions or programs, one can, by using traditional econometric tools, apply the data on health outcomes from the surveillance systems as inputs to economic analysis. Surveillance also clearly leads to a cost saving if it prevents the need for expenditure on treating patients. FUTURE OF SURVEILLANCE Public health agencies, ministries of finance, and international donors and organizations need to transform surveillance from dusty archives of laboriously collected after-the-fact statistics to meaningful measures that provide accountability for local health status or that deliver real-time early warnings for devastating outbreaks. This future depends in part on developing consensus on critical surveillance content and developing commitment on the part of countries, funding partners, and multilateral organizations to invest in surveillance system infrastructure and to use surveillance data as the basis for decision making. This vision of the future assumes a coherent, integrated approach to surveillance systems that is based on matching the surveillance objective with the right data source and modality and on paying attention to country-specific circumstances while maintaining global attention to data content needs. Information technology and informatics can help in attaining this vision. Specifically, technology can facilitate the collection, analysis, and use of surveillance data, if data standards are developed and compatible systems are established. Data collection for surveillance would be an automatic by-product of any electronic systems used to support clinical care. Under this scenario, an automatic electronic message would be sent to the responsible public health jurisdiction with information about a health event (for example, death, disease, or injury), including all relevant information from the electronic health record about the patient, provider’s name, patient’s home address, risk factors, previous immunizations, and treatments. Even before this ideal capacity becomes widespread, technology such as cell phone–based systems could accelerate collection of key data (for example, occurrence of a viral hemorrhagic fever outbreak). The rapid penetration of cell phones in developing countries might obviate the need for prohibitively expensive landline-based systems. An accelerated system of wireless Internet access might also transform the capacities to which a local health post or a district health official might have access. These systems should also be considered as means for collecting information beyond traditional data. For example, telemedicine access can permit views of a rash illness to be shared with national or international medical specialists. Analysis of surveillance data can also be transformed by using available technology. Software that is Web-enabled, together with the advances in geographic information system software and global positioning devices, means that anyone with Internet access can potentially apply the latest version of software running on a distant server in the national capital to local data to generate up-to-date maps and graphs describing health status in that jurisdiction. Use of surveillance data can also be transformed. Sophisticated algorithms can be applied to data as it is collected to determine when (and how) an alert should be sent to local, national, or even international health officials to indicate a need for immediate investigation. Increasingly sophisticated visual display techniques and creation of custom channels with data of particular relevance to groups of data users are just some of the tools already being used to put public health content on the desktop of anyone with broadband, secure Internet access. Realization of this future vision does not require technology beyond what is already feasible, but the following factors are needed: • the organizational and political will to develop and coordinate the needed systems and standards that will enable those systems • appropriate attention to individual privacy and system security • removal of the financial and logistical barriers to broadband Internet access • a strategic multisectoral approach to accelerating national infrastructure among the poorest developing nations. GLOBAL SURVEILLANCE NETWORKS Globally, infectious disease surveillance is implemented through a loose network that links parts of national health care systems with the media, health organizations, laboratories, and institutions focusing on particular disease conditions. WHO has described a “network of networks” (U.S. GAO 2001) that links existing regional, national, and international networks of laboratories and medical centers into a surveillance network (figure 53.5). Government centers of excellence (for example, CDC, the French Pasteur Institutes, and FETPs) along with WHO country and regional offices also contribute to disease and health condition reporting. Military networks, such as the U.S. Department of Defense’s Global Emerging Infectious Disease System, and Internet discussion sites, such as ProMed (http://www.promedmail.org) and Epi-X (http://www.cdc.gov/ epix), also supplement the reporting networks. In 1997, WHO started the Global Outbreak Alert and Response Network, and Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1011 ©2006 The International Bank for Reconstruction and Development / The World Bank 151 UNHCR and UNICEF country offices WHO collaborating centers and laboratories Epidemiology training networks National public health authorities Military laboratory networks WHO regional and country offices Global Public Health Intelligence Network Internet discussion sites Nongovernmental organizations Media press Formal Informal Source: U.S. GAO 2001. a. UNHCR represents the United Nations High Commissioner for Refugees. b. UNICEF represents the United Nations Children’s Fund. Figure 53.5 Global Infectious Disease Surveillance Frameworks it was formally adopted by WHO member states in 2000. The network has more than 120 partners around the world and identifies and responds to more than 50 outbreaks in developing countries each year (Heymann and Rodier 2004). The International Health Regulations are the only binding international agreements on disease control. The regulations provide a framework for preventing the international spread of disease through effective national surveillance coupled with the international coordination of response to public health emergencies of global concern by using the guiding principle of maximum protection, minimum restriction (WHO 2003a). The current regulations apply only to cholera, plague, and yellow fever; they require WHO member states to notify WHO of any cases of these diseases that occur in humans within their territories and then give further notification when the territory is free of infection. The regulations are being revised to include the development of national core capacities and national focal persons who have the competencies of graduates of FETPs and allied training programs. Programs established to improve the capacity of both epidemiologists and laboratorians to collect, use, and interpret surveillance and outbreak data (for example, the collaborative WHO program in foodborne diseases called the WHO Global Salm-Surv) are also important components in developing global surveillance networks. RESEARCH AGENDA IN PUBLIC HEALTH SURVEILLANCE Developing nations share surveillance needs with the rest of the world, yet they are challenged by economic limitations, weak public health infrastructure, and the overwhelming challenges of poverty and disease. As a result, countries in the developing world often depend on the research efforts of others, or they collaborate with others to conduct the research necessary for their surveillance needs. Within individual countries, surveillance systems are essential in measuring disease and injury burden as a first step in establishing public health priorities that lead to policies and programs. 1012 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 152 The major research question for surveillance is how to develop and maintain a cadre of competent, motivated surveillance and response workers in developing countries. Other questions include how to design and maintain surveillance systems for these problems, especially morbidity systems for chronic diseases. Standard methods can be used to evaluate existing surveillance systems, which, in turn, will help define surveillance needs (Romaguera, German, and Klaucke 2000). Developing countries have used the IDSR strategy, which provides an efficient approach to data collection and analysis. Unfortunately, the majority of developing countries have limited surveillance systems for noninfectious diseases; instead, existing data systems (for example, vital records, motor vehicle crash records, or insurance claims data) are potential sources of surveillance data. In other settings, even these data sources are scarce, and approaches such as verbal autopsies and recurrent surveys might be alternatives (White and McDonnell 2000). Surveillance for risk factors is another challenge, and BRFSSs need to be validated and applied more widely in developing countries. Surveillance for injuries, environmental hazards (such as traffic intersections that are associated with high rates of injuries), and exposures to chemical or biological agents is a key public health concern with few examples of effective application anywhere in the developed or less developed parts of the world. Rigorous research is required in this field (Thacker and others 1996). The burgeoning use of electronic data systems and the almost universal availability of the Internet provide a tremendous opportunity for more timely and comprehensive surveillance in all parts of the world. Yet in this rapidly emerging field, critical needs exist, including the following: • • • • • competent, motivated health workers data standards (Lober, Trigg, and Karras 2004) global policies and practices for international surveillance useful software (Dean 2000) evaluation of the effectiveness of all these applications. New approaches that must be evaluated by using standard methods (Romaguera, German, and Klaucke 2000) include the following: • IDSR for infectious diseases • syndromic surveillance (CDC 2004b) for terrorism and emergency response • laboratory-based surveillance methods to enhance diagnostic accuracy and increase timeliness of recognition of outbreaks and interventions (Swaminathan and others 2001). • conduct surveillance for multiple competing risk factors that lead to a single condition (for example, smoking, cholesterol, hypertension, and overweight for heart disease) • conduct surveillance for the adverse effects of drugs • interpret ecologic data relative to data collected on individuals (Greenland 2004) • measure cost-effectiveness of alternative approaches to surveillance (for example, integrated compared with categorical approaches) • link data sources effectively (for example, hazard, exposure, and outcome data for environmental diseases) • build and sustain human infrastructure in developing countries • strengthen evidence-based decision-making cultures in ministries of health and finance. CONCLUSION Public health surveillance is an essential tool for ministries of finance, ministries of health, and donors to effectively and efficiently allocate resources and manage public health interventions. To be useful, public health surveillance must be approached as a scientific enterprise, applying rigorous methods to address critical concerns in this public health practice (Thacker, Berkelman, and Stroup 1989). Although the surveillance needs in the developing world appear to differ from those in the developed world, the basic problems are similar. In a time when we are confronted with SARS and avian influenza, the need to integrate global networks is undeniable, and research in how these concerns are addressed is essential. Collaboration among practitioners, researchers, nations, and international organizations is necessary to address the global needs of public health surveillance. ACKNOWLEDGMENTS The authors would like to acknowledge ORISE Fellow Danielle Backes of the Division of International Health, Coordinating Office for Global Health, U.S. Centers for Disease Control and Prevention, for her invaluable help in preparing this chapter. REFERENCES Binder, S., and L. M. Sanderson. 1987. “The Role of the Epidemiologist in Natural Disasters.” Annals of Emergency Medicine 16 (9): 1081–84. Many research questions remain about surveillance methodology, including how to do the following: Birkhead, G. S., and C. M. Maylahn. 2000. “State and Local Public Health Surveillance.” In Principles and Practices of Public Health Surveillance, ed. S. M. Teutsch and R. E. Churchill, 270. New York: Oxford University Press. • Buehler, J. W., R. S. Hopkins, J. M. Overhage, D. M. Sosin, and V. Tong. 2004. “Framework for Evaluating Public Health Surveillance Systems use data for forecasting or temporal and spatial analysis for aberration detection Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1013 ©2006 The International Bank for Reconstruction and Development / The World Bank 153 for Early Detection of Outbreaks: Recommendations from CDC Working Group.” MMWR Recommendations and Reports 53 (RR-5): 1–11. CDC (U.S. Centers for Disease Control and Prevention). 1983. “Surveillance of Health Status of Kampuchean Refugees—Khao I-Dang Holding Center, Thailand, December 1981–June 1983.” Morbidity and Mortality Weekly Report 32 (31): 412–15. ———. 2001. “Outbreak of Ebola Hemorrhagic Fever, Uganda, August 2000–January 2001.” Morbidity and Mortality Weekly Report 50 (5): 73–77. ———. 2003a. “Efficiency of Quarantine during an Epidemic of Severe Acute Respiratory Syndrome—Beijing, China, 2003.” Morbidity and Mortality Weekly Report 52 (43): 1037–40. ———. 2003b. “Prevalence of Selected Risk Factors for Chronic Disease— Jordan, 2002.” Morbidity and Mortality Weekly Report 52 (43): 1042–44. ———. 2004a. Partnerships in Excellence: Charting the Future in Global Health. Atlanta: CDC, Epidemiology Program Office, Division of International Health. ———. 2004b. “Syndromic Surveillance: Reports from a National Conference, 2003.” Morbidity and Mortality Weekly Report 53 (Suppl.): 18–22. Clavel-Arcas, C., R. Chacon, and A. Concha-Eastman. 2004. “Hospital Based Injury Surveillance Systems in Nicaragua and El Salvador, 2001–2002.” Paper presented at the Seventh World Conference on Injury Prevention and Safety Promotion, Vienna, June 6–9. Concha-Eastman, A., and A. Villveces. 2001. Guidelines for the Design, Implementation and Evaluation of Epidemiological Surveillance Systems on Violence and Injuries. Washington, DC: Pan American Health Organization. Darby, Paul. 2003. “The Economic Impact of SARS.” Publication 434-05, Conference Board of Canada, Ottawa. Dean, A. D. 2000. “Computerizing Public Health Surveillance Systems.” In Principles and Practices of Public Health Surveillance, ed. S. M. Teutsch and R. E. Churchill, 229–52. New York: Oxford University Press. Effler, P., M. Ching-Lee, A. Bogard, M. C. Ieong, T. Nekomoto, and D. Jernigan. 1999. “Statewide System of Electronic Notifiable Disease Reporting from Clinical Laboratories: Comparing Automated Reporting with Conventional Methods.” Journal of the American Medical Association 282: 1845–50. Elbasha, E. H., T. D. Fitzsimmons, and M. I. Meltzer. 2000. “Costs and Benefits of a Subtype-Specific Surveillance System for Identifying Escherichia coli O157:H7 Outbreaks.” Emerging Infectious Diseases 6 (3): 293–97. Fan, E. X. 2003. SARS: Economic Impacts and Implications. ERD Policy Brief Series 15. Manila: Asian Development Bank. Foege, W. H., R. C. Hogan, and L. H. Newton. 1976. “Surveillance Projects for Selected Diseases.” International Journal of Epidemiology 5 (1): 29–37. Greenland, S. 2004. “Ecologic Inference Problems in the Analysis of Surveillance Data.” In Monitoring the Health of Populations, ed. R. Brookmeyer and D. F. Stroup, 315–40. New York: Oxford University Press. Herikstad, H., Y. Motarjemi, and R. V. Tauxe. 2002. “Salmonella Surveillance: A Global Survey of Public Health Serotyping.” Epidemiology and Infection 129 (1): 1–8. Heymann, D. L. 2004. Control of Communicable Diseases Manual. Washington, DC: American Public Health Association. Heymann, D. L., and G. Rodier. 2004. “Global Surveillance, National Surveillance, and SARS.” Emerging Infectious Diseases 10 (2): 173–75. Holder, Y., M. Peden, and E. Krug. 2001. Injury Surveillance Guidelines. Geneva: World Health Organization. Holtzman, D. 2003. “Analysis and Interpretation of Data from the U.S. Behavioral Risk Factor Surveillance System (BRFSS).” In Global Behavioral Risk Factor Surveillance, ed. D. V. McQueen and P. Puska. New York: Kluwer Academic Press. Kelly, A. E., A. C. Haddix, K. S. Scanlon, C. G. Helmick, and J. Mulinare. 1996. “Cost Effectiveness of Strategies to Prevent Neural Tube Defects.” In Cost-Effectiveness in Health and Medicine, ed. M. R. Gold, J. E. Siegel, L. B. Russell, and M. C. Weinstein, 313–48. New York: Oxford University Press. Kew, O., V. Morris-Glasgow, M. Landaverde, C. Burns, J. Shaw, Z. Garib, and others. 2002. “Outbreak of Poliomyelitis in Hispaniola Associated with Circulating Type 1 Vaccine-Derived Poliovirus.” Science 296 (5566): 356–59. Lober, W. B., L. Trigg, and B. Karras. 2004. “Information System Architectures for Syndromic Surveillance.” Morbidity and Mortality Weekly Report 53 (Suppl.): 203–8. Mack, T. M. 2005. “The Ghost of Pandemics Past.” Lancet 365 (9468): 1370–72. McGeehin, M. A., J. R. Qualters, and A. S. Niskar. 2004. “National Environmental Public Health Tracking Program: Bridging the Information Gap.” Environmental Health Perspectives 112 (14): 1409–13. Meehan, P. J., N. E. Rosenstein, M. Gillen, R. F. Meyer, M. J. Kiefer, S. Deitchman, and others. 2004. “Responding to Detection of Aerosolized Bacillus Anthracis by Autonomous Detection Systems in the Workplace.” MMWR Recommendations and Reports 53 (RR-7): 1–12. Meltzer, M. I. 2001. “Introduction to Health Economics for Physicians.” Lancet 358 (9286): 993–98. Murray C. J. L., and A. D. Lopez, eds. 1996. Global Burden of Disease and Injury Series, Volume I: The Global Burden of Disease—A Comprehensive Assessment of Mortality and Disability from Diseases, Injuries, and Risk Factors in 1990 and Projected to 2020. Cambridge, MA: Harvard University Press. Narasimhan, V., H. Brown, A. Pablos-Mendez, O. Adams, G. Dussault, G. Elzinga, and others. 2004. “Responding to the Global Human Resources Crisis.” Lancet 363 (9419): 1469–72. Noe, R., J. Rocha, C. Clavel-Arcas, C. Aleman, M. E. Gonzales, and C. Mock. 2004. “Occupational Injuries Identified by an Emergency Department Based Injury Surveillance System in Nicaragua.” Injury Prevention 10 (4): 227–32. Nsubuga, P., N. Eseko, W. Tadesse, N. Ndayimirije, C. Stella, and S. McNabb. 2002. “Structure and Performance of Infectious Disease Surveillance and Response, United Republic of Tanzania, 1998.” Bulletin of the World Health Organization 80 (3): 196–203. Nsubuga, P., S. M. McDonnell, B. Perkins, R. Sutter, L. Quick, M. E. White, and others. 2002. “Polio Eradication Initiative in Africa: Influence on Other Infectious Disease Surveillance Development.” BMC Public Health 2 (1): 27. Pappaioanou, M., M. Malison, K. Wilkins, B. Otto, R. A. Goodman, R. E. Churchill, and others. 2003. “Strengthening Capacity in Developing Countries for Evidence-Based Public Health: The Data for DecisionMaking Project.” Social Science and Medicine 57 (10): 1925–37. Peden, M., K. McGee, and G. Sharma. 2002. The Injury Chartbook: A Graphical Overview of the Global Burden of Injuries. Geneva: World Health Organization. Peiris, J. S., K. Y. Yuen, A. D. Osterhaus, and K. Stohr. 2003. “The Severe Acute Respiratory Syndrome.” New England Journal of Medicine 349 (25): 2431–41. Romaguera, R., R. R. German, and D. N. Klaucke. 2000. “Evaluating Public Health Surveillance.” In Principles and Practices of Public Health Surveillance, ed. S. M. Teutsch and R. E. Churchill, 176–93. New York: Oxford University Press. 1014 | Disease Control Priorities in Developing Countries | Peter Nsubuga, Mark E. White, Stephen B. Thacker, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 154 Sturm, R. 2003. “Increases in Clinically Severe Obesity in the United States, 1986–2000.” Archives of Internal Medicine 163 (18): 2146–48. Swaminathan, B., T. J. Barrett, S. B. Hunter, and R. V. Tauxe. 2001. “PulseNet: The Molecular Subtyping Network for Foodborne Bacterial Disease Surveillance, United States.” Emerging Infectious Diseases 7 (3): 382–89. Thacker, S. B., and R. L. Berkelman. 1988. “Public Health Surveillance in the United States.” Epidemiologic Reviews 10: 164–90. Thacker,S.B.,R.L.Berkelman,and D.F.Stroup.1989.“The Science of Public Health Surveillance.” Journal of Public Health Policy 10 (2): 187–203. Thacker, S. B., and D. F. Stroup. 1998a. “Public Health Surveillance.” In Applied Epidemiology: Theory to Practice, ed. R. C. Brownson and D. B. Petitti, 105–35. New York: Oxford University Press. ———. 1998b. “Public Health Surveillance and Health Services Research.” In Epidemiology and Health Services, ed. H. K. Armenian and S. Shapiro, 61–82. New York: Oxford University Press. Thacker, S. B., D. F. Stroup, R. G. Parrish, and H. A. Anderson. 1996. “Surveillance in Environmental Public Health: Issues, Systems, and Sources.” American Journal of Public Health 86 (5): 633–38. Ungchusak, K., P. Auewarakul, S. F. Dowell, R. Kitphati, W. Auwanit, P. Puthavathana, and others. 2005. “Probable Person-to-Person Transmission of Avian Influenza A (H5N1).” New England Journal of Medicine 352 (4): 333–40. White, M. E., S. M. McDonnell, D. H. Werker, V. M. Cardenas, and S. B. Thacker. 2001. “Partnerships in International Applied Epidemiology Training and Service, 1975–2001.” American Journal of Epidemiology 154 (11): 993–99. WHO (World Health Organization). 1993. “Epidemiological Surveillance of Communicable Disease at the District Level.” WHO Regional Committee for Africa, 43rd session, AFR/RC43/18. WHO, Geneva. ———. 1998. “Integrated Disease Surveillance: Regional Strategy for Communicable Diseases.” WHO Regional Committee for Africa, 48th session. AFR/RC48/8. WHO, Geneva. ———. 2000a. Global Strategy for the Prevention and Control of Noncommunicable Diseases. Report by the Director-General, 53rd World Health Assembly. Provisional Agenda Items 2.11: A53/14. Geneva: WHO. ———. 2000b. “An Integrated Approach to Communicable Disease Surveillance.” Weekly Epidemiological Record 75 (1): 1–7. ———. 2002. Global Public Goods for Health: The Report of Working Group 2 of the Commission on Macroeconomics and Health. Geneva: WHO. ———. 2003a. Revision of the International Health Regulations World Health Assembly. Resolution WHA 56.28. Geneva: WHO. ———. 2003b. “Severe Acute Respiratory Syndrome (SARS).” Weekly Epidemiological Record 78 (12): 81–83. USAID (U.S. Agency for International Development). 2005. “Infectious Disease and Response Strategy 2005.” USAID, Washington, DC. ———. 2003c. The SURF Report 1. Surveillance of Risk Factors Related to Noncommunicable Diseases: Current State of Global Data. Geneva: WHO. U.S. GAO (U.S. General Accounting Office). 2001. Challenges in Improving Infectious Disease Surveillance Systems. GAO-01-722. Washington, DC: U.S. General Accounting Office. WHO Report of an International Commission. 1978. “Ebola Haemorrhagic Fever in Zaire, 1976.” Bulletin of the World Health Organization 56 (2): 271–93. Vartiainen, E., C. Sarti, J. Tuomilehto, and K. Kuulasmaa. 1995. “Do Changes in Cardiovascular Risk Factors Explain Changes in Mortality from Stroke in Finland?” British Medical Journal 310 (6984): 901–4. World Bank. 2001. World Development Report 2000–2001: Attacking Poverty. New York: Oxford University Press. Wald, N. J. 2004. “Folic Acid and the Prevention of Neural-Tube Defects.” New England Journal of Medicine 350 (2): 101–3. White, M. E., and S. M. McDonnell. 2000. “Public Health Surveillance in Low and Middle Income Countries.” In Principles and Practices of Public Health Surveillance, ed. S. M. Teutsch and R. E. Churchill, 287–315. New York: Oxford University Press. ———. 2004. World Development Indicators 2004. Washington, DC: World Bank. http://devdata.worldbank.org/hnpstats/query/default. html. Zacher, M. 1999. Global Epidemiological Surveillance: Global Public Goods. New York: Oxford University Press. Public Health Surveillance: A Tool for Targeting and Monitoring Interventions | 1015 ©2006 The International Bank for Reconstruction and Development / The World Bank 155 ©2006 The International Bank for Reconstruction and Development / The World Bank 156 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1017 Chapter 54 Information to Improve Decision Making for Health Sally K. Stansfield, Julia Walsh, Ndola Prata, and Timothy Evans The new source of power is not money in the hands of a few, but information in the hands of many. —John Naisbitt and Patricia Aburdene, Megatrends 2000 This chapter focuses on the collection and management of public health information, in contrast to clinical information, which concerns individual patient care encounters. Even when aggregated, clinical data are necessary, but not sufficient, to inform efforts to improve the health of populations. While substantial attention has been focused on these facility-based clinical consultations and the health management information system (HMIS) used to track the relevant data, we focus here on the broader health information system (HIS) needed to inform decisions at individual, facility, district, and national levels. Considered here are the routine data collection systems upon which program management, planning, monitoring, and evaluation depend. Information needs for specific tasks, such as for research or for program evaluation, are discussed in the chapters on research (chapters 4 and 7). Other chapters in this volume refer to information needs to enable disease control or to evaluate programs and improve the delivery of interventions. Those interested in these issues should also pay special attention to chapter 53 and chapters 70–73. This chapter bridges the global and the local issues; it makes the case for strengthening the evidence base for action through comprehensive health information systems that include census, vital events, monitoring, public health surveillance, resource tracking, facility-based service statistics, and household surveys. INTRODUCTION From infancy on, we receive information that gives form to our thinking and problem solving. The method by which a phenomenon is measured shapes societal perceptions of it and the collective efforts to affect it. Likewise, the choices we make in the collection and use of information for health will determine our effectiveness in detecting problems, defining priorities, identifying innovative solutions, and allocating resources for improved health outcomes. Despite those fundamental realities, there has been little awareness to date of the ramifications that greater information use can have for advancing health, and even less attention has been given to systems needed to provide timely, accurate, and relevant information. An example of the formative power of information for policy change lies in the history of the United Nations’ Standard System of National Accounts, created by Richard Stone more than 50 years ago. The annual reporting of these accounts by most countries shapes our impressions of the relative position of nations, defines our views of the differential opportunities offered to their citizens, and drives the content of national and global political discourse (Jolly 2002). Another example is the measurement of disability-adjusted life years (DALYs), which has shaped priorities for investment in global health over the past decade. 1017 ©2006 The International Bank for Reconstruction and Development / The World Bank 157 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1018 However, data or information alone will not transform outcomes. Data, which are simple measures of characteristics of people and things, have little inherent meaning or value. Analysis of the data enables the identification of patterns, thereby creating information. Finally, the use of information to generate recommendations, rules for action, and behavior change signifies the creation of knowledge that is used to make decisions and change human behavior. Good decisions on effective policies, services, and behaviors require timely, accurate, and relevant information. Health information is required for strategic planning and the setting of priorities; clinical diagnosis and management of illness or injury; quality assurance and quality improvement for health services; detection and control of emerging and endemic disease; human resource management; procurement and management of health commodities (including drugs, vaccines, and diagnostics); regulation of toxic exposures; program evaluation; research; and other types of policies and programs (Walsh and Simonet 1995). Citizens require such information to choose healthy behaviors, to demand effective policies and services, and to hold their governments accountable for the allocation and use of resources for health. Internationally, information is required to meet transnational needs, such as for the detection and control of consequences of epidemics and infectious diseases, results-based management of development assistance programs, and advocacy for increased financing for health. Several recent trends further enhance the pressures to deliver better health information. Global epidemics, such as of severe acute respiratory syndrome (SARS) and “bird flu,” have amply demonstrated the need and potential benefits of sensitive and transparent systems for tracking health events. Donors, including the Global Fund and the Global Alliance for Vaccines and Immunization (GAVI), increasingly demand performance measures and detailed evidence to justify new requests for support. “Basket” funding and sectorwide approaches place further responsibilities on countries to define their own priorities. Decentralization and devolution of budgetary controls have shifted much of this growing burden to the periphery, requiring districts to provide local evidence as a basis for decisions. Tracking progress toward the Millennium Development Goals for health requires empowering countries to measure key indicators and produce evidence-based strategic plans to achieve and document that progress. Furthermore, nearly every chapter in this volume cites the need for better information, including through research dependent on a health information system, to accelerate improvements in health.1 Yet there is a striking disconnect between the need for information and the ability to respond to that need. To collect, collate, analyze, and communicate the necessary information in a timely and understandable fashion requires organized processes and procedures and a comprehensive HIS. However, donor-driven and disease-specific initiatives have actually undermined efforts to develop a comprehensive HIS by creating separate, parallel, and often duplicative systems to meet the need for each funding source. Health information and the systems for its supply are a public good, meeting the defining criteria of being nonexcludable (in that, once the information is in the public domain, it is difficult to withhold from users) and nonrival (in that consumption of the information does not lessen its availability for use by others). As a public good, the supply of health information is the primary responsibility of governments: national governments for information within these jurisdictions, and international agencies and national governments together for international comparative information and global summary data. Harmonizing the data collection, standards, best practices, and other elements of a national and global HIS has several advantages. Standardization enables economies of scale for training, hardware and software, and processes. Routine health information is a summative good in that the collation of each contribution produces a cumulative increase in the value of the public good, strengthening the credibility and importance of that information. Furthermore, standardization of systems improves the reliability and comparability of information, both within nations and across national and regional boundaries (Cibulskis and Hiawalyer 2002). SYSTEMS AND SUPPLY OF HEALTH INFORMATION To create an effective HIS, governments must finance the system, create the necessary policy environment (for example, through legislation and regulation), and develop systems and services for the collection, collation, dissemination, and use of health information. A substantial portion of the national health information is fully within the control of government health officials. However, information from the private health sector and other parts of the government is also required. Table 54.1 lists some of the data required and their sources. A principal challenge is the integration of these intra- and extrasectoral functions into a single, comprehensive HIS. Direct Expenditures for Health Information As for most public goods, the production of health information is mostly financed by government appropriation. Budget support for the HIS comes through both the ministry of health and a national statistics office (NSO) in most countries. The NSO is usually responsible for collecting information through the national census and most household surveys. For the least developed countries especially, bilateral and multilateral donors are essential sources of finance, particularly for HIS planning, infrastructure development, and training. In Africa, it has been estimated that grants or loans from donors account for between 20 and 70 percent of the financing for statistical 1018 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 158 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1019 Table 54.1 Health Information from Sources Outside the Health Sector Health information Responsible agency Census and national surveys: Income and poverty distribution Household expenditure for health Coverage with health interventions National statistical office National expenditures for health, economic development indicators, and industrial production and distribution data Ministry of finance Employment data: Human resources for health Occupational health information Ministry of labor Import data: Pharmaceuticals and vaccines Capital equipment and health commodities Ministry of trade Food production and security information and nutritional status data Ministry of agriculture Military health service statistics Ministry of defense Patterns of transportation injury (including motor vehicle accidents) Ministry of transportation Literacy rates and school health program information Ministry of education in the management of health information. There should be protection from political interference and full empowerment of the health statistics office to make public statements in response to criticisms of reports and the underlying methods. Ethical practices for protecting privacy and confidentiality must be well understood, and procedures should be in place to deal with breaches in these standards. Accuracy and reliability should be stated as expectations and ensured through periodic review of data collection methods and through benchmarking with internationally credible definitions of indicators. A client orientation should be instilled and users of data regularly consulted in defining outputs and formats for the presentation of data. Systems for Collection, Management, and Analysis Source: Authors. systems overall. Revenue generated by selling statistical products and services accounts for 10 to 20 percent of the financing for national statistical systems (Economic Commission for Africa 2003). User fees or taxes for use of information products and services can partially offset the costs of developing and maintaining the information system. In many countries, taxes and tariffs on computer equipment and government regulation of communications and Internet use remain barriers to public access to health information. Cost must not be a barrier to use of health information for the public good. Information Policy Sound information systems require a legislative and regulatory environment that encourages and supports effective HIS development. At the global level, many efforts have been made to establish international standards and policy frameworks for statistical data (United Nations Statistical Commission 1994). These policy frameworks are used to establish mandates for collection of basic health data (such as a decennial census or surveillance for reportable infectious diseases), to ensure the independence of official statistical agencies, to reinforce professional ethics, and to create norms for data quality and dissemination. Another key policy intervention, less tangible though equally critical, is the creation of a culture of quality and transparency Most developing countries have no comprehensive strategy for information management, reflecting the fractal nature of donor and national investments in these systems. Interventions to improve the HIS in the least developed countries, often donor driven, have often focused only on a specific subsystem, primarily for health service statistics, and have neglected other components of the HIS. An effective HIS requires an overarching architecture that defines the data elements, processes, and procedures for collection, collation, presentation, and use of information for decision making throughout the health sector (see box 54.1). This information architecture promotes comparison and integration of data elements from a variety of subsystems. As O’Carroll (2003) points out, such a comprehensive design enables phased system development, reduces redundancy, increases efficiency, and improves interoperability. Interoperability is critical to ensuring, for example, that census data, vital statistics, and health facility data can be integrated to generate rates, ratios, cost-effectiveness estimates, and other information required to compare options for health investment. The Pan American Health Organization (PAHO) has led the Regional Core Health Data Initiative “to facilitate speedy access to basic information on the health situation in the countries of the Region.” This initiative has involved an international consultation and agreement on the priority data, collection methods, and indicators. The initiative has shown that it is possible to create a regional database of essential, consistent, valid, standardized, timely, and regular information. PAHO has used the information to set its priorities, whereas countries have applied the results to design health programs and to allocate resources to upgrade their information systems. In the future, the plan is to expand the systems to subnational districts (PAHO 2004). Other WHO regions, including the Asia Pacific, are instituting similar initiatives with Web-based publication of core health indicators. Data Collection. No single mechanism for data collection is adequate to meet the needs for public health decision making. Information to Improve Decision Making for Health | 1019 ©2006 The International Bank for Reconstruction and Development / The World Bank 159 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1020 Box 54.1 The Health Metrics Network: Harmonizing Investment in HIS Development Developing countries, multilateral agencies, bilateral donors, and technical resource agencies have recently come together to form a global Health Metrics Network (HMN) that is designed to provide guidance for the development of the HIS, both in meeting national information needs and in producing the required indicators for tracking progress toward global goals. The HMN will provide the first consensus technical framework for HIS architecture and a plan for development of national health information systems. This HMN Framework includes a blueprint for iterative improvements in the HIS; descriptions of core data collection subsystems (census, surveys, vital events monitoring, service statistics, and resource tracking); and procedures for management and dissemination of information. Source: Authors. These needs can be met using a combination of the six key health information subsystems: census, household surveys, public health surveillance, vital events monitoring, health service statistics, and resource tracking. Surveys are conducted on a sample in order to limit costs, whereas the other subsystems are more often designed to cover the entire population. In most developing countries, public health surveillance—except for certain disease-specific efforts—is conducted through passive reporting from health facilities. Especially where utilization rates are low, this facility-based surveillance may be considered a sample or “sentinel”surveillance strategy.Vital events monitoring is, ideally, universal; however, many countries use a phased introduction of vital events monitoring that makes it functionally a sentinel or sample-based data collection effort during the transition to universal coverage. A national census every 10 years is an irreplaceable component of a national information system because it provides denominator data for so many indicators and sampling frames for subsequent sample surveys. The major costs of a census come from activities to establish the census maps, enumerate populations, enter data, and analyze the results. The cartographic costs can often be shared with other government departments, because the resulting updated maps can be instrumental in carrying out other critical public functions. Sample surveys of households are a mainstay of health information collection in the developing world. They provide data on service utilization; coverage of health interventions (for example, immunization); morbidity (self-reported illness or disability); pregnancy outcomes; mortality levels, differentials, and trends; and causes of death (through associated verbal autopsy; that is, expanded interviews in the case of death to determine cause on the basis of signs and symptoms before death). Surveys are, almost without exception, funded externally in the least developed countries and are not seen within the country as being part of a health information “system.” They are, in fact, generally undertaken to compensate for the lack of information available through routine systems (AbonZahr and Boerma 2005). The investment in surveys has thereby enabled donors and developing countries to sustain their neglect of the development of comprehensive and sustainable national health information systems. The United Nations Population and Statistics Divisions and the European Statistical Office (EUROSTAT) also support household survey work. Differences in methodologies among these surveys are currently a barrier to the comparison of results. The World Bank’s Managing for Development Results Roundtable, held in Marrakech, Morocco, in 2004, recommended harmonization of these surveys to eliminate duplication. Nonetheless, surveys offer an important source of information that transcends most of the selection bias that is inherent in service statistics. Especially in the least developed countries, where vital events registration systems and census taking are embryonic or nonexistent, surveys represent the only source of unbiased information about demography, socioeconomic status, coverage, morbidity, mortality, health expenditures, and other characteristics of the population. Where substantial proportions of the population use private health services, household surveys are particularly important. Even industrial countries rely on periodic community-based sample surveys for immunization coverage, for health service utilization rates, and for information on household health expenditures (Perrin, Kalsbeek, and Scanlan 2004). The World Health Organization (WHO) recommends using periodic surveys to monitor coverage, such as for immunization programs, especially in view of the shortcomings of service statistics for obtaining these measures (Murray and others 2003). Some household surveys collect biological and clinical specimens, such as blood, saliva, urine, and self-collected vaginal swabs, or they check swabs for anemia, HIV, disease antibodies, vitamin A, and other conditions. However, the performance characteristics of most diagnostic technologies (for example, 1020 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 160 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1021 cost, ease of field use, sensitivity, and specificity) are designed for clinical use and do not lend themselves readily to use in population surveys, especially in remote areas of developing countries. Moreover, the collection of diagnostic information along with individual identifiers introduces complex ethical issues in the notification of people with treatable conditions, the financing of any required treatments, and the use of the specimens for other studies (Ties Boerma, Holt, and Black 2001). Public health surveillance has been defined as the “ongoing systematic collection, analysis, and interpretation of data on specific health events affecting a population, closely integrated with the timely dissemination of these data to those responsible for prevention and control” (Thacker and others 1996). In developing countries, surveillance usually focuses primarily on a set of notifiable diseases, mainly infectious, which health care providers and laboratories are often required by law to report. Some nations also track risk factors for important diseases, injury events, adverse drug reactions, cancers, and pregnancy outcomes. Surveillance may be intensified over a period of years to enable targeting of special interventions for the control or elimination of diseases such as polio, tetanus, or measles. Active surveillance or screening of populations for target diseases may also be used in specific circumstances, such as during peak seasons for the disease or during natural disasters, when the potential for epidemics may be high. Most passive surveillance data, however, are incomplete. Reliance on surveillance for reportable diseases diagnosed in health facilities omits diseases diagnosed among those who go to private providers or who are too poor to go to any health facilities. Even in health facilities, reportable diseases are often underrecognized or cannot be confirmed in laboratories that have inadequate resources. Sentinel surveillance methods and registries maintained in a few selected sites may be more representative of the entire population and more cost-effective in identifying and reporting the target diseases or health conditions; however, an outbreak may go undetected in a geographic area without a sentinel site. Special regional surveillance may also be used where populations are vulnerable to special health risks. The Vigisus project in Brazil, for example, has developed a system of epidemiologic and environmental surveillance for the prevention and control of disease among indigenous populations in the Amazon region (http://www.br.undp.org/ propoor/BRA97028a.htm). Despite the methodological hazards, public health surveillance is essential for both national and global planning and preparedness, especially in view of the risks of regional expansion (for example, of meningitis and polio) or global spread of recent epidemics (for example, of SARS and bird flu). Vital events monitoring is the continuous, compulsory, and (in most cases) universal civil registration of key vital events, such as births, deaths (sometimes including fetal deaths), marriages, divorces, and migrations. In many countries, vital events monitoring systems function poorly and may be found only as remnants of past colonial administrations. In 2003, 115 of 192 WHO member states reported mortality data with causes of death, capturing about one-third of global deaths, or 18.6 million deaths per year. In South Asia, only 60 percent of births are registered (22.5 million), and in Africa, only 30 percent (17 million). Alternatives to universal registration include the sample registration systems used in China and India and the demographic surveillance sites in Tanzania. The International Network of Field Sites with Continuous Demographic Evaluation of Populations and Their Health in Developing Countries (INDEPTH), an association of longitudinal vital and health statistics surveillance sites in 17 countries, can provide technical support and training for development and management of these demographic surveillance sites (http://www. indepth-network.org/). The UN Statistics Division has developed principles and recommendations for vital statistics systems to guide countries in their development (http:// unstats.un.org/unsd/demographic/sources/civilreg/civilreg methods.htm.) Vital events monitoring systems may also be enhanced to determine causes of death, whether those deaths occur within health facilities or in the community. When deaths occur outside the health care system, a verbal autopsy, or structured interview of the relatives of the deceased, can assist in determining the cause of death. Verbal autopsies can, however, be used reliably to diagnose only those few conditions that have characteristic clinical signs or patterns of signs that can be recognized by family members or by the health workers who review the interview data. WHO is now developing standardized tools for verbal autopsy that will enhance the sensitivity and specificity of these instruments and permit comparisons over time and across geographics. Health service statistics are critical management tools for both preventive and curative services. The statistics are collected at each level: community outreach service points, primary care facilities, and district and regional referral hospitals. Information from clients and providers documents the quantity and quality of services and enables managers to detect and solve problems in order to improve health outcomes and efficiencies. This health information subsystem must be “flexible and capable of adapting to local needs, while at the same time allowing for standardization of health care quality assurance indicators, and subsequent ability to measure and compare the quality performance of health facilities nationwide” (DuranArenas and others 1998). A principal barrier to improving service quality in many health care facilities is the lack of reliable systems for managing and retrieving individual patient records. Service statistics are especially powerful when they can be compared with population-based measures from censuses and surveys to estimate rates and ratios, such as disease incidence or Information to Improve Decision Making for Health | 1021 ©2006 The International Bank for Reconstruction and Development / The World Bank 161 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1022 service coverage rates. Most service statistics subsystems track data only from public sector providers and facilities. Future improvements must implement systems and incentives to ensure reporting of service data from the private sector. The resource-tracking subsystem must enable measurement and management of human resources; facilities; commodities (pharmaceuticals, vaccines, and other consumables); and finances. Human resource tracking provides a mechanism for licensing health service providers and accrediting health facilities. Licensure and accreditation can be paired with incentives to ensure service quality and private sector contributions to achieving public health goals. The national health account (NHA) framework provides methods for measuring total national expenditures for health from household, public, private, and donor sources. NHA data document the sources of health financing, the amount spent for services, the distribution of funds across services and interventions, and the distribution of health benefits from those services and interventions. An NHA framework tracks the flow of funds, for example, from the ministry of health to health providers and government service programs or from households to pharmacies and private providers. These internationally comparable data enable benchmarking of performance among countries (Peters and others 2000). Of the 68 countries that have implemented NHAs, only one-third have used the framework more than once. However, 19 of 21 countries studied can report at least one instance in which the NHA system has informed and shaped policies (De and others 2003). For example, South Africa’s NHA analysis documented a higher per capita health expenditure in the richest districts, leading to intensified efforts to mitigate these inequities (Abt Associates 2003). Information and Communications Technologies. The rapid evolution of information and communications technologies (ICT) over the past 30 years has immense implications for the potential speed, cost, and effectiveness of an HIS. But a “digital divide” persists, with poor countries failing to benefit fully from these ICT advances. Lack of access to reliable power sources, absence of Internet connectivity, inability to procure computer equipment and appropriate software, and inadequate technical support are some of the barriers. African users account for only 1 percent of the world’s Internet traffic, 80 percent of which is in South Africa (http://www3.sn.apc.org/africa). Although less than 0.001 percent of the Internet use in Africa is among health professionals, this usage is growing rapidly. Internet access in health facilities can make the HIS more effective and efficient by enabling instantaneous transmittal of data to central locations. Internet access in facilities can also speed data transmission and improve clinical outcomes by providing access to evidence-based decision support for clinical care (Godlee and others 2004; McLellan 2001). Even in remote areas where no telephone or cable access exists, satellite technology can provide access to e-mail. Several countries, such as Bolivia and Peru, have successfully used satellite telephone technologies to enable continuous Web-based updating of health databases. Because the effectiveness of epidemic control often depends on timely detection and reporting of outbreaks, e-mail and telephone technologies have shown particular promise for use in disease surveillance. In Peru, for example, 100 percent reporting was achieved and sustained within six months of rollout of a pilot surveillance system using cellular telephones (Lescano and others 2003). The system is to be expanded to national coverage this year. Although individual citizens will not soon have equal access to ICT, these technologies can immediately be better used to improve public health. Automation of data entry and analysis can ease data capture, validation, analysis, and transmittal of health information. District managers can generate reports with tables and charts and transmit them to central levels, which can then apply this knowledge to improve local management. Special prompts and “exception reports” can alert managers to unexpected findings that require double-checking or immediate interventions (for example, outbreaks of infectious disease, low immunization coverage, or other management problems). Use of free software, such as the U.S. Centers for Disease Control and Prevention’s Epi Info, can lower costs, but often these software packages require substantial adaptation to local needs, along with additional training and technical support. Acquisition of computer equipment should be viewed not as a one-time capital expenditure but as a long-term commitment to buy periodic upgrades, maintenance, and technical support. Experience shows that purchase of inexpensive software and computers, such as in the Eastern Cape Province of South Africa, may actually increase overall costs when they require early replacement with more adequate alternatives. Geographic information system (GIS) technologies have also been successfully used in districts in several countries to enable mapping and visual representation of the geographic distribution of risk factors, disease, and services. A desktop GIS viewer and mapping software are available in several shareware versions, including the WHO’s “Health Mapper,” so that maps can be produced at little cost. Other potentially promising technologies include electronic scanners and personal digital assistants for data capture (http://www.healthnet.org) and global positioning systems to facilitate the mapping process. The principal barriers to improved information systems, however, are human, not technological. Substantial investment in training and technical support must accompany the introduction of any new technology. If the HIS is not functionally solid, introducing ICT will likely only worsen existing problems. 1022 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 162 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1023 Dissemination and Use of Health Information Information is a means to the end of improving health, but the availability of reliable information does not guarantee its use or improved decision making. Because decisions are often driven as much by politics as by evidence, it is critical to design information systems to meet the needs of decision makers and to create a culture of evidence that provides incentives and accountability for evidence-based decision making. Extensive dissemination promotes widespread use and accountability. The many users of information include the following: • • • • • • • • • • health ministries at national, regional, and district levels researchers and evaluators legislative and policy analysts nongovernmental organizations and consumer organizations advocacy groups private sector health providers and insurers communities, including groups of patients journalists donors and international agencies concerned with health individuals and families. The literature on health information systems is replete with complaints of the neglect of existing information, yet remarkably little is known regarding the effectiveness of interventions to improve the use of information. The NHA experience (De and others 2003) suggests that policy makers are most likely to use information when it contributes to and informs a preferred government direction, especially if that information is not available to stakeholders outside the health ministry. But systems and dissemination patterns for information can be engineered to ensure that clients, providers, and managers will seek and use information to inform decisions. Standard procedures can be developed to ensure analysis and use of data at the level at which it is collected. Training of health workers can be designed to include both basic and refresher training in the analysis and interpretation of data that are relevant to each job. Expectations of information use can be built into routine job requirements, including use of evidence for planning, data requirements for periodic reporting to supervisors, and use of information during performance reviews. Groups of managers can be convened across districts or regions for benchmarking, in which each manager presents and compares performance data and is rewarded for transparency and learning. These practices will result only from intense training in analysis and use. For example, Loevinsohn (1994) demonstrated that fewer than half of midlevel managers were able to use the information system even to identify best- and worst-performing districts. Nonetheless, if managers use the information, and if improved efficiency and coverage with interventions is the result, the HIS becomes exceedingly cost-effective. Information will “allow the public, their elected representatives, or donors to determine whether they are obtaining value for money” (Cibulskis and Hiawalyer 2002; see also Mackay 1998). Providing full access to the media will help to accelerate expectations of evidence-based decision making and accountability. Civil society, including nongovernmental organizations, should be the principal users of information to create and sustain citizen demand for quality services. The Healthy Communities Foundation’s “dashboard” of lead indicators of health system performance exemplifies one promising example of the visual display of data (http://whatcom.healthycities.org/demo/ aboutus.htm). Such dissemination and use of health information has enhanced government accountability for improved health in Papua New Guinea, where reports of local government performance in improving health systems transformed election results (G. Hiawalyer, personal communication). BENEFITS, COSTS, AND COST-EFFECTIVENESS OF IMPROVED INFORMATION There is broad agreement that information—plus the knowledge it enables—creates value. Yet it is challenging, indeed, to quantify the added value of information. Information, after all, is necessary but never sufficient to achieve improved outcomes. Other resources—human, material, and financial—are required for change. Nonetheless, it is possible to define the interventions necessary to improve health information and to draw on a few studies to estimate the cost and cost-effectiveness of these investments. Strengthening of Systems The steps involved in strengthening HIS include securing funding for a review of the current HIS and planning reforms and then using that plan to secure funding for implementing the reforms. The reforms depend on legislation and regulations that delineate the requirements, incentives, and disincentives for collecting the needed information. Finally, the review of the current HIS includes a situational analysis and outline of a plan that involves a comprehensive information architecture that is linked to both national and international needs. The HMN Framework includes assessment and planning tools and HIS standards that will guide strengthening of systems. Full implementation will likely take at least 36 months, and the effects on decision making and health outcomes will be detectable only after approximately five years. Benefits and Effectiveness of Improved Information The value of health information can be characterized in terms of cost savings; system efficiencies (for example, increased coverage or quality of services); or improved health outcomes Information to Improve Decision Making for Health | 1023 ©2006 The International Bank for Reconstruction and Development / The World Bank 163 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1024 (for example, DALYs saved or improved health equity). Information can also be used to increase overall resources for health. Publications such as this volume, World Development Report 1993: Investing in Health (World Bank 1993), and the report of the Commission on Macroeconomics and Health (2000), are important examples of evidence that has been used to change health policies and increase resources for health. The industrial world holds examples of the use of information to make service provision more effective and efficient. A quality improvement and evidence-based decision assistance program for diabetes patients in the United States created a net savings of US$510,133,2 primarily by averting hospitalizations (Petrakos 1998). The U.S. Institute of Medicine estimates that a computerized system for managing physician orders for medications costing US$1 million to US$2 million could “pay for itself in three to five years” and prevent injury to hundreds of patients per year (Kohn, Corrigan, and Donaldson 2001). There are also promising examples of the benefits and effectiveness of improved information from developing countries. Quality improvements driven by better information in Bolivia resulted in a 300 percent increase in hospital utilization rates (Pappaioanou and others 2003). In rural Mali, populations enrolled in a community-based information system calculated delivery costs for childhood immunization to be US$1.47 per child, compared with US$2.79 per child among populations not registered (Zayan, Berggren, and Doumbia 1992). Better information can also improve efficiencies in the management of pharmaceutical resources. For example, implementing a subnational information system in the Eastern Cape province of South Africa led to improved access to pharmaceuticals, with a 39 percent reduction in stockouts of essential drugs. Such improvements undoubtedly lead to better health outcomes, which may result in increased productivity and consequently an increase in the growth rate of the gross domestic product (Jamison, Sachs, and Wang 2001; Nordhaus 2002). Costs of Improved Health Information Few studies have documented the costs of an HIS. Kleinau (2000) estimated the resource requirements for health service statistics, the most expensive of the six subsystems. Using similar assumptions, we have calculated updated costs. This estimate includes only the public sector facilities, not private sector reporting systems. Reporting from private providers would likely include a more limited set of reported data: diseases, vaccinations, possibly staffing, and minimal utilization. Table 54.2 summarizes the total annual costs and per capita costs of the six health information subsystems. The costs of a facility-based services statistics subsystem of the HIS (table 54.3) can be assumed in most developing countries to include routine public health surveillance, because these data are obtained at health facilities when ill patients are Table 54.2 Cost of Essential HIS Subsystems Total cost (US$ million) Per capita cost (US$) Low income High income Low income High income Health service statistics 4.8 25.9 0.16 1.66 Public health surveillance (included with health service statistics) 0 0 0 HIS subsystem 0 Census 7.5 30.0 0.25 1.0 Household surveys 0.6 1.0 0.02 0.03 Vital events surveillance 1.5 6.0 0.05 0.20 Resource tracking 1.5 3.0 0.05 0.10 15.9 65.9 0.53 2.99 Total Source: Authors. Note: Table is based on a population of 30 million. Household survey costs are based on the experience of the demographic and health surveys during 2001–2003 (Macro International, personal communication). Costs vary by sample size and by length of the survey instrument; Macro International estimates, an average cost of US$100 per survey participant. A sample of 6,000 is assumed for the low-income setting, and a sample size of 10,000 is assumed for the high-income setting. Cost estimates for vital events monitoring are based on demographic surveillance sites. In the high-income setting, the annual costs are assumed to quadruple. Resource-tracking costs are based on the experience of national health accounts (Abt Associates, personal communication), and the Egyptian Budget Tracking system. Similar costs are estimated for human resources and commodities. brought for treatment. The additional costs of program-specific surveillance (for example, in support of polio eradication or tetanus elimination programs) could be assumed with a minor marginal investment in addition to facility-based and community-based information systems, including for vital events surveillance. The calculated range for per capita annual costs of a comprehensive HIS—US$0.53 to US$2.99—compares closely to the estimates from a country setting in which those data have been obtained, including a low-resource country (Tanzania) with a per capita cost of approximately US$0.50 (Rommelmann and others 2004) and a high-resource country (Mexico) with a per capita cost of approximately US$1.00. The Health Metrics Network (HMN) Technical Task Force South Africa has also estimated costs of the HIS at approximately US$26 million (165 million rand) for a population of 43 million, yielding a per capita cost of US$0.60. The highest range of the estimate would apply in countries with higher salaries and a more comprehensive HIS. Estimations of the Cost-Effectiveness of Interventions to Improve Health Information The Tanzania Essential Health Interventions Program (TEHIP) is perhaps the best source of evidence for the cost-effectiveness of improved health information. The project was designed to 1024 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 164 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1025 Table 54.3 Annual Costs of the Facility-Based Services Statistics Subsystem of an HIS HIS cost Low-resource setting High-resource setting Primary care facility One person (salary US$4,514/year) spends 10 percent of time at each of 6,000 facilities (US$2,708,400) Two people (salary US$10,351/year each) spend 20 percent of time at each of 6,000 facilities (US$24,842,400) First referral level One person (salary US$4,514/year) spends 25 percent of time at each of 1,000 facilities (US$1,128,500) Two people (salary US$10,351/year each) spend 75 percent of time at each of 1,000 facilities (US$15,526,500) District hospital Two people (salary US$4,514/year each) spend 20 percent of time at each of 300 facilities (US$541,680) Two people (salary US$10,351/year each) spend 100 percent of time at each of 300 facilities (US$6,210,600) Regional level Three people (salary US$10,962/year each) spend 50 percent of time at each of 15 facilities (US$246,645) Three people (salary US$25,134/year each) spend 100 percent of time at each of 15 facilities (US$1,131,030) National level Six people (salary US$10,962/year each) spend 50 percent of time (US$32,886) Ten people (salary US$25,134/year each) spend 100 percent of time (US$251,340) Subtotal (personnel) US$4,658,111 US$47,961,870 Primary care facility US$100/year US$400/year First referral level US$250/year US$1,000/year District hospital US$500/year US$2,000/year Regional US$1,500/year US$5,000/year National US$5,000/year US$30,000/year Subtotal (supplies) US$7,350 US$38,400 Personnel Data collection instruments and supplies Information technology: computers and software Primary care facility 0 0 First referral level 0 0 District hospital 0 20 percent use of each of two computers with software at US$1,100 at each of 300 facilities (US$132,000) Regional level 0 Two dedicated computers with software at US$1,100 at each of 15 facilities (US$33,000) National level 50 percent use of each of four computers with software at US$1,100 (US$2,200) 10 dedicated computers with software at US$1,100 (US$11,100) Subtotal (information technology) US$2,200 US$176,100 Training cost US$180,000 US$1,730,000 Total cost US$4,847,661 US$49,906,370 Per capita cost US$0.16 US$1.66 Source: Authors. Note: Based on a model country with a total population of 30 million. test how evidence can be used to decentralize health sector planning at the district level and to what extent evidence-based priority setting would result in improved health outcomes. The project budgeted for a marginal investment of US$2.00 per capita for the information and for health interventions, although only US$0.80 per capita was actually spent. The slightly increased investment covered training in the use of the information to set priorities and to better manage the most cost-effective interventions. The information systems included a district burden-of-disease intervention priority profile, district health accounts, a district cost information system, and district health service mapping. Management and technical support strengthened the district and regional health sector use of the information for management and administration. Communities participated in the ownership and management of health facilities. The cost-effectiveness estimates in this Information to Improve Decision Making for Health | 1025 ©2006 The International Bank for Reconstruction and Development / The World Bank 165 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1026 Box 54.2 The Tanzania Essential Health Interventions Program TEHIP is a partnership between Tanzania’s Ministry of Health and the International Development Research Centre. The project was established to determine the feasibility of an evidence-based approach to health planning at the district level. Testing the premise of the World Bank’s (1993) World Development Report 1993: Investing in Health, TEHIP enabled district health planners in two of Tanzania’s 117 districts to collect and use burden-ofdisease and cost-effectiveness data to get the best value for money from national investments in health. Interventions included door-to-door collection of data and training or technical support for managers in the analysis and use of the data for decision making. TEHIP districts allocated services to high-burden diseases, resulting in a tripling of clinic utilization rates and increased treatment effectiveness. With a per capita increase in spending of only US$0.80, district health managers achieved a 47 percent reduction in child mortality rates. Source: Authors. Table 54.4 The Effectiveness of Evidence-Based Resource Allocation in Improving Health Year Number of children ⬍ 5 years Probability of dying (birth to 5 years) Mortality rate (⬍ 5 years) Total deaths 1999 31,000 135.5 34 2000 31,500 119.0 25 2001 32,000 110.0 2002 32,661 114.0 Deaths averted DALYs gained/ death 1,054 — — — — 791 263 41 10,850 11,511 25 803 251 41 10,332 10,643 26 853 202 41 8,303 8,304 29,487 30,458 Total DALYs gained Total DALYs DALYs discounted at 3 percent Source: Authors. Note: 1999 is baseline year; therefore, no deaths were averted. section are based solely on the declines in mortality of children under five years of age, even though adult mortality also decreased. To ensure a conservative estimate of the costs of the HIS, we used a per capita cost of US$2.00—higher than the actual investment for TEHIP and at the high end of the range of costs for a comprehensive HIS estimated in table 54.2— US$0.53 to US$2.99. All costs were ascribed to the information system, because there were no improvements in the interventions themselves. Expenditures and deaths before 2002 were discounted by 3 percent annually (see box 54.2). The demographic and epidemiologic data were taken from the Rufiji district, where the most complete data were available. The estimate of the number of children under age five (32,661) is based on the 2002 census results. The Ministry of Health, census, and Rufiji Demographic Surveillance System estimates range from 31,000 to 36,000 children for 2003. Because of this discrepancy, the decline in the total fertility rate, from 5 to 4.7 (5 percent), is taken into account in estimating the number of children less than five years of age for 1999 to 2001. The probability of dying before five years of age declined by 15.6 percent, and because of declining fertility, each year has 1.5 percent Table 54.5 Costs of Evidence-Based Resource Allocation for Improving Health Total cost at US$2 per capita (US$) Discounted cost (US$) Year Population 1998 186,809 373,618 420,510 1999 191,012 382,024 417,448 2000 196,515 393,030 416,966 2001 202,176 404,352 416,482 2002 208,000 Total costs 416,000 416,000 1,969,024 2,087,406 Source: Authors. fewer children than the preceding year. DALYs saved from each child death averted is estimated at 41.2. The resulting calculations of effectiveness are summarized in table 54.4. The estimates of cost are based on population size projected back from the 2002 census results, assuming an average annual growth rate of 2.8 percent. Costs incurred in 1998 are included because we assume that it takes at least two years (1998 to 2000) of improving the HIS before health benefits accrue. 1026 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 166 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1027 Cost per DALY saved (US$) 800 700 600 500 400 300 200 100 0 5 10 20 30 Percentage increase in hepatitis B immunization coverage attributable to investment in HIS 40 High-cost setting (US$2.00 per capita) Low-cost setting (US$0.80 per capita) capita less than US$997; World Bank 2002). Figure 54.1 shows that, for the high-prevalence countries (Miller and McCann 2000), the investment in a comprehensive HIS is highly costeffective (US$159 to US$126 per DALY saved for low-cost settings and US$757 to US$597 per DALY saved for high-cost settings), even if the investment results in only minor increases in immunization coverage. A similar analysis for countries with a lower prevalence rate of hepatitis B demonstrates that the cost per DALY saved is higher, but the investment in an HIS still yields a savings of DALYs at a cost that is well below the GNP per capita for the majority of the low-income countries. These calculations of the cost-effectiveness of investments in an HIS are highly conservative, because they consider health benefits within a single population group (children, in the case of TEHIP) or a single disease problem (hepatitis B). They therefore underestimate the true cost-effectiveness of investment in an HIS, which can drive improvements in program efficiency and effectiveness across a broader range of health interventions. Source: Authors. Figure 54.1 Cost-Effectiveness of Health Information Systems: Cost per DALY Saved Because of Increases in Coverage Attributable to HIS FINANCING OF IMPROVED HEALTH INFORMATION Using these figures for effectiveness (table 54.4) and cost (table 54.5), we find that the cost-effectiveness of the HIS that results in improved evidence-based resource allocation and child health may be conservatively estimated at US$68.50 per DALY gained (US$2,087,406 to gain 30,457 DALYs). Even in the poorest countries, this is well below the gross national product (GNP) per capita benchmark for what is considered worthwhile for government investment in health. This analysis for the TEHIP project is based solely on child deaths averted. But the improvement in health information would also yield substantial benefits for adult populations. For example, HIS-driven increases in coverage with hepatitis B vaccine have varied between 5 and 33 percent (Miller and McCann 2000). These increases in coverage with hepatitis B immunization will result in incremental reductions in death and disability among adults attributable to hepatitis B–induced cirrhosis and liver cancer, thereby averting the loss of substantial numbers of DALYs in low-income countries (World Bank 2002). Hepatitis B vaccine is a cost-effective addition to an existing immunization program, with a cost per death averted of US$11 to US$15 (US$193 to US$262 per DALY saved). But efficiency and coverage can be substantially improved with an additional investment in the HIS. The cost per DALY saved by incremental investment in the HIS can be calculated using estimates of costs of the HIS from table 54.2, plus the estimates of cost and deaths averted because of immunization from Miller and McCann (2000) for populations in all low-income countries (GNP per The annual per capita cost, estimated earlier, of US$0.53 to US$2.99 for a comprehensive HIS, represents a substantial portion of the current per capita health expenditure for many developing countries. These figures include capital and recurrent costs, although they do not include the costs of any external technical assistance. Because most countries have already made a substantial investment in a HIS, the actual incremental costs to improve the existing HIS likely are much less. Salaries, which account for more than 90 percent of HIS costs, are expenditures that are already being made in most settings, so the marginal cost of HIS improvements would be primarily the initial development costs of planning, training, technical assistance, and information technology upgrades. Furthermore, the costs of HIS improvements may be fully offset or even exceeded by the savings from the resulting improvements in efficiencies in the health care system. Existing funding is adequate to strengthen systems substantially in all low-income and lower-middle-income countries primarily through the major international initiatives (Global Fund to Fight AIDS, Tuberculosis, and Malaria; President’s Emergency Plan for AIDS Relief; and Multi-country AIDS Program of the World Bank). All these funders recommend that 3 to 7 percent of grants and loans be allocated to monitoring and evaluation. Several bilateral development agencies and the multilateral development banks will provide financing for HIS reform, including the U.S. Agency for International Development (USAID) through the MEASURE (Monitoring and Evaluation to Assess and Use Results) Project, which is Information to Improve Decision Making for Health | 1027 ©2006 The International Bank for Reconstruction and Development / The World Bank 167 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1028 designed to improve and institutionalize the collection and use of data for health policy development and program monitoring. The HMN offers some financial assistance to countries that are preparing for and planning HIS reform and will assist countries in negotiating financing packages that blend loan funding with grants from bilateral donors to implement those reforms. Several international agencies support strengthening systems for national statistics that extend beyond the health sector. STATCAP (Statistical Capacity Building), which is a new lending program offered by the Partnership in Statistics for Development in the 21st Century (PARIS21) through the World Bank, supports the development of national statistical systems. The separate Trust Fund for Statistical Capacity Building offers smaller amounts of grant funding to prepare the statistical master plan that is required for obtaining a STATCAP loan. Although short-term project funding can often be secured for system development, the resulting system and its recurrent costs must be within the country’s capacity to sustain it, both technically and financially. opment of the district-level HIS. The HMN will create an alliance of countries committed to a parsimonious consensus technical framework and encourage donors to cooperate with and strengthen the HMN-sanctioned HIS architecture in participating countries. The predictors of success in developing and maintaining an HIS are as follows: • high-level commitment to HIS development and the linked changes in management • a champion of HIS reform who engages the stakeholders and can work across sectors • an information architecture that is simple, is structured to drive decision making at the level that data are collected, provides incentives and accountability for performance, and links health information subsystems • investment in training and increased status for the people who manage the HIS. RESEARCH AND DEVELOPMENT IMPLEMENTATION OF CHANGE: LESSONS OF EXPERIENCE Underinvestment is the root cause of the nearly universal weaknesses in the HIS in developing countries. This failure is reflected in the poorly paid and undervalued HIS staff; in the irregular and unreliable transmittal of data from the periphery; in the underreporting of events, including births, deaths, and morbidity; and in the failures to base planning and decision making—at both the district and the central levels—on credible evidence (Azubuike and Ehiri 1999). When the need for HIS improvement is identified, ministries of health should explicitly state the characteristics they need in a reformed system and quantify the expected benefits. A common mistake made in implementing HIS change is failing to recognize the associated need for change in management processes and organizational culture. In contrast, recent HIS reforms in Niger (Mock and others 1993) and Uganda (Gladwin, Dixon, and Wilson 2003) have had unprecedented success because they have been aligned with broader management reforms and changes in organizational culture. Failure to adjust management roles with HIS changes can constrain effectiveness, such as when HIS managers are not given the necessary increased status and authority to demand reports and trigger corrective actions (Gladwin, Dixon, and Wilson 2003). Failure to invest adequately in training, especially in skills for presentation and communication of results, may also inhibit the use of health information. The demand from international organizations and global programs, such as the Expanded Program on Immunization and Stop TB, for reports on vast numbers of indicators has retarded the smooth devel- An effective HIS delivers routine information that enables informed policy making and management but also promotes health research. Routine information systems may serve as a research platform, but the HIS itself should also be a subject of research. Research should drive the continual refinement of HIS methods and tools, thereby ensuring expanding and welldocumented returns on our investments in health. The instruments and methods of the HIS must be continually refined to improve its effectiveness and reduce its costs. For the phased introduction of vital events monitoring, for example, there is a pressing need for the development and validation of methods for projecting subnational results to national rates of birth and death. More research is needed to develop and test new methods for rapid assessment in order to obtain timely and affordable information to solve management problems. As field-appropriate and cost-effective diagnostic technologies are developed, research should be performed to document the utility of obtaining biomarkers in household surveys. Documenting improved outcomes and lower costs will provide evidence for policy makers on the effectiveness of HIS investments. To better decide how to improve the HIS, decision makers will need documentation of the costs and effects of introducing ICT in support of the HIS. Existing and emerging technologies should be tested for their cost and effectiveness in assisting field-based data capture, instantaneous data transmission, GIS-based mapping of indicators, and compelling presentation for decision making by policy makers, managers, and other stakeholders. Research and development efforts are needed to devise software—or preferably shareware—that is specifically tailored to support the consensus technical framework developed by the HMN. 1028 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 168 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1029 Bailey and Pang (2004) point out the need for more research in the developing world to better understand users’ information needs. In fact, research is needed to better document the entire information value chain, with special attention to improving the identification of information needs, to overcoming the natural disincentives to information sharing, and to enabling better use of information for constructive change. At present, there is still a need to improve the access to information and knowledge in the developing world. However, the future will bring the larger challenge of improving the management and use of information and the knowledge such information can bring. Research in the HIS will be instrumental in both accelerating equitable access to information and improving the management and use of knowledge for improved health. CONCLUSIONS More than ever before, it is in the mutual interest of the developing and industrial worlds to invest in strengthening systems for collection and management of health information (Stansfield 2005). The trend toward “basket” funding for health and sectorwide approaches makes the need for priority setting all the more acute. Priority setting depends on accurate information. The success of efforts to reduce poverty and health inequity will depend on the existence of information systems to detect those problems, facilitate the design of solutions, and track progress toward eliminating the problems. Countries and donors must, therefore, accelerate and harmonize their investments in information systems. Within countries, the trend toward decentralization of authority for management of health resources has led to further challenges for the HIS, as well as to greater reliance on the information it provides to inform decision making. It is clear from the instructive failures of underresourced systems that the accuracy and value of information reported to the national level will depend on that information’s perceived value in the periphery. Information is relevant only if it is used to solve a local problem or if it helps to generate innovation that solves a local problem (Bailey and Pang 2004). Therefore, the decentralization of authority will be successful only with better information systems to support decisions at the periphery, and evidence-based decision making will be possible only if authority can be devolved to the periphery. This decentralization, along with increasing cooperation and collaboration across sectors to improve health outcomes, makes it all the more critical to present data in simpler ways that are understandable and compelling to a broader and nontechnical audience. Although historically neglected, investments in comprehensive development of the HIS will clearly deliver good value for money. Improvements in the HIS can accelerate broad improvements in health if they are engineered to reflect, reinforce, and even drive health sector reforms. Even more compellingly, investments in the HIS can make health the “thin edge of the wedge,” giving governments and politicians a positive experience with information sharing and overcoming the natural disincentives to transparency and accountability. HIS investments hold the promise, therefore, not only of transforming public health, but also of accelerating progress toward good governance in every sector. NOTES 1. Sauerborn and Lippeveld (2000) have defined a health information system as the “set of components and procedures organized with the objective of generating information that will improve health management decisions at all levels of the health system.” 2. The dollar amounts given are quoted from the references and are not adjusted for current dollar value. REFERENCES AbonZahr, C., and T. Boerma. 2005. “Health Information Systems: The Foundations of Public Health.” Bulletin of the World Health Organization 83 (8): 578–83. Abt Associates. 2003. Primer for Policymakers—Understanding National Health Accounts: The Methodology and Implementation Process. Bethesda, MD: Partners for Health Reformplus Project, Abt Associates. Azubuike, M. C., and J. E. Ehiri. 1999. “Health Information Systems in Developing Countries: Benefits, Problems, and Prospects.” Journal of the Royal Society for the Promotion of Health 119 (3): 180–84. Bailey, C., and T. Pang. 2004. “Health Information for All by 2015?” Lancet 364 (9430): 223–24. Boerma, Ties, J., E. Holt, and R. Black. 2001. “Measurement of Biomarkers in Surveys in Developing Countries: Opportunities and Problems.” Population and Development Review 27 (2): 303–14. Cibulskis, R. E., and G. Hiawalyer. 2002. “Information Systems for Health Sector Monitoring in Papua New Guinea.” Bulletin of the World Health Organization 80 (9): 752–58. Commission on Macroeconomics and Health. 2000. De, S., T. Dmytraczenko, D. Brinkerhoff, and M. Tien. 2003. Has Improved Availability of Health Expenditure Data Contributed to Evidence-Based Policymaking? Country Experiences with National Health Accounts. Technical Report 022. Bethesda, MD: Partners for Health Reformplus Project, Abt Associates. Duran-Arenas, L., C. C. Rivero, S. F. Canton, R. S. Rodriquez, F. Franco, R. W. Luna, and J. Catino. 1998. “The Development of a Quality Information System: A Case Study of Mexico.” Health Policy and Planning 13 (4): 446–58. Economic Commission for Africa. 2003. “Workshop on Organization and Management of Statistical Systems.” Report on workshop in Addis Ababa, December 8–12. http://www.unstats.un.org/unscl/ methods/statorg/workshops/AddisAbaba/presentation_session8b_ financing.pdf. Gladwin, J., R. A. Dixon, and T. D. Wilson. 2003. “Implementing a New Health Management Information System in Uganda.” Health Policy and Planning 18 (2): 214–24. Godlee, F., N. Pakenham-Walsh, D. Ncayiyana, B. Cohen, and A. Packer. 2004. “Can We Achieve Health Information for All by 2015?” Lancet 364 (9430): 295–300. Information to Improve Decision Making for Health | 1029 ©2006 The International Bank for Reconstruction and Development / The World Bank 169 DCP_1017-1030.qxd 2/10/06 12:14 PM Page 1030 Jamison, D. T., J. Sachs, and J. Wang. 2001. “Mortality Changes and Economic Welfare in Sub-Saharan Africa, 1960–2000.” Commission on Macroeconomics and Health, Background Paper for Working Group 1, World Health Organization, Geneva. Pappaioanou, M., M. Malison, K. Wilkens, B. Otto, R. A. Goodman, R. E. Churchill, and others. 2003. “Strengthening Capacity in Developing Countries for Evidence-Based Public Health: The Data for Decision Making Project.” Social Science and Medicine 57: 1925–37. Jolly, R. 2002. “Statisticians of the World Unite: The Human Development Challenge Awaits.” Journal of Human Development 3 (2): 263–72. PARIS21 (Partnership in Statistics for Development in the 21st Century). 2004. “Meeting the Data Challenge: A Funding Proposal for PARIS21 and the Trust Fund for Statistical Capacity Building for 2004 to 2006.” PARIS21, Paris. Kleinau, E. 2000. “Management of Health Information Systems.” In Design and Implementation of Health Information Systems, ed. T. Lippeveld, R. Sauerborn, and C. Bodart. Geneva: World Health Organization. Kohn, L., J. Corrigan, and M. Donaldson, eds. 2001. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press. Perrin, E. B., W. D. Kalsbeek, and T. M. Scanlan, eds. 2004. Toward a Health Statistics System for the 21st Century. Summary of a workshop, Division of Behavioral and Social Sciences and Education, National Research Council and Committee on National Statistics, National Academy of Sciences. Washington, DC: National Academy Press. Lescano, A. G., M. Ortiz, R. Elgegren, E. Gozzer, E. Saldarriaga, I. Soriano, and others. 2003. “Alerta DISAMAR: Innovative Disease Surveillance in Peru.” Paper presented at the American Society of Tropical Medicine and Hygiene, Philadelphia, December 5. Peters, D. H., A. E. Elmendorf, K. Kandola, and G. Chelleraj. 2000. “Benchmarks for Health Expenditures, Services, and Outcomes in Africa during the 1990s.” Bulletin of the World Health Organization 78 (6): 761–68. Loevinsohn, B. P. 1994. “Data Utilization and Analytical Skills among MidLevel Programme Managers in a Developing Country.” International Journal of Epidemiology 23 (1): 194–200. Petrakos, C. 1998. “Finding a Cure: Disease Management Aids the Search for Better Outcomes.” Modern Physician, September 1, 2004. http://www.modernphysician.com. Mackay, K. 1998. “Public Sector Performance: The Critical Role of Evaluation.” In Public Sector Performance—The Critical Role of Evaluation: Selected Proceedings of a World Bank Seminar, ed. K. Mackay, ix–xvi. Washington, DC: World Bank. Rommelmann, V., P. Setel, Y. Hemed, H. Mponezya, G. Angeles, and T. Boerma. 2004. “Costs and Results of Information Systems for Poverty Monitoring, Health Sector Reform, and Local Government Reform in Tanzania.” MEASURE Evaluation, Adult Morbidity and Mortality Project, University of Newcastle upon Tyne, U.K., and University of North Carolina, Chapel Hill. McLellan, F. 2001. “Information Technology Can Benefit Developing Countries.” Lancet 358 (9278): 308. Miller, M. A., and L. McCann. 2000.“Policy Analysis of the Use of Hepatitis B, Haemophilus influenzae type B, Streptococcus PneumoniaeConjugate and Rotavirus Vaccines in National Immunization Schedules.” Health Economics 9: 19–35. Mock, N., J. Setzer, I. Sliney, G. Hadizatou, and W. Bertrand. 1993. “Development of Information-Based Planning in Niger.” International Journal of Technology Assessment in Health Care 9 (3): 360–68. Murray, C. J. L., B. Shengelia, N. Gupta, S. Moussavi, A. Tanjon, and M. Thieren. 2003. “Validity of Reported Vaccination Coverage in 45 Countries.” Lancet 362 (9389): 1022–27. Nordhaus, W. 2002. “The Health of Nations: The Contribution of Improved Health to Living Standards.” Cowles Foundation Discussion Paper 1355, Yale University, New Haven. O’Carroll, P. W. 2003. “The Context for Public Health Informatics.” In Public Health Informatics and Information Systems, ed. P. W. O’Carroll, W. Yasnoff, M. E. Ward, L. H. Ripp, and E. L. Martin. New York: Springer. Sauerborn, R., and T. Lippeveld. 2000. “Why Health Information Systems?” In Design and Implementation of Health Information Systems, ed. T. Lippeveld, R. Sauerborn, and C. Bodart. Geneva: World Health Organization. Stansfield, S. 2005. “Structuring Information Systems to Improve Health.” Bulletin of the World Health Organization 83 (8): 562. Thacker, S. B., D. F. Stroup, R. G. Parrish, and H. A. Anderson. 1996. “Surveillance in Environmental Public Health: Issues, Systems, and Sources.” American Journal of Public Health 86 (5): 633–38. Walsh, J. A., and M. Simonet. 1995. “Data and Data Needs for Health Sector Reform.” Health Policy 32 (1–3): 295–306. World Bank. 1993. World Development Report 1993: Investing in Health. Washington, DC: World Bank. ———. 2002. World Development Indicators. Washington, DC: World Bank. Zayan, A., W. Berggren, and F. Doumbia. 1992. The Price of Immunization and the Value of Information. Westport, CT: Save the Children. PAHO (Pan American Health Organization). 2004. “Ten Year Evaluation of the Regional Core Health Data Initiative.” Epidemiological Bulletin 25 (3): 1–7. 1030 | Disease Control Priorities in Developing Countries | Sally K. Stansfield, Julia Walsh, Ndola Prata, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 170 Chapter 58 School-Based Health and Nutrition Programs Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, Kathleen Beegle, Amaya Gillespie, Lesley Drake, Seung-hee Frances Lee, Anna-Maria Hoffman, Jack Jones, Arlene Mitchell, Delia Barcelona, Balla Camara, Chuck Golmar, Lorenzo Savioli, Malick Sembene, Tsutomu Takeuchi, and Cream Wright The paradigmatic shift in the past decade in our understanding of the role of health and nutrition in school-age children has fundamental implications for the design of effective programs. Improving the health and nutrition of schoolchildren through school-based programs is not a new concept. School health programs are ubiquitous in high-income countries and most middle-income countries. In low-income countries, these programs were a common feature of early, particularly colonial, education systems, where they could be characterized as heavily focused on clinical diagnosis and treatment and on elite schools in urban centers. This situation is changing as new policies and partnerships are being formulated to help ensure that programs focus on promoting health and improving the educational outcomes of children, as well as being socially progressive and specifically targeting the poor, girls, and other disadvantaged children. This evolution reflects five key changes in our understanding of the role of these programs in child development. • First, ensuring good health at school age requires a life cycle approach to intervention, starting in utero and continuing throughout child development. In programmatic terms this requirement implies a sequence of programs to promote maternal and reproductive health, management of childhood illness, and early childhood care and development. Promoting good health and nutrition before and during school age is essential to effective growth and development. • Second, operations research shows that the preexisting infrastructure of the educational system can often offer a more cost-effective route for delivery of simple health interventions and health promotion than can the health system. Low-income countries typically have more teachers than nurses and more schools than clinics, often by an order of magnitude. • Third, empirical evidence shows that good health and nutrition are prerequisites for effective learning. This finding is not simply the utopian aspiration for children to have healthy bodies and healthy minds, but also the demonstration of a systemic link between specific physical insults and specific cognitive and learning deficits, grounded in a new multisectoral approach to research involving public health and epidemiology, as well as cognitive and educational psychology. • Fourth, the provision of quality schools, textbooks, and teachers can result in effective education only if the child is present, ready, and able to learn. This perception has additional political momentum as countries and agencies seek to achieve Education for All (EFA) by 2015 and address the Millennium Development Goals of universal basic education and gender equality in education access. If every girl and boy is to be able to complete a basic education of good quality, then ensuring that the poorest children, who suffer the most malnutrition and ill health, are able to attend and stay in school and to learn while there is essential. 1091 ©2006 The International Bank for Reconstruction and Development / The World Bank 171 Morbidity as a proportion of peak value 1 0.9 0.8 Ascaris 0.7 0.6 School-age children 0.5 0.4 Schistosoma haematobium Cerebral malaria 0.3 0.2 0.1 Diarrhea 0 0 5 10 15 Age (years) 20 25 30 Source: Bundy and Guyatt 1996. Figure 58.1 Age Distribution of Infection-Specific Morbidity • Finally, education, including education that promotes positive health behaviors, contributes to the prevention of HIV/AIDS—the greatest challenge for generations to come. School health and nutrition programs that help children complete their education and develop knowledge, practices, and behaviors that protect them from HIV infection as they mature have been described as a “social vaccine” against the disease. Because of the success of child survival programs, the number of children reaching school age (defined as 5 to 14 years of age) is increasing and is estimated to be 1.2 billion children, with 88 percent living in less developed countries (U.S. Census Bureau 2002). As figure 58.1 illustrates, the pattern of disease is age specific. A large body of evidence shows that these conditions affect cognition, learning, and educational achievement (see Jukes, Drake, and Bundy forthcoming; Pollitt 1990 for reviews of this extensive literature). This chapter focuses on the health, nutrition, and education of the school-age child and on the programs that can be implemented at school age to promote positive outcomes. INFECTIOUS DISEASE AND SCHOOL-AGE CHILDREN A range of infectious diseases affect school-age children. Helminth Infections Between 25 and 35 percent of school-age children are estimated to be infected with one or more of the major species of worms (Bundy 1997; see also chapter 24). The most common and important infections are caused by geohelminths (the roundworm Ascaris, the whipworm Trichuris, and the two species of hookworms Ancylostoma and Necator) and by the schistosomes (Schistosoma spp.), which give rise to a wide range of chronic but largely nonspecific symptoms. The most intense worm infections and related illnesses occur at school age (Partnership for Child Development 1998b, 1999) and account for some 12 percent of the total disease burden and 20 percent of the loss of disability-adjusted life years (DALYs) from communicable disease among schoolchildren (World Bank 1993). Infected schoolchildren perform poorly in tests of cognitive function; when they are treated, immediate educational and cognitive benefits are apparent only for children with heavy worm burdens or with concurrent nutritional deficits. Treatment alone cannot reverse the cumulative effects of lifelong infection or compensate for years of missed learning, but studies suggest that children are more ready to learn after treatment for worm infections and may be able to catch up if this learning potential is exploited effectively in the classroom (Grigorenko and others forthcoming). In Kenya, treatment reduced absenteeism by one-fourth, with the largest gains for the youngest children who suffered the most ill health (Miguel and Kremer 2004). Malaria Up to 5 percent of children infected with malaria early in life have residual neurological sequelae (Snow 1999). In areas of unstable transmission, malaria accounts for 10 to 20 percent of all-cause mortality among school-age children (Bundy and others 2000), and those who have suffered repeated attacks have poorer cognitive abilities. In Kenya, primary school students miss 11 percent of school days because of malaria, equivalent to 4 million to 10 million days per year (Brooker and others 2000). Oral antimalarial treatment reduced school absenteeism by 50 percent in Ghana (Colbourne 1955); the use of insecticidetreated bednets in Tanzania reduced malaria and increased attendance (Shiff and others 1996). Girls in The Gambia were more than twice as likely to enroll in primary school if they had received malaria prophylaxis in early childhood (Jukes and others submitted). HIV/AIDS Although school-age children have the lowest infection prevalence of any age group (figure 58.2), an estimated 3.8 million children under 15 years of age have been infected with HIV and more than two-thirds have died (UNAIDS 2002). Even uninfected children suffer physically, socially, and psychologically through death or illness in their family (World Bank 2002). The proportion of orphans, most of whom are of school age, has risen from 2 to 15 percent in some African countries, with 1092 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 172 Acute Respiratory Infection a. Male Cases Acute respiratory infection, the most common acute infection in school-age children globally, is a significant cause of absenteeism. Research in industrial countries (Cohen and Smith 1996) finds that flu infection affects attention and reaction time; colds primarily affect hand-eye coordination, as well as reduce the ability to tolerate high levels of noise and other distractions common to the classroom. Percent infected in each age group, as a percentage of cases 45 40 35 30 25 20 MALNUTRITION, NONINFECTIOUS DISEASE, AND HEALTH AND EDUCATION 15 10 Malnutrition and noninfectious disease also affect school-age children. 5 0 0–4 5–14 15–19 20–29 30–39 Age group 40–49 50⫹ Malnutrition Stunting (low height for age) is a physical indicator of chronic or long-term malnutrition, whereas underweight (low weight for age) is an indicator of both chronic and acute malnutrition. Both are common in school-age children (figure 58.3). Girls who are better nourished are more attentive and more involved during class, and boys have improved classroom behavior and increased activity levels. One Z-score increase in height for age is associated with an increase of 0.1 standard deviation (SD) in tests of arithmetic and language. Stunted children enroll in school later than other children. School foodservice programs have been successful in improving school attendance. b. Female Cases Percent infected in each age group, as a percentage of cases 50 45 40 35 30 25 20 15 10 Z-score 0 5 0 0–4 5–14 15–19 Botswana Tanzania 20–29 30–39 Age group Côte d’Ivoire Zimbabwe 40–49 50⫹ Malawi Source: UNAIDS epidemiological fact sheets 2000. Note: Figure shows percentage of males (top) and females (bottom) infected with HIV in each age group (as a percentage of all HIV-infected males and females, respectively), for five countries in Africa. Infection peaks at a younger age in women than in men, and the lowest prevalence of infection occurs in school-age children. ⫺0.5 ⫺1.0 ⫺1.5 ⫺2.0 ⫺2.5 ⫺3.0 Figure 58.2 Age Prevalence of HIV/AIDS ⫺3.5 6 AIDS accounting for 50 percent of this increase. The number of orphans is expected to reach more than 25 million by 2010. School-age children with HIV infections have lower IQ levels and poorer academic achievement, language, and visual motor functioning. These deficits can be reduced or reversed with antiretroviral therapy. The improvement is greater for children of school age than for younger children. 7 8 9 10 11 12 13 Age (years) Ghana Tanzania 14 India Vietnam 15 16 17 18 Indonesia Source: Data from Partnership for Child Development 1998a. Note: Z-scores of less than ⫺2 indicate stunting. Figure 58.3 Mean Z-Scores of Height-for-Age of Boys in Five Countries School-Based Health and Nutrition Programs | 1093 ©2006 The International Bank for Reconstruction and Development / The World Bank 173 Short-Term Hunger Hunger, which reduces ability to perform school tasks, is readily reversed by feeding. Children age 11 to 13 years in Jamaica improved their scores on arithmetic tests after one semester of receiving breakfast at school because they attended more regularly and studied more effectively (Simeon 1998). Missing breakfast impairs performance to a greater extent for children of poor nutritional status, who also benefit most from food intervention (Pollitt, Cueto, and Jacoby 1998; Simeon and Grantham McGregor 1989). Micronutrient Deficiency Micronutrient deficiencies may take several different forms, each with negative impacts on children’s ability to perform well in school. Iron Deficiency. Iron deficiency, the most common form of micronutrient deficiency in school-age children, is caused by inadequate diet and infection, particularly by hookworm and malaria (Hall, Drake, and Bundy 2001). More than half the school-age children in low-income countries are estimated to suffer from iron deficiency anemia (Partnership for Child Development 2001). Children with iron deficiency score 1 to 3 SD worse on educational tests and are less likely to attend school. Iron supplementation reduces these deficits. Iodine Deficiency. Iodine deficiency affects an estimated 60 million school-age children; studies indicate prevalence rates between 35 and 70 percent. Iodine deficiency is related to lowered general cognitive abilities and tests scores. No conclusive evidence shows that iodine supplementation improves cognitive abilities in this age group (Huda, GranthamMcGregor, and Tomkins 2001). Vitamin A Deficiency. Vitamin A deficiency affects an estimated 85 million school-age children. The deficiency, which causes impaired immune function and increases risk of mortality from infectious disease, is an important cause of blindness. Recent studies suggest that this deficiency is also a major public health problem in school-age children. Multiple-micronutrient supplements have improved cognitive function and short-term memory in schoolchildren and have reduced absenteeism caused by diarrhea and respiratory infections. Obesity An estimated 17.6 million children worldwide are overweight. Obesity is associated with underperformance in education. In low-income countries obesity is still rare, but the prevalence in the children of many middle-income countries is similar to that in the United States. ESTIMATING THE BURDEN OF DISEASE The cost per DALY of school health programs has been estimated at US$20 to US$34, implying that the programs are at least as cost-effective as many other public health “best buys” (Bobadilla and others 1994). However, current methods of estimating the burden for school-age children result in a significant underestimation of both the developmental consequences of disease and malnutrition at school age and the overall benefits for health and development of school health and nutrition programs. There are two key reasons for this underestimation. The first issue relates to time scales. Many serious diseases in adulthood, including heart disease and carcinomas, are a consequence of unhealthy practices established in early life. This later burden can be substantially and cost-effectively averted by early intervention, particularly by school-based life-skills programs. For example, in the United States (Del Rosso and Marek 1996), US$1 invested can avert US$18.80 spent on the later problems caused by tobacco and US$5.70 on problems of drug and alcohol abuse. DALY estimates cannot capture these downstream consequences of upstream intervention and instead attribute the disease burden to the adult age group in which it appears. This kind of estimate is particularly misleading in the case of HIV/AIDS, for which prevention education at school age is effective in averting later infection and disease (World Bank 2002), and in the case of estimates of intergenerational effects, in which ensuring the health of an adolescent girl may help secure the health of her baby born a few years later. The second issue is illustrated by experience with helminth infections. In 1990, the burden was first estimated at 18 million DALYs, close to the value for tuberculosis, measles, and malaria. This estimate reflected the ubiquity of infection and the longterm consequences of cognitive impacts. In 2001, the estimate was only 4.7 million DALYs (WHO 2003), and during the intervening years one estimate put the value as low as 2.6 million. This extraordinary variability is caused in part by different emphases on the cognitive and health impacts and illustrates how, for very common conditions, even minor changes in disability weight can affect the overall values. This variability also reflects the importance of a sectoral perspective, because the low estimates reflect a focus on health, whereas the higher estimates include impact on educational achievement and child development. The scale of the burden of disease in terms of cognition is illustrated by estimating the impact of stunting, anemia, and helminths on the cognition of the estimated 562 million schoolage children in developing countries. According to typical 1094 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 174 deficits in test scores attributable to these diseases, the total global loss of points ranges from 600 million to 1.8 billion IQ points, an additional 15 million to 45 million cases of mental retardation (defined here as IQ less than 70), and a loss of between 200 million and 524 million years of primary schooling (Jukes, Drake, and Bundy forthcoming). Although the precision of these striking figures may be open to debate, they clearly show that even minor cognitive deficits resulting from ubiquitous conditions can result in an extraordinarily large scale of effect. INTERVENTIONS In light of the significant effects of ill health and malnutrition on educational outcomes, the role of effective health promotion and simple school-based programs to deliver low-cost interventions becomes increasingly important (Bundy and others 1992). Other chapters provide information on the integrated management of childhood illness, early child development, and adolescent health (see chapters 63, 27, and 59, respectively). The focus here is on ill health and malnutrition at school age and the role of the formal and nonformal education sector in delivering interventions. Developing a Programmatic Approach The focus of school health and nutrition programs in lowincome countries has shifted significantly over the past two decades away from a medical approach that favored elite schools in urban centers and toward an approach that improves health and nutrition for all children, particularly the poor and disadvantaged. This change began in the 1980s, when research showed not only that school health and nutrition programs were important contributors to health outcomes but also that they were essential elements of efforts to improve education access and completion, particularly for the poor. In an effort to reconceptualize the relationship between health and education, the United Nations Education, Scientific, and Cultural Organization (UNESCO) hosted a series of workshops on this topic in the 1980s (Bundy 1989; Halloran, Bundy, and Pollitt 1989) and supported one of the first authoritative reviews of the area (Pollitt 1990). Similarly, the United Nations Development Programme, in conjunction with the Rockefeller, Edna McConnell Clark, and J. S. McDonnell Foundations supported the creation of the Partnership for Child Development to strengthen the evidence base across the education and health sectors and to support the dissemination of information (Berkley and Jamison 1990; Bundy and Guyatt 1996). This paradigm shift coincided with the World Conference on Education for All in Jomtien, Thailand, in 1990 and led to renewed efforts by countries and agencies to develop more effective programmatic approaches to school health and nutrition. The United Nations Population Fund (UNFPA) has pioneered population and family life education (PopEd) as an intrinsic part of school curricula. In 1994, the International Conference on Population and Development placed specific emphasis on school health, including reproductive and sexual health. Efforts at country level have addressed PopEd both within the school system and outside, and the concept has evolved to include references to family life education, sex education, HIV/AIDS awareness and prevention, and life-skills programs. Today, approximately 84 countries have UNFPAsupported school health programs. In 1995, the World Health Organization (WHO) launched its Global School Health Initiative to foster the development of health-promoting schools (HPSs) (WHO 1996). The concept started in Europe in the early 1990s, based on the Ottawa Charter of Health Promotion (WHO 1986; European Commission 1996), which recognized that health is created by caring for oneself and others, by being able to make decisions and have control over one’s life and circumstances, and by creating conditions that support health for all. WHO’s European Regional Office, the Council of Europe, and the Commission of the European Communities widely promoted the concept of HPSs to foster healthy lifestyles and develop environments conducive to health (European Commission, WHO Europe, and Council of Europe 1996). Although definitions vary among regions, countries, and schools, an HPS may be characterized as one that is constantly strengthening its capacity as a healthy setting for living, learning, and working. The initiative fosters the development of HPSs by the following: • consolidating research and expert opinion to describe the nature and effectiveness of school health programs • building capacity to advocate for the creation of HPSs and to apply the components to priority health issues • strengthening collaboration and national capacities to assess the prevalence of important health-related behaviors and conditions and to plan and implement policies and programs that improve health through schools • creating networks and alliances, including regional networks. The key elements of how this approach is interpreted today are listed in table 58.1. In the mid 1990s, the United Nations Children’s Fund (UNICEF) began promoting the Child-Friendly Schools framework as a holistic way to promote children’s rights as expressed in the Convention on the Rights of the Child (UNICEF 1990) and children’s access to education as stated in the World Declaration of Education for All (UNESCO 1990). This approach included a gender-sensitive component, which was further strengthened when girls’ education became the first priority in UNICEF’s Medium Term Strategic Plan, 2002–5. Another key element is skills-based health education, School-Based Health and Nutrition Programs | 1095 ©2006 The International Bank for Reconstruction and Development / The World Bank 175 Table 58.1 Characteristics of Agency-Specific School Health and Nutrition Programs, within the FRESH framework FRESH framework Health-promoting schools (WHO) Child-friendly schools (UNICEF) Policy Respects an individual’s well-being and dignity Respects and realizes the rights of every child Provides multiple opportunities for success Acts to ensure inclusion, respect, and equality of opportunity for all children Acknowledges good efforts and intentions as well as personal achievements PopEd (UNFPA) Global school feeding campaign (World Food Program) Creates a supportive and enabling policy environment for reproductive health and HIV prevention for young people Focuses on the poorest and most food-insecure communities. Protects young people from early and unwanted pregnancy, sexually transmitted diseases, sexual abuse, and violence Serves as platform for essential package approach that includes water, sanitation, and environmental measures Strengthens HIV/AIDS and sexual Provides education that is affordable and reproductive health education programs and accessible Supports learning through good nutrition Gives priority to girls and AIDS-affected children Is gender sensitive and girl friendly Is flexible and responds to diversity Sees and understands the whole child in a broad context Enhances teacher capacity, morale, commitment, and status School environment Is healthy Is healthy, safe, and secure Provides opportunities for physical education and recreation Is protective emotionally and psychologically Education Provides skills-based health education Promotes quality learning outcomes Fosters health and learning Promotes access to education Provides skills-based health education, including life skills relevant to children’s lives Services Provides school health services Promotes physical health Provides nutrition and foodsafety programs Promotes mental health Ensures access to youth-friendly sexual and reproductive health services Provides food Targets young people in school and out of school Promotes community and school partnerships Promotes and supports deworming Provides programs for counseling, social support, and mental health promotion Provides health promotion programs for staff Includes school and community projects and outreach Supportive partnerships Engages health and education Is child centered officials, teachers, teachers’ Is family focused unions, students, parents, health Is community based providers, and community leaders in efforts to make the school a healthy place Ensures active participation of parents, youths, community leaders, and organizations Source: Summarized from World Bank Fresh Toolkit (2000), WHO (1996), and personal communications from Arlene Mitchell and Sheldon Shaeffer (May 2005). including life skills, which has been promoted through UNICEF with partner organizations as part of HPSs, childfriendly schools, and the framework for Focusing Resources on Effective School Health (FRESH). Research shows that this approach is more effective than traditional strategies, which tend to be didactic and to focus on scientific information alone. In contrast, skills-based health education uses the experiences of students as the starting point and explores the links between knowledge, attitudes, and the interpersonal skills required to promote health and learning (UNICEF, WHO, World Bank, UNFPA, UNESCO 2003). The approach is interactive, activity based, and flexible so that it can be used to address a range of health and social issues, including HIV/AIDS, sanitation, drug use, violence and bullying, nutrition, and cross-cutting issues such as gender and culture. Some key elements of how the child-friendly schools approach is interpreted currently, including its focus on healthy and protective learning environments, are listed in table 58.1. 1096 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 176 Also during the 1990s, the World Bank Human Development Network sought to support countries in implementing school health and nutrition programs (Del Rosso and Marek 1996; World Bank 1993) and launched an International School Health Initiative with the aim of raising awareness among decision makers in the education sector. Thus, the 1990s were characterized by the creation of a number of apparently separate programs to promote and support school health. However, analysis at the country level revealed that although the various agency initiatives used different “prisms” to view school health—public health for WHO, quality education for UNESCO, and child rights for UNICEF—the core activities for all the programmatic approaches were essentially the same. FRESH Framework A major step forward in international coordination and cohesion was achieved when the FRESH framework was launched at the World Education Forum in Dakar in April 2000 (World Bank FRESH Toolkit 2000). Among the early partners in this effort were the Education Development Centre, Education International, the Partnership for Child Development, UNESCO, UNICEF, the World Food Programme (WFP), WHO, and the World Bank. This partnership recognizes that the goal of universal education cannot be achieved while the health needs of children and adolescents remain unmet and that a core group of cost-effective activities can and must be implemented across the board to meet those needs and to deliver on the promise of EFA. The expanded commentary on the Dakar Framework for Action reflects the recommendations of this partnership and describes three ways in which health relates to EFA: as an input and condition necessary for learning, as an outcome of effective quality education, and as a sector that must collaborate with education to achieve the goal of EFA. In the follow-up to the Dakar Forum, UNESCO designated FRESH as an interagency flagship program that will receive international support as a strategy to achieve EFA. The FRESH framework, which is based on good practice recognized by all the partners, provides a consensus approach for the effective implementation of health and nutrition services within school health programs. The framework proposes four core components that should be considered in designing an effective school health and nutrition program and suggests that the program will be most equitable and cost-effective if all of these components are made available, together, in all schools: • Policy: health- and nutrition-related school policies that are nondiscriminatory, protective, inclusive, and gender sensitive and that promote the nutrition and physical and psychosocial health of staff, teachers, and children • School environment: access to safe water and provision of separate sanitation facilities for girls, boys, and teachers • Education: skills-based education, including life skills, that addresses health, nutrition, HIV/AIDS prevention, and hygiene issues and that promotes positive behaviors • Services: simple, safe, and familiar health and nutrition services that can be delivered cost-effectively in schools (such as deworming services, micronutrient supplements, and nutritious snacks that counter hunger) and increased access to youth-friendly clinics. The FRESH framework further proposes that these four core components can be implemented effectively only if they are supported by strategic partnerships between the following groups: • health and education sectors, especially teachers and health workers • schools and the community • children and others responsible for implementation. Adopting this framework does not imply that these core components and strategies are the only important elements; rather, implementing all of these in all schools would provide a sound initial basis for any pro-poor school health program. The common focus has encouraged concerted action by the participating agencies. It has also provided a common platform on which countries, agencies, donors, and civil society can support all programs, including agency-specific programs (table 58.1). Another important consequence of the FRESH consensus framework has been to offer a common point of entry for new efforts to improve health in schools, as illustrated by the three examples in box 58.1. This consensus approach has increased significantly the number of countries implementing school health reforms. The simplicity of the approach, combined with the enhanced resources available from donor coordination, has helped ensure that these programs can go to scale. Annual external support from the World Bank for these actions approaches US$90 million, targeting some 100 million schoolchildren. Common Interventions Table 58.2 lists some specific interventions commonly combined within the school health intervention package, but it should be recognized that not all of these interventions will be needed or be appropriate for all locations. Some interventions are synergistic: for example, worm infection will be addressed by the provision of latrines, the promotion of hand washing, relevant health and hygiene education, and deworming services. Similarly, HIV/AIDS infection among youths will be addressed by ensuring girls’ participation in school, offering School-Based Health and Nutrition Programs | 1097 ©2006 The International Bank for Reconstruction and Development / The World Bank 177 Box 58.1 Three Efforts to Improve Health in Schools The Multiagency Effort to Accelerate the Education Sector Response to HIV/AIDS in Africa This effort, coordinated by a Working Group of the UNAIDS Inter-Agency Task Team on HIV/AIDS and Education, promotes the FRESH framework specifically and helps education systems do the following: • adopt policies that avoid HIV/AIDS discrimination and stigmatization • provide a safe and secure school environment • provide skills-based health education, including life skills, in schools to promote positive behaviors and healthy lifestyles • improve access to youth-friendly health services. More than 36 countries and a similar number of agencies, bilateral donors, and nongovernmental organizations have collaborated in this effort since November 2002. The Global School Feeding Campaign of the WFP This campaign has gone beyond providing food aid to develop a programmatic link between nutrition and education. Working with partners, including national governments, parent-teacher and other community organizations, UNICEF, WHO, the World Bank, UNESCO, and the Food and Agriculture Organization, the campaign promotes the following: • policies that make food aid conditional on girls’ participation in education • an essential package that includes school sanitation and water and environmental improvement • nutrition education that improves the quality of students’ diets and HIV prevention education • nutrition services that include food, deworming, and alleviation of short-term hunger. Some 70 countries have begun to implement these principles and activities since 2002. The Partnership for Parasite Control Led by WHO and involving a broad range of development partners, this initiative promotes public and private efforts to include deworming in school health services, following a resolution of the 54th World Health Assembly to provide by 2010 regular deworming treatment to 75 percent of schoolage children at risk (an estimated target population of 398 million). Of 41 target countries in Africa, 19 have begun school-based deworming programs since 2001. Source: Authors. skills-based health education (including life skills), offering peer education, providing access to health clubs, and providing access to treatment for sexually transmitted infections (STIs) at clinics. It is also apparent that whereas some interventions promote multiple outcomes—for example, skills-based health education and life-skills development can help promote positive behaviors that prevent STIs and substance abuse—other interventions may have a single focus, such as iron supplementation to avoid anemia. Nevertheless, it is apparent that out-of-school children cannot benefit from many of the important components of school-based programs, such as skills-based health education and life-skills development programs to prevent HIV/AIDS. Reaching these children requires more flexible approaches that combine the best of nonformal, informal, and communitybased approaches (see chapter 59). Out-of-School Children A key issue in addressing the costs of the new approach to school health and nutrition programs is the significant savings offered by using the school system infrastructure rather than that of the health system as the key delivery mechanism. The school system provides not only a preexisting mechanism, so costs are at the margins, but also a system that aims at being pervasive and socially progressive. Some important interventions, especially in terms of health education, may be virtually cost free; they require only policy changes that result in doing things differently. More than 100 million school-age children are out of school; 60 percent are girls (UNESCO 1993). School health programs in Guinea and Madagascar have demonstrated that many of these children will take advantage of simple services, such as deworming, provided in schools (Del Rosso and Marek 1996); the school acts essentially as a community center. It also has been demonstrated that deworming programs in schools benefit out-of-school children by reducing disease transmission in the community as a whole (Bundy and others 1990). COST-EFFECTIVENESS OF INTERVENTION 1098 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 178 Table 58.2 Common Interventions within a School Health Program FRESH category Intervention Expected outcome Policy 1. Child rights, avoidance of discrimination and stigmatization, gender sensitive, child centered 1. Inclusion of all children 2. Inclusion of pregnant girls and mothers in education 2. Specific inclusion of girls 3. Enforcement of code of practice for teacher behavior zero tolerance policy 3. Avoidance of harassment and abuse 4. Collaboration between health and education sectors 4. Effective implementation 1. Access to safe water 1,2,3,5. Reduced infection 2. Hand washing 4. Reduced drop out of adolescent girls Environment 3. Provision of sanitation 4. Gender-separate sanitation 5. Garbage disposal Education 1. Curriculum addressing health, hygiene, and nutrition 2. Life-skills program 2. Lifelong positive behaviors such as avoidance of HIV/AIDS and substance abuse 3. Peer education program 4. Health-promoting clubs Services 1. Improved knowledge and skills to promote good health, hygiene, and nutrition 3, 4. Reinforcement of positive behaviors 1. Deworming for intestinal worms and schistosomiasis 1. Reduction in worm infection 2. Prompt recognition and treatment of malaria 2. Reduction in impact of malaria 3. Insecticide-treated nets 3. Reduction in incidence of malaria 4. Micronutrient supplements 4. Reduction in anemia and malnutrition 5. Breakfast, snacks, and meals 5. Avoidance of hunger 6. First-aid kits 6. Management of injuries 7. Referral to youth-friendly clinics 7. Access to specific treatment 8. Counseling and psychosocial support 8. Mental health Source: Authors. Table 58.3 Annual per Capita Costs of School-Based Health and Nutrition Interventions Delivered in Schools Condition Intervention Cost (US$) Intestinal worms Albendazole or mebendazole 0.03–0.20 Schistosomiasis Praziquantel 0.20–0.71 Vitamin A deficiency Vitamin A supplementation Iodine deficiency Iodine supplementation Iron deficiency and anemia Iron folate supplementation Refractive errors of vision Spectacles Clinically diagnosed conditions Physical examination Undernutrition, hunger School feeding 0.04 0.30–0.40 0.10 2.50–3.50 11.50 21.60–151.20, 21.26–84.50a Sources: Del Rosso and Marek 1996; Partnership for Child Development 1999; WHO 2000. a. For South America and Africa, costs are standardized for 1,000 kilocalories for 180 days. Annual costs of providing some common school-based interventions to students are given in table 58.3. This table illustrates two important points. First, some of the most widely needed interventions can be provided at remarkably low cost. Second, significant diversity exists in the cost of interventions, which is affected by factors such as local capacity, location and remoteness of communities, and community values and opinions; hence, these factors must be borne in mind when identifying a school health package. (See chapter 41 for details of the costs of sanitation provision.) Not illustrated in the table is the cost advantage of using the existing school infrastructure for delivery. Estimates for delivery of simple interventions (such as anthelmintic pills or micronutrient supplements) suggest that the teacher-delivery approaches listed here may be one-tenth of the cost of the more traditional mobile health teams and yet equally effective (Guyatt 2003). As with all education innovations, however, the additional cost of teacher orientation and training (inservice as well as preservice) needs to be factored into the costs of using the education system for delivery of health services. ECONOMIC BENEFITS OF INTERVENTION The most obvious benefit of school health interventions is arguably through the economic returns of improved adult health outcomes. Studies have increasingly documented a causal effect of adult health (broadly defined) on labor force participation, School-Based Health and Nutrition Programs | 1099 ©2006 The International Bank for Reconstruction and Development / The World Bank 179 wages, and productivity in developing countries; Strauss and Thomas (1995) present an overview of economic studies in this area. For example, height has been shown to affect wage-earning capacity as well as participation in the labor force for both women and men (Haddad and Bouis 1991). The effect of health on productivity and earnings may be strongest where low-cost health interventions produce large effects on health, such as lowincome settings where physical endurance yields high returns in the labor market. For a 1 percent increase in height, Thomas and Strauss (1997) find a 7 percent increase in wages in Brazil compared with a 1 percent increase in the United States. However, the apparent benefits of school health and nutrition programs will be underestimated when measured using only mortality or health-related disability metrics because these measures do not capture the impact of ill health on cognitive development or educational outcomes. Evidence over the past decade suggests these impacts have effect sizes in the range 0.25 to 0.4 SD and have implications for the child’s education and for life beyond school, including future earning potential. We investigate those implications by considering the economic benefits in terms of IQ and school attendance and by comparing school health programs with traditional education interventions. Economic Benefits of Long-Term Improvements in IQ School health interventions can yield considerable economic benefits through returns to wages and productivity if they translate into improved cognitive functioning and IQ in adulthood. For the United States, Zax and Rees (2002) estimate conservatively that an increase in IQ of 1 SD is associated with an increase in wages of more than 11 percent, falling to 6 percent when controlling for other covariates. Similar estimates for the relationship between IQ and earnings have been made for Indonesia (Behrman and Deolalikar 1995) and Pakistan (Alderman and others 1997) and in a review of developing countries (Glewwe 2002). In South Africa, an increase of 1 SD in literacy and numeracy scores was associated with a 35 percent increase in wages (Moll 1998). Extrapolating these results, a 0.25 SD increase in IQ, which is a conservative estimate of the benefit resulting from a school health intervention, would lead to an increase in wages of from 5 to 10 percent. Economic Benefit of Improved School Attendance School health interventions can raise adult productivity not only through higher levels of cognitive ability, but also through their effect on school participation and years of schooling attained. Healthier children are more likely to attend, and modest improvements in examination scores can be associated with continuation in schooling. Malaria chemoprophylaxis given in early childhood in The Gambia led to an increase of more than one year in primary schooling. In preschool children in Delhi, iron supplementation was associated with an increase of 5.8 percent in rates of participation at the preschool level (Bobonis, Miguel, and Sharma 2004). In western Kenya, deworming treatment improved primary school participation by 9.3 percent, with an estimated 0.14 additional years of education per pupil treated (Miguel and Kremer 2004). On the basis of crude estimates of returns to schooling, an increase of 9.3 percent in participation rates results in a return of US$44. Miguel and Kremer (2004) conclude that these benefits still outweigh the costs even if increased school participation leads to greater costs in teacher compensation through the need for additional teachers. They note that the benefit-cost ratio remains over 10 even if the rate of return to an additional year of schooling is as low as 1.5 percent. These results suggest that for realistic estimates of returns to schooling, the net present discounted value of lifetime earnings is likely to be high compared to the costs of treatment even for small gains in school participation.1 In the absence of studies estimating the direct link between school health interventions and school participation, the relationship can be estimated indirectly by considering the effect of interventions on test scores and the implications that improved test scores have for school participation. Improvements in cognitive function can be converted into an equivalent number of years of schooling. For example, Jukes and others (2002) found that heavy schistosomiasis was (nonsignificantly) associated with a decrease in arithmetic scores of 1.35 marks (0.25 SD). An extra year of schooling was associated with an increase in arithmetic scores of 2.24 marks (0.42 SD). Thus, the negative effect of heavy schistosomiasis was equivalent to missing just over half a year of schooling. The cognitive gains from an extra year of schooling can also be estimated retrospectively: in a study of adults in South Africa, each additional year of primary schooling was associated with a 0.1 SD increase in cognitive test scores (Moll 1998). According to these estimates, a typical increase of 0.25 SD associated with school health and nutrition programs is equivalent to an additional 2.5 years of schooling. Liddell and Rae (2001) assessed the direct effect of test scores on grade progression in Africa. Each additional SD scored in first-grade exams resulted in children being 4.8 times as likely to reach seventh grade without repeating a year of schooling.2 According to these estimates, an increase of 0.25 SD in examination scores, which is typically achieved by school health and nutrition programs, will make children 1.48 times3 as likely to complete seventh grade, which implies that the extra cumulative years of schooling attributable to the school health intervention average 1.19 years per pupil. The previous estimates for added years of schooling owing to school health interventions range from seven months to two years. Increased years of schooling are associated with, among other outcomes, higher worker productivity and generally higher productivity in nonmarket production activities, including greater farmer 1100 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 180 efficiency and productivity (Jamison and Lau 1982; Psacharopoulos and Woodhall 1985; Strauss and Thomas 1995). Psacharopoulos and Patrinos (2002) summarize a wide range of studies that focus on individual wage earnings. For Sub-Saharan Africa, they find a 12 percent rate of return to one additional year in school, compared with 10 percent for Asian countries. These returns are very high, even allowing for a portion of the return to years of schooling to be capturing ability and factors other than schooling itself (Card 2001). Education brings benefits beyond improved earnings. One year of extra education for girls can lead to a reduction of from 5 to 10 percent in infant mortality (Schultz 1993). Five extra years of education for women in Africa could reduce infant mortality by up to 40 percent (Summers 1994). tional cash-transfer approach is, in both cases, apparently at the lower end of effectiveness and the higher end of cost. Economic Benefits of Programs Policy and Economic Issues in Defining Sectoral Roles in Intervention The educational gains from school health and nutrition programs should be considered in the context of alternative educational inputs, such as improving teacher salaries and qualifications, reducing class size, improving school facility infrastructure, and providing instructional materials. Many studies relate student outcomes to school characteristics, but few of these studies provide information on the relative or actual costs of the educational inputs. The costs, however, are substantially greater than for the school health interventions considered here. Despite the higher costs, the evidence from the few randomized evaluations that have been conducted suggests that the scale of effect of additional education inputs is typically low (see discussion in Miguel and Kremer 2004). A review of studies showed that instructional materials (such as additional textbooks) had the highest productivity, raising student test scores significantly more than other inputs for each dollar spent. However, even these interventions have only a weak effect. In a randomized experiment in Kenya, for example, providing textbooks had no effect on the bottom three quintiles of students and raised test scores by only 0.2 SD for the upper two quintiles. Relating these results to the findings in the previous section and to the annual per pupil costs, school health interventions appear very cost-effective compared to the highestproductivity, more traditional education inputs. Recently, conditional cash-transfer programs have been viewed as potentially very cost-effective methods to increase school enrollment. These programs are generally large in scope, representing a commitment of between 0.1 and 0.2 percent of gross national income. The Progresa program in Mexico is estimated to have increased enrollment by 3.4 percent and to have increased schooling by 0.66 years, with an average cash transfer (for grades 3 to 8) of about US$136 per child per school year (assumed to be 180 days). Gains from a similar program in Nicaragua were estimated at 0.45 years of school at a cost of US$77 per year. If we compare these results with those presented for school health and nutrition programs, the condi- IMPLEMENTATION OF PROGRAMS AND LESSONS FROM EXPERIENCE The FRESH framework provides strategic guidance, but the practical design of actual programs reflects differences in local needs and capacity. Successful and equitable programs in lowand middle-income countries are characterized by a focus on school-based delivery, on a public health paradigm that minimizes the need for clinical intervention and reliance on health service facilities, and on participation of the public sector and civil society locally. A negative correlation between income level and both ill health and malnutrition is clearly demonstrated both in cross-country comparisons and within countries (see de Silva and others 2003), partly because poverty promotes both disease and an inadequate diet. Similarly, children who are not enrolled in school come from households with lower income levels (Filmer and Pritchett 2001). This fact suggests that school health services that are pro-poor and specifically linked to efforts to achieve universal participation in education will have a greater return. Early school health programs, particularly in colonial Africa, were intended to serve the minority of children who had access to school in urban centers or elite boarding facilities. They relied on specific infrastructures and services—such as mobile health teams, school visits, school nurses, and in-school clinics—that were additional to the normal range of health service provision. This approach has proven difficult to make universally available, even in middle-income countries. A school nurse program in KwaZulu-Natal, for example, achieved inadequate coverage (18 percent of the target population) and little referral or follow-up treatment of cases of ill health detected, despite a relatively high investment of US$11.50 per student targeted per year (World Bank FRESH Toolkit 2000). As shown in the following examples, using the FRESH framework approach reduces costs significantly and enhances both coverage and outcomes. An important element of the new approach to school health is a focus on minimizing the need for clinical diagnosis. Mass delivery of services, such as deworming and micronutrient supplementation, is preferable on efficacy, economic, and equity grounds to approaches that require diagnostic screening (Warren and others 1993). Sectoral Roles in Implementation Table 58.4 gives examples from low- and middle-income countries of how the four core components of FRESH are being supported by different approaches. In about 85 percent School-Based Health and Nutrition Programs | 1101 ©2006 The International Bank for Reconstruction and Development / The World Bank 181 ©2006 The International Bank for Reconstruction and Development / The World Bank 182 1102 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others Table 58.4 Nine Low- and Middle-Income Countries and How They Use FRESH Program approach Country examples Public sector: public sector– supported and –implemented Outcomes (Costs per child per year) Policy Environment Health education Health services Guinea, Ghana, and Tanzania In all three countries, the Ministry of Education (or in Ghana, its executive body, the Ghana Education Service) implements the program under the guidance of the Ministry of Health, on the basis of a formal policy agreement. In Tanzania, the Ministries of Community Development and of Local Government are also parties to the agreement. The existing inservice teacher training and supply-line infrastructures are used to prepare teachers and supply the necessary materials. Separate sanitation facilities for girls and boys in all new schools; access to potable water in all schools. Health, hygiene, and nutrition education as part of the formal curriculum. Deworming (for both schistosomiasis and intestinal worms) provided by teachers twice a year; in Guinea, this service is followed by iron folate supplementation. In three years, in Guinea— 1.1 million students, in Ghana—577 schools and 83,000 students (US$0.54), in Tanzania—353 schools and 113,000 students (US$0.89). Parastatal support for public sector intervention Madagascar The Community Nutrition Programme provides training and support to the Ministry of Education on the basis of a formally agreed-on health policy for the education sector. In all schools in the 43 poorest districts (44 percent of all districts), the program prepares teachers and provides materials. In addition, the program also provides Parent-Teacher Associations (PTAs) with access to a social fund to support construction of facilities. Each PTA can request up to US$500, with a 20 percent community contribution based on an annual parental contribution of US$0.16. Access to potable water and hand-washing facilities, in all schools; where requested by PTAs, construction of latrines, wells, fences, and sports facilities. A formal health education curriculum, supported by community information, education, and communication (IEC). Twice-yearly deworming and iron folate (for three months) delivered by teachers; test kits to confirm iodization of local sources of salt; where requested by PTAs, provision of food preparation facilities. In three years, 14,000 teachers trained in 4,585 schools, 430,000 students (US$0.78 to US$1.08 per capita per year). Social fund: public sector support for community intervention Tajikistan The Ministry of Labor and Social Protection, with the Ministries of Education and of Health, have developed a memorandum of understanding that sets out health policies for the education sector. The program channels resources through PTAs, which identify and assist needy children. A training program, delivered by NGOs, prepares PTA members to develop proposals of up to US$5,000 for their school to support activities selected from a menu of items. Provision of sanitation facilities, potable water, and sports facilities. Training of teachers in health promotion. Training of teachers to provide first aid, micronutrients, and deworming; provision of food preparation facilities. The program targets the 100,000 neediest children in all 200 schools in the six poorest districts of Tajikistan (US$1 per capita per year). ©2006 The International Bank for Reconstruction and Development / The World Bank 183 Private sector: community payment for NGO-implemented intervention Indonesia The NGO Yayasan Kusuma Buana has a formal agreement with the education department in Jakarta and three other major cities to train teachers, perform diagnostic tests, and provide medicines and materials. The NGO offers Papanicolaou smear tests and referral services to teachers. Unit costs are low because parasite diagnosis involves mass screening in a central laboratory (approximately 2,500 diagnoses per day) and medicines are obtained at preferential rates from two commercial partners. Not included in program. Nutrition and hygiene education as part of the curriculum. Stool examination by the laboratory and deworming by teachers as necessary twice a year; iron folate provided by teachers twice a year (for three months). The program has been in existence for 17 years and currently reaches 627 schools and 161,000 students, at a cost to parents of US$0.10 annually. NGO implementation with financial support from public sector Burkina Faso, Malawi, and the Philippines The international NGO Save the Children U.S.A. implements school health and nutrition activities in nonformal schools created with support from government, local communities, and private donations. Separate sanitation facilities for girls and boys and access to potable water. Health, hygiene, and nutrition education as part of the curriculum supported by extracurricular IEC activities Deworming and micronutrient supplementation (vitamin A and iron) provided by teachers annually. In three years, in Burkina Faso, 42,000 students plus nonenrolled children in 171 schools (US$2). In four years, in Malawi, 122,000 children in 181 schools (US$3). In four years in the Philippines, 23,000 children in 53 schools (US$6). Source: Authors. School-Based Health and Nutrition Programs | 1103 of programs reviewed, school health and nutrition programs are delivered and funded by the public education sector, with a formal role for the health sector in design and supervision. Although this public sector “mainstream” model has proven the most popular approach, it is not the only successful one. In some cases, the public sector has identified appropriate options and developed operational manuals but then has used a social fund to provide direct support to communities and has used schools to select and implement the most relevant actions locally, often with the assistance of nongovernmental organizations (NGOs). In other cases, services have been contracted out by the public sector, and in some middle-income countries, the move toward a demand-led approach has resulted in a private sector service. The private sector approach has proven sustainable over nearly two decades in urban Indonesia but may require a technical infrastructure and local market base that are inappropriate for predominantly rural low-income countries. The approach is modeled on a program initiated in Japan in 1948, which relied on private sector technicians, working independently at first but later formalized within the Japan Association of Parasite Control, who conducted stool examinations and then treated infected individuals for a per capita fee equivalent to approximately US$0.74 in 2004. At its peak, the private sector program conducted some 12 million examinations annually, implying a turnover of nearly US$9 million at today’s prices. The prevalence of roundworm infection fell from a high of 73 percent in 1949 to less than 0.01 percent by 1985. Although a private sector response is effective in some circumstances, overall the characteristics of school health and nutrition programs make a compelling case for public sector intervention. First, treatment externalities may create external benefits to others in addition to the benefit for the treated individual. This situation is clearly the case for communicable disease interventions, especially against worm infection. Second, some forms of intervention (such as vector control, health education campaigns, epidemiological surveillance, and interventions that have strong externalities) are almost pure public goods; that is, no one can be excluded from using the goods or service they deliver,and thus the private sector is unlikely to compete to deliver these goods. Finally, there is typically little private demand for general preventive measures,such as information on the value of washing hands. None of these factors is an argument against a private sector role in service delivery, but they do suggest that private sector demand is likely to be greater in middle-income populations and where public sector actions have created a demand. Roles of Key Stakeholders in Implementation There are many ways to approach the delivery of school health, but these diverse experiences suggest common features—in particular, the consistency in the roles played by government and nongovernmental agencies as well as other partners and stakeholders (table 58.5). In nearly every case, the Ministry of Education is the lead implementing agency, reflecting both the goal of school health programs in improving educational achievement and the fact that the education system provides the most complete existing infrastructure for reaching schoolage children. However, the education sector must share this responsibility with the Ministry of Health, particularly because the latter has the ultimate responsibility for health of children. It is also apparent that the program’s success depends on the effective participation of numerous other stakeholders, including civil society, and especially the beneficiaries and their parents or guardians. The children and their families are the clients of these programs, and their support for program implementation is critical to the program’s success. Key Issues in Designing Effective Programs The diverse experiences of school health programming suggest some key elements that are common contributors to success in many programs. • Focus on education outcomes. Making explicit links among school health programs and learning and education sector priorities (especially EFA and gender equity) helps ensure the commitment of the sector to program support and implementation. • Develop a formal, multisectoral policy. Education sector actions in health require the explicit agreement of the health sector. This potential tension can be resolved by defining sectoral responsibilities at the outset; failure to enter into dialogue has led, in Africa and Central Asia, to some health sectors resisting teacher delivery of deworming drugs, despite WHO recommendations. • Initiate a process of wide dissemination and consultation. Because there are multiple stakeholders, implementers, enablers, and gatekeepers, a process of consultation is necessary to establish ownership and to identify obstacles before they constrain progress. The process should involve at least community-based organizations, NGOs, faith-based organizations, pupils, and teacher associations. In one country in East Africa, lack of prior agreement on the content of sexuality education delayed implementation for more than three years. • Use the existing infrastructure as much as possible. Building on existing curriculum opportunities and the network of formal and nonformal teachers will accelerate implementation and reduce costs. Programs that rely on the development of new delivery systems—mobile school health teams, a cadre of school nurses—take longer to establish and are expensive and complicated to sustain and take to scale. 1104 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 184 Table 58.5 Roles of Agencies, Partners, and Stakeholders in School Health and Nutrition Programs Partner Roles Comments Ministry of Education Lead implementing agency Health and nutrition of schoolchildren is a priority for EFA. Lead financial resource Education policy defines school environment, curriculum, duties of teachers. Education sector policy Education system has a pervasive infrastructure for reaching teachers and school-age children. Lead technical agency Health of school-age children has lower priority than clinical services, infant health. Ministry of Health Health sector policy Health policy defines role of teachers in service delivery, procurement of health materials. Other public sector agencies (for example, Welfare, Social Affairs, local government) Support for education and health systems Private sector (for example, health service, pharmaceuticals, publications) Specialist service delivery Sector has major role in drug procurement and training materials production. Material provision Sector has specialist roles in health diagnostics. Civil society (for example, NGOs, faith-based organizations, PTAs) Training and supervision At the local level, organizations serve as gatekeepers and fund holders and may target implementation. Teacher associations Definition of teachers’ roles School health programs demand an expanded role for teachers. Community (children, teachers, parents) Partners in implementation Communities are gatekeepers for the content of health education (especially moral and sexual content) and for the role of nonhealth agents (especially teachers) in health service delivery; pupils are active participants in all aspects of the process at the school level. Fund holder Ministries of local government are often fund holders for teachers and schools and for clinics and health agents. Ministries of Welfare and Social Affairs provide mechanisms for providing social funds. Local resource provision Organizations provide additional resource streams, particularly international NGOs. Definition of acceptability of curriculum and teachers’ roles Supplementation of resources • Use simple, safe, and familiar health and nutrition interventions. Success in rapidly reaching all schools depends on stakeholder acceptance, which is more likely if the interventions are already sanctioned by local and international agencies and are already in common use by the community. • Provide primary support from public resources. Compelling arguments exist for public investment in school health programs: the contribution to economic growth, the high rate of return, the large externalities, and the fact that the majority of interventions are public goods. • Be inclusive and innovative in identifying implementation partners. Although public resources are crucial for school health programs, contributions from outside the public sector can be vital. NGOs have proven effective in supporting public sector programs through training and supervision, particularly at local levels.Although market failure appears to have largely precluded the private sector from effectively implementing national programs in low-income countries, examples of successful contributions do occur, particularly in dense urban populations and in middle-income countries. Communities supplement program finance at the margins. RESEARCH AND DEVELOPMENT AGENDA Reliable evidence suggests that ill health and malnutrition affect education access, participation, completion, and achievement, and that school-based health and nutrition programs can provide a cost-effective and low-cost solution. This evidence does not imply, however, that no uncertainties exist. Cost-Effectiveness of School-Based HIV/AIDS Prevention Substantial evidence suggests that skills-based health education, including life-skills development programs, can promote positive behaviors and reduce the risks of exposure to HIV infection, and that girls’ education programs have similar effects (Kirby 2002). Evidence also exists for a positive effect of completing education on HIV prevalence (de Walque 2004; World Bank 2002). What is lacking is direct evidence about the contribution that school-based prevention programs can make in reducing the incidence of HIV infection, as well as evidence for the relative cost-effectiveness of such programs compared with existing efforts to promote education completion and girls’ education. School-Based Health and Nutrition Programs | 1105 ©2006 The International Bank for Reconstruction and Development / The World Bank 185 Cost-Effectiveness of Malaria Programs Malaria occurs commonly in schoolchildren, particularly in areas of unstable transmission in Africa and Asia. It is a leading source of mortality in this age group and adversely affects education by reducing school attendance, cognition, learning, and school performance. Current school-based approaches focus on knowledge of the disease and the use of impregnated bednets but do not address the need for treatment of affected children. Yet presumptive treatment by teachers has been shown to significantly reduce mortality (Pasha and others 2003), and intermittent preventive treatment also shows considerable promise (Brooker and others 2000). There is a need to confirm the success of school-based treatment in different epidemiological settings and to address questions about the cost and sustainability of this approach. Cost-Effectiveness of Targeting Food Aid The high prevalence of malnutrition in children continues to be a major challenge for low-income countries. Providing food to children at school is often seen as an important part of the solution and is a major focus for food aid. However, the nutrition literature suggests that ensuring good nutrition earlier in life—certainly before 3 years of age, but perhaps earlier—is essential to ensuring an appropriate development trajectory throughout life (see chapter 27). Where food is limiting, it raises the question whether the first target should be preschool rather than school-age children. This debate has been blurred by admixing the nutrition outcomes with broader social and education issues. Clearly, providing a meal at school is socially desirable and can offer education benefits for children who otherwise would have to walk often long distances home to eat or remain hungry. It is also clear that schools represent an extensive and established network for providing nutrition interventions to very large numbers of children at a low cost per child. No comparable network exists to reach preschool children. However, from a nutritional perspective, it remains unclear whether ensuring good nutrition early in life has more effect on subsequent development—including educational achievement—than providing food at school age. In consequence, the clearest benefit of school health and nutrition programs is measurable in terms of education outcomes and their economic returns. The scale of benefit is significant: school health and nutrition interventions can add four to six points to IQ levels, 10 percent to participation in schooling, and one to two years of education. This scale of benefit can add 8 to 12 percent to labor returns and provide a rate of return that offers a strong argument for public sector investment. Compelling evidence suggests that education qua education can help protect individuals from HIV infection. Achieving EFA goals and combining this outcome with school health programs that help establish lifelong positive behaviors are now recognized as essential to the multisectoral prevention response to HIV/AIDS. The scale of the education benefit and the role of education in the fight against HIV/AIDS mean that school health and nutrition programs are today seen as a priority for both the education and the health sectors. This focus, in turn, has resulted in a shift toward public health rather than clinical intervention and toward school-based delivery rather than health system approaches. These policy changes enhance cost-effectiveness and social progressiveness, because delivery through the school system is an order of magnitude less costly than using health systems and in low-income countries is better targeted to the poor. These changes in emphases have coincided with significant technical and political policy reform. Technical consensus around the FRESH framework has encouraged countries and agencies to develop programs around a common coordinating principle, while the political imperative has been strengthened by the recognition that school health and nutrition programs are essential to achieving EFA and the Millennium Development Goals and are at the center of the preventative response to the HIV/AIDS pandemic. Although much of this change has evolved over the past two decades, significant acceleration has occurred since the World Education Forum in 2000. Today, a majority of low-income countries have recognized the need for school health and nutrition programs and are seeking to implement them. NOTES CONCLUSIONS The rationale for school-based health and nutrition programs and the approach to their implementation have undergone a paradigm shift over the past two decades. The traditional perception of these programs as seeking to improve the health of schoolchildren cannot be justified on the basis of mortality or public health statistics alone. Instead, it is increasingly recognized that a major—perhaps the major— impact of ill health and malnutrition on this age group is that on cognitive development, learning, and educational achievement. 1. These calculations assume the following: a return to an additional year of school is 7 percent; wage gains are earned over 40 years in the workforce, discounted at 5 percent per year with no wage growth; annual wage earnings are US$400 per year, which is below the estimated agricultural and nonagricultural annual wages for low-income countries (World Bank 2003). The opportunity costs of the additional schooling (child labor) have not been considered but are likely to be negligible. 2. These calculations assume that a pupil’s falling behind the equivalent of one year in test scores has the same effect on earnings as losing one year of schooling; that the advantage that third graders have over second graders, for example, is the same as the advantage someone who has studied for a total of three years has over someone who has studied for two years; and that the impact of first-grade examination scores on 1106 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 186 the probability of transition from one class to the next is the same at each grade level. 3. If an increase of 1 SD in exam scores leads to children being 4.8 times as likely to reach seventh grade, the increased likelihood of reaching seventh grade because of a 0.25 SD increase can be calculated as EXP (0.25 ⫻ LN(4.8)). Filmer, D., and L. H. Pritchett. 2001. “Estimating Wealth Effects without Expenditure Data—or Tears: An Application to Educational Enrolments in States of India.” Demography 38: 115–32. Glewwe, P. 2002. “Schools and Skills in Developing Countries: Education Policies and Socioeconomic Outcomes.” Journal of Economic Literature 40: 436–82. Guyatt, H. L. 2003. “The Cost of Delivering and Sustaining a Control Program for Schistosomiasis and Soil-Transmitted Helminthiasis.” Acta Tropica 86: 267–74. REFERENCES Alderman, H., J. R. Behrman, S. Khan, D. R. Ross, and R. Sabot. 1997. “The Income Gap in Cognitive Skills in Rural Pakistan.” Economic Development and Cultural Change 46: 97–122. Behrman, J. R., and A. B. Deolalikar. 1995. “Are There Differential Returns to Schooling by Gender? The Case of Indonesian Labor-Markets.” Oxford Bulletin of Economics and Statistics 57: 97–117. Berkley, S., and D. Jamison. 1990. A Conference on the Health of School Age Children. New York: United Nations Development Programme and Rockefeller Foundation. Processed. Bobadilla, J. L., P. Cowley, P. Musgrove, and H. Saxenian. 1994. “Design, Content, and Financing of an Essential National Package of Health Services.” Bulletin of the World Health Organization 72: 653–62. Bobonis, G., E. Miguel, and C. Sharma. 2004. “Iron Deficiency Anemia and School Participation.” Working paper, University of California at Berkeley. Brooker, S., H. Guyatt, J. Omumbo, R. Shretta, L. Drake, and J. Ouma. 2000. “Situation Analysis of Malaria in School-Aged Children in Kenya: What Can Be Done?” Parasitology Today 16: 183–86. Bundy, D. A. P. 1989. “New UNESCO International Project.” Parasitology Today 5: 168. ———. 1997. “This Wormy World: Then and Now.” Parasitology Today 13: 407–8. Bundy, D. A. P., and H. L. Guyatt. 1996. “Schools for Health: Focus on Health, Education, and the School-Aged Child.” Parasitology Today 12: 1–16. Bundy, D. A. P., A. Hall, G. F. Medley, and L. Savioli. 1992. “Evaluating Measures to Control Intestinal Parasitic Infections.” World Health Statistics Quarterly 45: 168–79. Bundy, D. A. P., S. Lwin, J. S. Osika, J. McLaughline, and C. O. Pannenborg. 2000. “What Should Schools Do about Malaria?” Parasitology Today 16: 181–82. Bundy, D. A. P., M. S. Wong, L. L. Lewis, and J. Horton. 1990. “Control of Geohelminths by Delivery of Targeted Chemotherapy through Schools.” Transactions of the Royal Society of Tropical Medicine and Hygiene 84: 115–20. Card, D. 2001. “Estimating the Return to Schooling: Progress on Some Persistent Econometric Problems.” Econometrica 69: 1127–60. Cohen, S., and A. Smith. 1996. “Psychology of Common Colds and Other Infections.” In Viral and Other Infections of the Human Respiratory Tract, ed. S. Myint and D. Taylor-Robinson. London: Chapman and Hall. Grigorenko, E., R. Sternberg, D. Ngorosho, C. Nokes, M. Jukes, K. Alcock, and D. Bundy. Forthcoming. “Effects of Antiparasitic Treatment on Dynamically-Assessed Cognitive Skills.” Journal of Applied Developmental Psychology. Haddad, L., and H. Bouis. 1991. “The Impact of Nutritional Status on Agricultural Productivity: Wage Evidence from the Philippines.” Oxford Bulletin of Economics and Statistics 53: 45–68. Hall, A., L. J. Drake, and D. A. P. Bundy. 2001. “Public Health Measures to Control Helminth Infections.” In Nutritional Anemias, ed. U. Ramakrishnan. Boca Raton, FL: CRC Press. Halloran, M. E., D. A. P. Bundy, and E. Pollitt. 1989.“Infectious Disease and the UNESCO Basic Education Inititative.” Parasitology Today 5: 359–62. Huda, S. N., S. M. Grantham-McGregor, and A. Tomkins. 2001. “Cognitive and Motor Functions of Iodine-Deficient but Euthyroid Children in Bangladesh Do Not Benefit from Iodized Poppy Seed Oil (Lipiodol).” Journal of Nutrition 131: 72–77. Jamison, D., and L. J. Lau. 1982. Farmer Education and Farm Efficiency. Baltimore: Johns Hopkins University Press. Jukes, M. C. H., L. L. Drake, and D. A. P. Bundy. Forthcoming. Leveling the Playing Field: The Importance of School Health and Nutrition in Achieving Education for All. Washington, DC: World Bank. Jukes, M., C. A. Nokes, K. L. Alcock, J. Lambo, C. Kihamia, A. Mbise, and others. 2002. “Heavy Schistosomiasis Associated with Poor Short-Term Memory and Slower Reaction Times in Tanzanian School Children.” Tropical Medicine and International Health 7: 104–17. Jukes, M. C. H., M. Pinder, E. L. Grigorenko, H. Banos Smith, E. Bariau-Meier, G. Walraven, and others. Submitted. “The Impact of Malaria Chemoprophylaxis in Early Childhood on Cognitive Abilities and Educational Outcomes 14 Years Later: Follow-up of a Randomised Controlled Trial in The Gambia.” Kirby, D. 2002. “The Impact of Schools and School Programs upon Adolescent Sexual Behavior.” Journal of Sexual Research 39: 27–33. Liddell, C., and G. Rae. 2001. “Predicting Early Grade Retention: A Longitudinal Investigation of Primary School Progress in a Sample of Rural South African Children.” British Journal of Educational Psychology 71: 413–28. Miguel, E., and M. Kremer. 2004. “Worms: Identifying Impacts on Education and Health in the Presence of Treatment Externalities.” Econometrica 72: 159–217. Moll, P. G. 1998. “Primary Schooling, Cognitive Skills and Wages in South Africa. “ Economica 65: 263–84. Colbourne, M. J. 1955. “The Effect of Malaria Suppression in a Group of Accra School Children.” Transactions of the Royal Society of Tropical Medicine and Hygiene 4: 356–69. Partnership for Child Development. 1998a. “The Anthropometric Status of School Children in Five Countries in the Partnership for Child Development.” Proceedings of the Nutrition Society 57: 149–58. Del Rosso, J. M., and T. Marek. 1996. Class Action: Improving School Performance in the Developing World through Better Health and Nutrition School. Washington, DC: World Bank. ———. 1998b: “The Health and Nutritional Status of School Children in Africa: Evidence from School-Based Health Programmes in Ghana and Tanzania.” Transactions of the Royal Society of Tropical Medicine and Hygiene 92: 254–61. de Walque, D. 2004. “How Does the Impact of an HIV/AIDS Information Campaign Vary with Educational Attainment? Evidence from Rural Uganda.” Working paper, World Bank, Washington, DC. European Commission, WHO (World Health Organization) Europe, and Council of Europe. 1996. Facts about the European Network of Health Promoting Schools. ———. 1999. “The Cost of Large-Scale School Health Programmes Which Deliver Anthelmintics to Children in Ghana and Tanzania.” Acta Tropica 73: 183–204. ———. 2001. “Anaemia in Schoolchildren in Eight Countries in Africa and Asia.” Public Health Nutrition 4: 749–56. School-Based Health and Nutrition Programs | 1107 ©2006 The International Bank for Reconstruction and Development / The World Bank 187 Pasha, O., J. Del Rosso, M. Mukaka, and D. Marsh. 2003. “The Effect of Providing Fansidar (Sulfadoxine-Pyrimethamine) in Schools on Mortality in School-Age Children in Malawi.” Lancet 361: 577–78. UNESCO (United Nations Educational, Scientific, and Cultural Organization). 1990. “World Declaration on Education for All.” http://www. unesco.org/education/information/nfsunesco/pdf/JOMTIE_E.PDF. Pollitt, E. 1990. Malnutrition and Infection in the Classroom. Paris: United Nations Educational, Scientific, and Cultural Organization. ———. 1993. World Education Report. Paris: UNESCO. Pollitt, E., S. Cueto, and E. R. Jacoby. 1998. “Fasting and Cognition in Welland Undernourished Schoolchildren: A Review of Three Experimental Studies.” American Journal of Clinical Nutrition 67: 779S–84S. Psacharopoulos, G., and H. Patrinos. 2002. Returns to Investment in Education: A Further Update. Vol. 2881. Washington, DC: World Bank. Psacharopoulos, G., and M. Woodhall. 1985. Education for Development: An Analysis of Investment Choices. New York: Oxford University Press. Schultz, T. P. 1993. “Returns to Women’s Schooling.” In Women’s Education in Developing Countries: Barriers, Benefits, and Policy, eds. E. King and M. Anne Hill. Baltimore: Johns Hopkins University Press. Shiff, C., W. Checkley, P. Winch, Z. Premji, J. Minjas,and P. Lubega. 1996. “Changes in Weight Gain and Anaemia Attributable to Malaria in Tanzanian Children Living under Holoendemic Conditions.” Transactions of the Royal Society of Tropical Medicine and Hygiene 90: 262–65. Simeon, D. T. 1998. “School Feeding in Jamaica: A Review of Its Evaluation.” American Journal of Clinical Nutrition 67: 790S–94S. Simeon, D. T., and S. Grantham McGregor. 1989. “Effects of Missing Breakfast on the Cognitive Functions of School Children of Differing Nutritional Status.” American Journal of Clinical Nutrition 49: 646–53. Snow, R. W. 1999. “Estimating Mortality, Morbidity, and Disability Due to Malaria among Africa’s Non-Pregnant Population.” Bulletin of the World Health Organization 77: 624–40. Strauss, J., and D. Thomas. 1995. “Human Resources: Empirical Modeling of Household and Family Decisions.” In Handbook of Development Economics, ed. J. Behrman and T. N. Srinivasan, vol. IIIA. Amsterdam: Elsevier. Summers, L. H. 1994. Investing in All the People: Educating Women in Developing Countries. Vol. 45. Washington, DC: World Bank. Thomas, D., and J. Strauss. 1997.“Health and Wages: Evidence on Men and Women in Urban Brazil.” Journal of Economics 77: 159–85. ———. 2003. EFA Global Monitoring Report 2003/4: Gender and Education for All: The Leap to Equality. Paris: UNESCO. UNICEF (United Nations Children’s Fund). 1990. “Convention on the Rights of the Child.” http://www.unicef.org/crc/crc.htm. UNICEF, WHO (World Health Organization), World Bank, UNFPA (United Nations Population Fund), and UNESCO. 2003. Skills for Health: Skills-Based Health Education, Including Life Skills—An Important Component of a Child Friendly/Health Promoting School. U.S. Census Bureau. 2002. “Global Population Profile.” http://www. census.gov. Warren, K. S., D. A. P. Bundy, R. M. Anderson, A. R. Davis, D. A. Henderson, D. T. Jamison, and others. 1993. “Helminth Infection.” In Disease Control Priorities in Developing Countries, ed. D. T. Jamison, W. H. Mosley, A. R. Measham, and J. L. Bobadilla, 131–60. New York: Oxford University Press. WHO (World Health Organization). 1986. Ottawa Charter. Geneva: WHO. ———. 1996. Global School Health Initiative. WHO/HPR/98.4. Geneva: WHO. ———. 2000. Elimination of Avoidable Disability Due to Refractive Errors. Report of an informal planning meeting. WHO/PBC/00.79. WHO: Geneva ———. 2003. World Health Report. http://www.who.int/whr/2003. World Bank. 1993. World Development Report: Investing in Health. New York: Oxford University Press. ———. 2002. Education and HIV/AIDS: A Window of Hope. Washington, DC: World Bank. ———. 2003. World Development Indicators. Washington, DC: World Bank. ———, with the Partnership for Child Development. 2000. “FRESH: Focusing Resources on Education and School Health. A FRESH Start to Improving the Quality and Equity of Education.” http://www. schoolsandhealth.org/FRESH.htm. UNAIDS (Joint United Nations Programme on HIV/AIDS). 2000. World Bank 2000. The FRESH Framework: A Toolkit for Task Managers. Human Development Network, World Bank, Washington, DC. ———. 2002. Report on the Global HIV/AIDS Epidemic: The Barcelona Report. Geneva: UNAIDS. Zax, J. S., and D. I. Rees. 2002. “IQ, Academic Performance, Environment, and Earnings.” Review of Economics and Statistics 84: 600–16. 1108 | Disease Control Priorities in Developing Countries | Donald A. P. Bundy, Sheldon Shaeffer, Matthew Jukes, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 188 Chapter 61 Natural Disaster Mitigation and Relief Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio Sudden-onset natural and technological disasters impose a substantial health burden, either directly on the population or indirectly on the capacity of the health services to address primary health care needs. The relationship between communicable diseases and disasters merits special attention. This chapter does not address epidemics of emerging or reemerging diseases, chronic degradation of the environment, progressive climatic change, or health problems associated with famine and temporary settlements. In line with the definition of health adopted in the constitution of the World Health Organization (WHO), the chapter treats disasters as a health condition or risk, which, as any other “disease,” should be the subject of epidemiological analysis, systematic control, and prevention, rather than merely as an emergency medicine or humanitarian matter. The chapter stresses the interdependency between long-term sustainable development and catastrophic events, leading to the conclusion that neither can be addressed in isolation. DISASTERS AS A PUBLIC HEALTH CONDITION According to the International Federation of Red Cross and Red Crescent Societies, internationally reported disasters in 2002 affected 608 million people worldwide and killed 24,532—well below the preceding decade’s annual average mortality of 62,000 (IFRC 2003). Many more were affected by myriad local disasters that escaped international notice. Disaster has multiple and changing definitions. The essential common element of those definitions is that disasters are unusual public health events that overwhelm the coping capacity of the affected community. This concept precludes the universal adoption of a threshold number of casualties or vic- tims. What would be a minor incident in a large country may constitute a major disaster in a small isolated island state. Not only are “quantitative definitions of disasters unworkably simplistic” as noted by Alexander (1997, 289), but when based on the economic toll or the number of deaths, they are also misleading with regard to the immediate health needs of the survivors or their long-term impact on the affected country. Classification of Disasters In the early 1970s, a series of well-publicized disasters (the civil war and resulting famine in Biafra, the cyclone in Bangladesh, and the earthquake in Peru) triggered the scientific interest of the international public health community. Disasters can be classified as natural disasters, technological disasters, or complex emergencies. The latter include civil wars and conflicts. These classifications are arbitrary and refer to the immediate trigger—a natural phenomenon or hazard (biological, geological, or climatic); a technologically originated problem; or a conflict. In reality, all disasters are complex events stemming from the interaction of external phenomena and the vulnerability of man and society. The human responsibility in so-called natural disasters is well acknowledged. The term natural disaster remains commonly used and should not be understood as denying a major human responsibility for the consequences. Disaster Terminology The following definitions are adapted from those proposed by the Secretariat of the International Strategy for Disaster Reduction (ISDR), a United Nations (UN) body established to sustain the efforts of the International Decade for Natural 1147 ©2006 The International Bank for Reconstruction and Development / The World Bank 189 Disaster Reduction (UN/ISDR 2004) and the WHO World Health Report 2002 (WHO 2002): US$ billions 700 • Hazards are potentially damaging physical events, which may cause loss of life, injury, or property damage. Each hazard is characterized by its location, intensity, frequency, and probability. • Vulnerability is a set of conditions resulting from physical, social, economic, and environmental factors that increase the susceptibility of a community to the effects of hazards. A strong coping capacity—that is, the combination of all the strengths and resources available within a community—will reduce its vulnerability. • Risk is the probability of harmful consequences (health burden) or economic losses resulting from the interactions between natural or human-induced hazards and vulnerable or capable conditions. In a simplified manner, risk is expressed by the following function: 600 12 500 10 400 8 300 6 200 4 100 2 0 0 Richest Nations Economic losses Figure 61.1 Disaster Losses, Total and as Share of Gross Domestic Product, in the Richest and Poorest Nations, 1985–99 8 6 4 Geographic Distribution of Risk. Natural disasters do not occur at random. Geological hazards (earthquakes and volcanic eruptions) occur only along the fault lines between two tectonic plates on land or on the ocean floor. However, the local population often does not recognize the implications (the risks), as shown in the December 2004 tsunami in the Indian Ocean. Rockslides 10 El Niño phenomenon Annual growth of GDP in Ecuador compared with preceding year (percent) 12 0 ⫺2.0 ⫺4.0 ⫺6.0 ⫺8.0 19 8 19 0 8 19 1 8 19 2 8 19 3 8 19 4 8 19 5 8 19 6 8 19 7 8 19 8 8 19 9 9 19 0 9 19 1 9 19 2 9 19 3 9 19 4 9 19 5 9 19 6 9 19 7 9 19 8 9 20 9 0 20 0 01 Health and relative economic losses of natural disasters disproportionately affect developing countries (Alexander 1997; UN/ISDR 2004). More than 90 percent of natural disaster– related deaths occur in developing countries. Even though the economic losses are far greater in industrial countries, the percentage of losses in relation to gross national product (GNP) in developing countries far exceeds that percentage in industrial countries (figure 61.1).At an individual level,a sudden reduction of US$5,000 from an annual income of US$50,000 is worrisome; however, the ongoing loss of US$50 from a monthly income of US$100 may be catastrophic. For this reason, statistics of economic damage and mortality alone are not true indicators of the effect of disasters on the health and development of people and communities. Disaster impact statistics show a global trend: more disasters occur, but fewer people die; larger populations are affected, and economic losses are increasing (IFRC 2000). Losses as percentage of GDP Source: UN/ISDR, 2004. 2 Distribution and Risk Factors Poorest Nations Earthquake North-East A public health approach to disaster risk management will aim to decrease the vulnerability by adopting prevention and mitigation measures to reduce the physical impact and to increase the coping capacity and preparedness of the health sector and community, in addition to providing traditional emergency care (response) once the disaster has occurred. 14 El Niño phenomenon Risk ⫽ ƒ (Hazards ⫻ Vulnerability) Percentage of GDP 16 Years Source: UN/ISDR, 2004. Figure 61.2 Annual Growth of Gross Domestic Product and Occurrence of Major Natural Disasters in Ecuador, 1980–2001 Hydrometeorological hazards do not follow a wellestablished distribution. Although the areas subject to seasonal flood, drought, or tropical storms (cyclones, hurricanes, or typhoons) are well known locally, global warming may possibly redraw the map of climatic disasters. As the National Research Council (1999, 34–35) notes, “This change is far from uniform. A pattern of response ‘modes’ appears to be involved, in which warming is concentrated in northern Asia . . . while large regions of the northern Pacific and North Atlantic Oceans and their neighboring shores have actually cooled.” El Niño–related fluctuations in relation to the gross domestic product (GDP) of Ecuador are shown in figure 61.2. 1148 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio ©2006 The International Bank for Reconstruction and Development / The World Bank 190 The risk of massive technological disasters, such as the catastrophic release of chemicals in Bhopal, India (methyl isocyanate), in December 1984, is serious in countries with significant industry (WHO 1992, 1996). Very few countries are immune to public health risks from hazardous chemical substances (from insecticides to industrial by-products) or discarded radioactive material from therapeutic or diagnostic use. Technological hazards increase rapidly with the unregulated industrialization of developing countries and the globalization of the chemical industry, suggesting that chemical emergencies may become a major source of disasters in the 21st century. locations, particularly urban areas. Following Hurricane Mitch in Tegucigalpa, Honduras, families that were relocated from flooded areas to safer (but inconveniently remote) ground were rapidly replaced by new illegal settlers. In 2003, families killed by a landslide in Guatemala had been warned about their vulnerability but were unable to afford resettlement in safer (and more costly) areas. Subsidies alone may not have prevented this effect, given the overarching issue of land ownership by a few in Central America. Factors Affecting Vulnerability. Vulnerability to all types of disasters—and to poverty—is linked to demographic growth, rapid urbanization, settlement in unsafe areas, environmental degradation, climate change, and unplanned development. Losses fall under three categories, which may have both direct and indirect components: Age The importance of age as a factor of vulnerability can be significant in situations where physical fitness is necessary for survival. The higher fatality among children, elderly, or sick adults following the 1970 tidal wave in Bangladesh (250,000 fatalities) and the 2004 tsunami in Asia (more than 180,000 dead or missing) illustrates this point. Gender Reports on immediate morbidity and mortality according to gender are not as conclusive. An Inter-American Development Bank paper indicated that 54 percent of the 3,045 people who died as a result of Hurricane Mitch in Nicaragua were male (IDB 1999). Stereotypes of gender vulnerability at the time of impact often do not apply. Depending on the type of disaster, far more significant vulnerability factors than gender or age are the time of day of the impact (and, therefore, the occupational activity of each group) and the structural vulnerability of housing, factories, and public buildings, including the location of the victims within the buildings. Following disasters, increased vulnerability of women is commonly noted in temporary settlements, where violence and sexual abuse are common. Specialized health care also may not be available (Armenian and others 1997). Poverty Economic vulnerability might play a much greater role than age and gender. What has been noted regarding the greater vulnerability of poor countries also holds true at the community and family levels. Disasters predominantly affect the poor. Poverty increases vulnerability because of the unequal opportunity for healthy and safe environments, poor education and risk awareness, and limited coping capacity. A notable exception was the 2004 tsunami in Banda Aceh, Indonesia, where the middle- and upper-class neighborhood close to the shore was particularly affected. A major example is the settlement of a large number of economically disadvantaged populations in highly vulnerable Short-Term Health Burden • lives and disabilities (both direct damage and an indirect consequence) • direct losses in infrastructure and supplies (direct impact) • loss or disruption in the delivery of health care, both curative and preventive (indirect impact). The immediate health burden is directly dependent on the nature of the hazard. National health budgets of developing countries are, in normal times, insufficient to meet the basic health needs of the population. In the aftermath of a major disaster, authorities need to meet extraordinary rehabilitation demands with resources that often have been drained by the emergency response (as distinct from the resources destroyed by the event). Beyond the immediate response, decision making in the allocation of resources among sectors is mostly influenced by the magnitude of the economic losses rather than by the health statistics (principally the disability-adjusted life year, or DALY, losses) or social costs. Earthquakes. As noted by Buist and Bernstein (1986), in the past five centuries, earthquakes caused more than 5 million deaths—20 times the number caused by volcanic eruptions. In a matter of seconds or minutes, a large number of injuries (most of which are not life-threatening) require immediate medical care from health facilities, which are often unprepared, damaged, or totally destroyed, as was the case in the earthquake in Bam, Iran, in 2003. In the aftermath of that earthquake, which resulted in 26,271 deaths, the entire health infrastructure of the city was destroyed. All traumas were evacuated by air to the 13 Iranian provinces long before the arrival of the first foreign mobile hospitals. Table 61.1 illustrates the accelerated pace with which priorities evolve and overlap in the first week following an earthquake. After a few weeks, national political solidarity and external assistance wane, and the local budgetary resources are drained. At the same time, health authorities face the overwhelming task of providing services to a displaced population, rehabilitating Natural Disaster Mitigation and Relief | 1149 ©2006 The International Bank for Reconstruction and Development / The World Bank 191 Table 61.1 Health Priorities Following Earthquakes Priority Time period Comments Search and rescue 0 to 48⫹ hours Returns are rapidly diminishing. Most effective work is done by local teams. Trauma care 0 to 48 hours: initial lifesaving carea 48 hours to 6 months: secondary care External assistance generally arrives too late for initial care. Traumas may include burns and crush syndrome, especially in urban areas. Paraplegics and amputees require long-term care. Routine medical emergencies and primary health care Resumes as soon as the need for acute lifesaving care subsides (within 24 hours) Emergencies include earthquake-related cardiovascular emergencies and premature births. Attention to the dead Varies. Not a public health issue but a social and political one Priorities are identification and ritual burial. Disease surveillance Urgent—within 48 hours, unsubstantiated rumors of impending epidemics will be circulating Surveillance is a sensitive public information and education issue. A simple, syndrome-based system is needed that will involve humanitarian organizations. Provision of safe water A predominant issue within 48 hours The challenge is to provide a sufficient quantity of reasonably safe water. Temporary shelter 48 hours to several months Sanitation and provision of health services is a main issue. Accommodating families near their residence is preferable to setting up camps. Provision of food 3 days to 6 weeks Food provision is a social or economic issue. Food stocks and agricultural output are not affected by earthquakes. Psychosocial care 7 days to 6 months Mental health assistance is best provided by local personnel, if available. Source: de Ville de Goyet 2001. a. Following the earthquake in Mexico City in 1985 (10,000 deaths), bed occupancy rates did not exceed 95 percent despite the loss of 5,829 hospital beds. health facilities, restoring normal services, strengthening communicable disease surveillance and control, and attending to the long-term consequences, such as permanent disabilities, mental health problems, and possibly long-term increases in rates of heart disease and chronic disease morbidity (Armenian, Melkonian, and Hovanesian 1998). Tsunamis. Earthquakes on the ocean floor may cause catastrophic tidal waves (tsunamis) on faraway shores. Waves caused by the seismic event crest at less than a meter in open seas, but they are travel several hundred kilometers per hour, so when they reach shallow waters, they can be 10 meters high. Damage on the coast can be extensive. Usually, the number of survivors presenting severe injuries is small in proportion to the number of deaths. Volcanic Eruptions. Volcanoes persist as a serious public health concern, though they are often overlooked by authorities and communities lulled by long periods of inactivity. Eruptions are preceded by a period of volcanic activity, which provides an opportunity for scientific monitoring, warning, and timely evacuation. Some issues, such as ash fall, lethal gases, lava flow, and projectiles, although of concern to the public, are of minimal health significance: Ash fall causes a significant burden on medical services but is unlikely to result in excess mortality or significant permanent problems. However, ash fall affects transportation, communications, water sources, treatment plants, and reservoirs. Studies by Bernstein, Baxter, and Buist (1986) following the 1980 eruption of Mount St. Helens (United States) reviewed the transient, acute irritant effects of volcanic ash and gases on the mucous membranes of the eyes and upper respiratory tract as well as the exacerbation of chronic lung diseases with heavy ash fall. Concentrations of volcanic gases are rapidly diluted to nonlethal levels, which lead to inconvenience but negligible morbidity for the general public. Lava flows present little health risk because of their very slow speed of progression. Mortality caused by ballistic projectiles from a volcanic eruption is minimal. Attention to these public concerns may distract the authorities from preparing for the greatest factors of mortality: the pyroclastic flows (Mount Pelé in Martinique, in 1902, with 29,000 deaths) and lahars. Lahars are mud flows or mud and ash flows caused by the rapid melting of a volcano’s snowcap, as in Colombia in 1985 (23,000 deaths), or caused by heavy rains on unstable accumulations of ash, as in the Philippines in 1991. Historically, pyroclastic explosions or lahars have caused about 90 percent of the casualties from volcanic eruptions. Potential contamination of water supplies by minerals from ash; displacement of large populations for an undetermined period of time (over five years in Montserrat, a small island in the Caribbean); accompanying sanitation problems; and mental health needs are of great public health significance (PAHO 2002a). Among the long-term problems, the risk of developing silicate pneumoconiosis requires further investigation.1 1150 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio ©2006 The International Bank for Reconstruction and Development / The World Bank 192 Climatic Disasters. Many communities and health services have learned to live with seasonal floods of moderate intensity. Periodically, the magnitude of the phenomenon exceeds the local coping capacity and overwhelms the resources of the health systems. The health burden associated with seasonal floods is well known locally: increased incidence of diarrheal diseases, respiratory infections, dermatitis, and snake bites. The actual risk of compromised water supplies depends on the level of contamination of the community’s water supply before the disaster, compared with contamination after the flooding. Saline contamination is a long-term issue following sea surges and tsunamis. Prolonged flooding endangers local agriculture and occasionally requires food assistance on a large scale. The primary factors of morbidity remain overcrowded living conditions and poor water and sanitation in temporary settlements and other areas where water and sanitation services have deteriorated or are suspended. Mortality and morbidity caused by tropical storms (hurricanes in the Atlantic Ocean and typhoons in the Pacific Ocean) result from, in increasing order of importance, high winds, heavy rainfall, and storm surge. When Hurricanes Mitch and George hit the Caribbean in 1998, traumatic injuries (lacerations or electrocution) caused by high winds of up to 150 miles per hour were relatively few; deaths from extensive rainfall (leading to flash floods and landslides) constituted the bulk of the more than 13,000 fatalities (PAHO 1999). In the Bangladesh delta, storm surges up to 6 meters traveled unimpeded over hundreds of kilometers and claimed between 250,000 and 500,000 lives in 1970 and up to 140,000 lives during five cyclones in the 1990s—primarily during one storm in 1991. Another cost is the need for specialized psychosocial assistance to large numbers of the population who survive the sustained violence of nature. Cumulative mortality caused by small, undocumented mudslides and rockslides from water-saturated, unstable slopes probably approach the toll from well-known landslides (earthquakes in Peru in 1970 and in El Salvador in 2001, and the rains in Caracas, Venezuela, in 1999). Morbidity problems are often minimal, as survivors in the path of the landslide are few. Impact on Communicable Diseases Disasters related to natural events may affect the transmission of preexisting infectious diseases. However, the imminent risk of large outbreaks in the aftermath of natural disasters is overstated. Among the factors erroneously mentioned is the presence of corpses of victims, many buried beneath rubble. Dead bodies from a predominantly healthy population do not pose a risk of increased incidence of diseases (Morgan 2004). Catastrophic incidence of infectious diseases seems to be confined to famine and conflicts that have resulted in the total failure of the health system. In the short term, an increased number of hospital visits and admissions from common diarrheal diseases, acute respiratory infections, dermatitis, and other causes should be expected following most disasters (Howard, Brillman, and Burkle 1996; Malilay and others 1996). This increase may reflect duplicate reporting (diarrhea cases were reported through both the emergency and the routine surveillance systems in Maldives after the 2004 tsunami), a temporary surge in surveillance, and medical attention available to an otherwise underserved population rather than representing a genuine change in the epidemiological situation. In the medium term, heavy rainfalls may affect the transmission of vectorborne diseases. Following an initial reduction as mosquito-breeding sites wash away, residual waters may contribute to an explosive rise in the vector reservoir. When associated with a breakdown of normal control programs, this rise in the vector reservoir may lead to epidemic recrudescence of malaria or dengue. Retrospective studies (Bouma and Dye 1997; PAHO 1998; UN/ISDR 2004, 156) all confirm a direct but delayed relationship between the intensity of rainfall (regardless of the existence of flooding) caused by the El Niño phenomenon and the incidence of malaria. Flooding has contributed to local outbreaks of leptospirosis (in Brazil and Jamaica, for example; PAHO 1982) and hepatitis A in Latin America and Africa (WHO 1994). In summary, what can be expected and prevented is a local surge in problems that the health services are normally used to handling. Long-Term Impact and Economic Valuation In addition to the delayed impact on transmission and control of endemic diseases and the burden of disabilities (paraplegia, amputation, burns, or chronic or delayed effects of chemical or radiological exposure), the health sector bears a significant share of the economic burden. Disasters must be seen in a systemic (that is, intersectoral) manner: what affects the economy will affect the health sector—and vice versa. After the emotional response of the first few days, decision makers in a crisis react primarily to political and economic realities, not to health indicators. Economic valuation of the social burden—that is, placing a monetary value on the cost—becomes a critical tool as the various sectors compete for scarce resources. The health sector, in particular, must learn how to use this tool in spite of being absorbed by its immediate relief responsibilities. Valuation of Disasters. The Economic Commission for Latin America and the Caribbean (ECLAC) has developed over the decades a methodology for the valuation of disasters (ECLAC 2003). This tool, intended for reconstruction, has also proved its usefulness by developing historical records of major events, particularly of the health burden expressed in economic terms. Natural Disaster Mitigation and Relief | 1151 ©2006 The International Bank for Reconstruction and Development / The World Bank 193 Valuation is made using all possible sources of information, from georeferenced satellite mapping and remote sensing to more conventional statistical data, direct observation, and surveys, with a reliance on information gathered immediately after the event. Economic valuation rests on the basic concepts of direct damage and indirect losses. Direct damage is defined as the material losses that occur as an immediate consequence of a disaster.2 Direct damage is measured first in physical terms. The physical loss includes assets, capital, and material things that can be counted: hospital beds lost, equipment and medicines destroyed, damaged or affected health service installations (number and type of installations, stocks of medicines, laboratory facilities, operating rooms, and so on), and pipes and water plants destroyed. The physical plant then is valued both in terms of discounted present value and estimated replacement cost. Reconstructing facilities with the same vulnerability and level of service as before would be unacceptable; the affected health infrastructure must be replaced by more resilient and efficient installations to ensure better and sustainable service. This need is most evident in developing countries where impacts tend to be concentrated in those most at risk (the poor, marginalized, and less resilient sectors of the population). Indirect effects refer to production of goods and services that will not occur as an outcome of the disaster, reduced income associated with those activities not occurring, and increased costs to provide those goods and services. In the case of health services, indirect effects encompass both the income losses associated with the diminished supply of health care services and the increased costs of providing the services following the disaster. Indirect effects are valued at the current market value of goods or services not produced and the costs associated with the necessary provision of services under emergency, disaster-related conditions. Both compo- nents of the cost of illness—the cost of treatment and the cost of lost opportunities (lost income and employment, loss of time and productivity)—are sharply increased. The social burden is heavier on the poorest, who are unable to adjust their willingness to pay to absorb the additional expenses of alternative (private) providers of care. The same approach applies to the economic valuation of lives lost. Kirigia and others (2004) found a statistically significant impact of disaster-related mortality on the GDP of African countries. One single disaster death reduced the GDP per capita by US$0.01828. Lost lives are given a higher economic value in places where productivity is high. Because economic valuation uses standard sectored procedures that allow comparability of results, it can be used in the decision-making process and for policy formulation since it identifies sectors, geographical areas, and vulnerable groups that are more severely affected economically. Over the years, a number of conceptual improvements have been made to allow for the measurement of aspects not included in national accounting systems—to bring attention to environmental losses as a cross-cutting issue; to highlight the contribution of specific groups, namely women, as agents for change; and to focus on the better management of both the emergency and the reconstruction processes. It is also a valuable tool for preparedness and mitigation of future damage. Table 61.2 summarizes the valuations made by ECLAC over the years for Latin America and the Caribbean in terms of deaths, affected populations, and economic losses (2003 values). Of interest are the decrease in the number of deaths and the increase in total damage (in particular, indirect damage) over time. The distribution of direct and indirect damage in the health sector also varies. According to ECLAC (2003), direct damage between 1998 and 2003 in Latin America ranged from 44.6 percent to 77.2 percent of total damage. Table 61.2 Impact of Disasters in Latin America and the Caribbean Population Date Damage (2003 US$ millions) Deaths Affected Total Direct Indirect 1972–80 38,042 4,229,260 9,376 5,420 3,956 1981–90 33,638 5,442,500 19,603 13,916 5,687 1991–2000 11,086 2,318,508 20,902 10,401 10,501 2001–2002 Total of major events 1972–2002 Overall estimate including small disastersa Average per year 120 4,828,470 4,498 2,270 2,228 82,886 16,818,738 54,379 32,007 22,372 103,608 21,023,422 67,974 40,009 27,965 3,454 700,781 2,266 1,334 932 Source: ECLAC 2003. a. The full image should include the recurrent small disasters that do not make the headlines but have a cumulative negative effect. Such disasters can be more pervasive and damaging to the development process because their economic, social, psychological, and political effects are hardly perceived. An estimate of the average losses of small disasters would be at least 25 percent greater than those of large disasters. 1152 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio ©2006 The International Bank for Reconstruction and Development / The World Bank 194 Specific Damage to the Health Infrastructure. Damage to housing, schools, channels of communication, industry, and so on contributes to the health burden. However, the following analysis focuses on the health infrastructure (understood as health care facilities, including hospitals, health centers, laboratories, and blood banks) and the drinking water and sanitation infrastructure. Damage to Hospitals and Health Installations Most data and examples presented here come from Latin America and the Caribbean because of the disaster reduction programs in the health sectors of those regions. In the past two decades, damage to approximately 260 hospitals and 2,600 health centers resulted in interruption of services at a direct cost of US$1.2 billion. In the 1985 earthquake in central Mexico, 5,829 beds were destroyed or evacuated (PAHO 1985), at a direct cost of US$550 million (ECLAC 1998). Hurricane Gilbert (1988) damaged 24 of the 26 hospitals on Jamaica, and the El Salvador earthquake (2001) resulted in the loss of 2,000 beds—40 percent of the country’s hospital capacity (PAHO 2002b). The health burden is not limited to the loss of medical care. The control of communicable diseases and other public health programs suffer from loss of laboratory support and diagnostic capabilities of hospitals. Further research on the actual impact of these losses, in terms of DALYs, is essential. A common misperception is that damage to critical health facilities is promptly repaired. Experience shows that damaged health infrastructure recovers at a slower pace than infrastructure in other service sectors, such as trade, roads, bridges, telecommunications, and even housing. For example, as a result of the earthquake that affected El Salvador in 1986, renovation of the general hospital, the most sophisticated referral hospital in the capital, was completed 15 years after the earthquake. The only national pediatric facility was fully rehabilitated and strengthened six years after the earthquake. Two years after the earthquake of 2001 in El Salvador, several key hospitals still remained vacated or services were transferred to unsuitable temporary facilities. The factors are many: low priority assigned to a nonproductive sector, the sector’s inexperience in developing comprehensive proposals for funding, conflicting attempts to use the reconstruction process to influence the ongoing reform and decentralization processes, the novelty of the engineering and design issues for safe hospital construction, the complicated negotiation process for loans, and the administrative inexperience of the health sector in executing large investment projects. Indeed, few large health installations have been built directly by developing countries in the past decades. Damage to Water and Sewage Systems The primary goal of water and sewage systems is to safeguard the public health of the population. For that reason, these systems are considered part of the health infrastructure. The developmental burden is significant. In the past 30 years in Latin America and the Caribbean alone, an estimated 400 urban water supply systems and 1,300 rural systems (in addition to 25,000 wells and 120,000 latrines) were severely damaged, at an estimated cost of almost US$1 billion—a major setback to efforts to expand coverage and improve those services. In severe flooding, the sudden interruption of these basic services coincides with the direct effect on the transmission of waterborne or vectorborne diseases. In the case of earthquakes, the number of people who are adversely affected by water shortage may far exceed those injured or suffering direct material loss. As in the case of health care facilities, the rehabilitation of public water systems is slow, particularly for communityowned or community-operated rural systems, which may not be repaired for decades. The foregoing demonstrates the need for water authorities to harmonize their short-term objectives, which are oriented almost exclusively to increasing the coverage of these services, with the long-term objective of reducing vulnerability to extreme natural hazards. INTERVENTIONS: FROM RESPONSE TO PREVENTION The immediate lifesaving response time is much shorter than humanitarian organizations recognize. In a matter of weeks, if not days, the concerns of both the population and authorities shift from search and rescue and trauma care to the rehabilitation of infrastructure (temporary restoration of basic services and reconstruction). In Banda Aceh, Indonesia, after the December 2004 tsunami, victims were eager to return to normalcy while external medical relief workers were still arriving in large numbers. Response and Rehabilitation Immediate emergency response is provided under a highly political and emotional climate. The public demands visible, albeit perhaps unnecessary, measures at the expense of proven low-key approaches. The international community, eager to demonstrate its solidarity or to exercise its“right of humanitarian intervention,”undertakes its own relief effort on the basis of the belief that local health services are unwilling or unable to respond. Donations of useless medical supplies and medicines and the belated arrival of medical or fact-finding teams add to the stress of local staff members who may be personally affected by the disaster. The cultural disregard of the humanitarian community to cost-effective approaches in times of disaster and the tendency to base decisions on perceptions and myths rather than on facts and lessons learned in past disasters contribute to making disaster relief one of the least cost-effective health activities. The responsibilities of the national or local health authorities are significant. Natural Disaster Mitigation and Relief | 1153 ©2006 The International Bank for Reconstruction and Development / The World Bank 195 Assessment of the Health Situation. A country’s ministry of health is expected to assess the health situation. To influence the course of humanitarian response, this assessment must be rapid and, therefore, simple; transparent in collaboration with the main actors—nongovernmental organizations (NGOs) and donors; and technically credible. The input of WHO, as the lead agency in health matters, is most valuable. Confusion should be avoided between assessing emergency needs and inventorying or valuating the damage. In the first hours or days, relief actors base their decision making on the ministry of health’s assessment of what is required and, more importantly, what is not required for emergency response. Later, the international community will request detailed data, such as the number of persons affected, buildings damaged, and monetary valuation. Mass Casualties Treatment. Following natural disasters, hospital capacity may be considerably reduced by actual damage to the facility or, in the case of a seismic event, an often unnecessary—but hard to reverse—evacuation. Triage of patients is required in order to first treat those likely to benefit most, rather than the terminally injured or those whose care can be delayed. Lifesaving primary care takes place in the first six hours (the golden rule of emergency medicine), making most of the foreign field hospitals irrelevant for intensive acute care of traumas (WHO and PAHO 2003). Effectiveness of immediate care will depend on local preparedness before the disaster, not on faraway resources. Strengthened Surveillance, Prevention, and Control of Communicable Diseases. Because the surveillance, prevention, and control of communicable diseases are strengthened, the anticipated massive outbreaks generally do not actually occur. Traditional surveillance systems that are based on periodic notification of diseases by the health services are inadequate in a crisis situation. Early warning requires flexible and simple syndrome-based monitoring in temporary settlements and health centers, with information collected not only by the official health services but also by the medical humanitarian organizations. Systems that do not include consultation with NGOs are unlikely to succeed. Disease control programs in place before the disaster are the fruit of local experience and external technical advice. In a disaster situation, there is generally no need to resort to new and expensive control measures. The key is to quickly resume, strengthen, and better monitor the routine control programs. No public health concerns justify the hurried disposal of corpses through mass burial or unceremonious incineration. This practice is socially and culturally damaging. In addition, improvised mass immunization campaigns, especially by external relief groups, should be discouraged in favor of opportunistically strengthening national routine immunization coverage, especially in temporary settlements. Environmental Health. Typical interventions in the aftermath of disasters include strengthening the monitoring and surveillance of water quality, vector control, excreta disposal, solid waste management, health education, and food safety. A first priority is water supply. It is often preferable to have a large quantity of reasonably potable water than a smaller amount of high-quality water (UNHCR 1998). Massive distribution of water purification disinfectants can be effective if the public is already familiar with their use and not confused by the availability of many different brands and concentrations of donated chemicals. Health education and hygiene promotion efforts target populations in shelters, temporary camps, collective kitchens, or prepared food distribution centers. The cost-effectiveness of the external relief effort could often be increased by shifting resources from the overattended medical response to the improvement of environmental health in temporary settlements. Transparent Management of Donations and Supplies. If donations and supplies are managed transparently during the emergency, the flow of assistance to the intended beneficiaries will be improved. Unsolicited and often inappropriate medical donations compete with valuable relief supplies for scarce logistical resources. Good governance is critical, and effective logistics cannot be improvised following a disaster. A humanitarian supply management system developed by PAHO and WHO successfully helped developing countries improve transparency and accountability in managing humanitarian supplies and donations (de Ville de Goyet, Acosta, and others 1996). Coordination of the Humanitarian Health Effort. Coordination of the humanitarian health effort is essential to maximize the benefit of the response effort and ensure its compatibility with the public health development priorities of the affected country. Effective coordination in the health sector must do the following: • Be comprehensive and include all external health actors. • Be based on mutual respect rather than regulatory authority alone. Dialogue and consultation are more effective than enforcement. • Benefit all parties, starting with the victims. It should aim to support and facilitate the activities of other partners. • Be evidence-based and transparent. Information is made to be shared and used, not jealously guarded. Coordination cannot be improvised in the aftermath of a disaster. Preparedness before the occurrence of the hazard is essential. 1154 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio ©2006 The International Bank for Reconstruction and Development / The World Bank 196 Emergency Preparedness of the Health Sector Prevention and Mitigation Effective response by national health authorities cannot be impromptu. Ministries of health that neglected to invest in capacity building before emergencies have generally experienced serious difficulties in exercising their technical and political leadership in the immediate aftermath of a disaster. Disaster preparedness is primarily a matter of building institutional capacity and human resources, not one of investing heavily in advanced technology and equipment. Building local coping capacity is one of the most costeffective ways to improve the quality of the national response and the external interventions. Disaster preparedness is not merely having a disaster plan written by experts. It must involve the following: The slogan “prevention is better than cure” was invented by the health sector. However, this sector has been slow to adopt the concept of preventing deaths and injuries from disasters through the mitigation (that is, reduction) of damage to its own facilities. As is unfortunately often the case, political action is often triggered only by a major disaster, such as the collapse of Hospital Juarez in Mexico in the earthquake of 1985; in that disaster 561 patients and employees died, (Poncelet 1997). Evaluating the damage (the past vulnerability) helps establish mitigation criteria for the future. The level of protection required for each health installation must be negotiated—from life protection, which prevents an immediate structural collapse to permit the evacuation of people; to investment protection, which minimized the economic losses; to operational protection, which guarantees the sustainability of services under any extreme circumstances. Though a commercial or office building may be structurally designed only to prevent loss of lives, key hospitals must remain operational during the times they are most needed. Local engineering and architectural experts play a key role in developing the knowledge, technical abilities, and costeffectiveness analysis to establish mitigation priorities. Technical mitigation guidelines prepared at a global level (PAHO, WHO, World Bank, and ProVention Consortium 2004) need to be adapted to local culture, conditions, and resources. Reducing the physical vulnerability of infrastructure can take place on three different occasions (UN/ISDR 2004, 324): • Identifying vulnerability to natural or other hazards. The health sector should seek information and collaborate with other sectors and institutions (civil protection, meteorology, environment, geology) that have the primary responsibility for collecting and analyzing this information. • Building simple and realistic health scenarios of a possible and probable occurrence. It is challenging enough to prepare for a moderate-size disaster; building and sustaining a culture of fear based on unrealistic worst-case scenarios may serve the corporate interests of the disaster community but not the interests of the public at large. • Initiating a participative process among the main actors to develop a basic plan that outlines the responsibilities of each actor in the health sector (key departments of the ministry of health, medical corps of the armed forces, private sector, NGOs, UN agencies, and donors). What matters is the process of identifying possible overlaps or gaps and building a consensus—not the paper plan itself. Disasters often present problems that are unforeseen in the most detailed plans. • Maintaining a close collaboration with these main actors. A good coordinator is one who appreciates and adapts to the strengths and weaknesses of other institutions. Stability is essential. Changes of key emergency staff members during a disaster situation or when a new administration or minister take over have occasionally complicated the tasks. • Sensitizing and training the first health responders and managers to face the special challenges of responding to disasters. Participation of external actors (UN agencies, donors, or NGOs) in designing and implementing the training is critical. The incorporation of disaster management in the academic curriculum of medical, nursing, and public health schools should complement the on-the-job training programs of the ministry of health, UN agencies, and NGOs. Well-designed disaster management training programs will improve the management of daily medical emergencies and accidents as well. • When reconstructing the infrastructure destroyed by a disaster. At that time, risk awareness is high, political will is present, and resources are available. • When planning new infrastructure. Reducing vulnerability is most cost-effective and politically acceptable when it is included at the earliest planning and negotiation stage, whether it involves a 1 to 2 percent additional cost for wind resistance or a 4 to 6 percent additional cost for earthquake resilience. Full resistance to any damage is prohibitively expensive. • Strengthening of existing facilities (retrofitting). This most expensive measure has been adopted by several developing countries (Chile, Colombia, Costa Rica, Mexico, Peru, and others) to protect their most critical health facilities. In the earthquake in Colombia in 1999, partial retrofitting of the main hospital is credited for saving the installation. Costs vary greatly (see table 61.3). Mitigation of Damage to Hospitals. Mitigation does not pretend to eliminate all possible damage from hazards but aims to ensure the continuing operation of the health facility at a level previously defined by the health authority. Hospitals Natural Disaster Mitigation and Relief | 1155 ©2006 The International Bank for Reconstruction and Development / The World Bank 197 Table 61.3 Retrofitting of Hospitals in Costa Rica Hospital Hospital Mexico Number of beds Duration of retrofitting (months) Cost of retrofitting (US$) Percentage of total value of the hospital 600 31 2,350,000 7.8 Children’s Hospital 375 25 1,100,000 4.2 Hospital Monseñor Sanabria 289 34 1,270,000 7.5 Source: PAHO and WHO 2000. should be subject to stricter norms than other less critical facilities that are designed to prevent only total collapse and loss of life. Hospital mitigation interventions fall into three categories: • Functional mitigation to ensure that the necessary supporting infrastructure services permit continuing operation: water, electricity, road access, communications, and so forth. Improving routine maintenance will facilitate operations under normal circumstances and in the event of extreme hazards. • Nonstructural mitigation to reduce losses and health injuries from falling or moving objects. Measures include, for instance, proper anchoring of equipment for earthquakes or strong winds or the location of only noncritical services on flood-prone floors. • Structural mitigation to ensure the safety of the structure itself (columns, beams, load-bearing walls). Given the high economic, health, and political costs represented by the avoidable loss of critical health facilities, health authorities and funding agencies should require that, in all new health infrastructure projects, natural hazards be a decisive factor for selecting the facility’s location and for formulating the specifications at the earliest stage of the process. Mitigation of Damage to Water Systems. Unlike hospitals, water supply systems are geographically extensive and thus are exposed to different types of hazards. The search for technical solutions is more complex, given the diversity of the water system’s components. Finally, in many countries, the health authorities have no jurisdiction over the construction or operation of those services owned or administered by many local or municipal agencies. Even a short disruption of water services may have serious and direct implications for the health of individuals, the operation of health services, and the community at large through its impact on business. A probabilistic model studied the disruptive potential of a water outage in the event of an earthquake in Los Angeles county in the United States. As noted by the authors, “water outage is more likely to be disruptive for businesses in some industries, such as health services, than for others” (Chang and Chamberlin 2004, 89). The health sector should, therefore, coordinate with the institutions in charge of constructing, operating, and maintaining water and sanitation services, both urban and rural, to promote reduction of the vulnerability of existing systems. The health sector should also ensure that health aspects and mitigation of damage be included in the regulatory framework and operating procedures of water and sanitation services. Protecting the water supply is feasible in developing countries. The Costa Rican Institute of Aqueducts and Sewage Systems reduced the vulnerability of one of the main aqueducts of the country, the Orosi Aqueduct. Over 10 years, Costa Rica invested almost US$1.5 million in studies and reinforcements, an amount equivalent to 2.3 percent of the total cost of the aqueduct. This investment would prevent a loss of nearly US$7.3 million in direct damages alone (FEMICA 2003). INTERVENTION COST, COST-EFFECTIVENESS, AND ECONOMIC BENEFITS The highly emotional and sensationalized climate of disaster response has long prevented the adoption of a costeffectiveness approach in decision making. When survival of both people and political institutions is threatened, perceptions and visibility tend to prevail over facts and analysis, resulting in a lack of evidence-based studies on costs and benefits. The willingness to spend hundreds of thousand of dollars per victim rescued from a collapsed building in a foreign country is a credit to the solidarity of the international community, but it also presents an ethical issue when, once the attention has shifted away, modest funding is unavailable for the mid-term survival of tens of thousands of victims. Cost-Effectiveness of Selected Humanitarian Interventions Emergency health interventions are more costly and less effective than time-tested health activities. Improvisation and rush inevitably come with a high price. The preferential use of expatriate health professionals; the emergency procurement and airlifting of food, water, and supplies that often are available locally or that remain in storage for long periods of time; and the tendency to adopt dramatic measures contribute to 1156 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio ©2006 The International Bank for Reconstruction and Development / The World Bank 198 making disaster relief one of the least cost-effective health activities. Search and Rescue. Few developing countries have established the technical capacity to search for and attend to victims trapped in confined spaces in the event of the collapse of multistory buildings. Industrial nations routinely dispatch search and rescue (SAR) teams. Costs are high and effectiveness is reduced by delayed arrival and quickly diminishing returns. Following the 1988 earthquake in Armenia, in the former Soviet Union, the U.S. SAR team extracted alive only two victims at a cost of over US$500,000. In Turkey in 1999, 98 percent of the 50,000 people pulled alive from the rubble were salvaged by relatives and neighbors. In Bam in 2003, the absence of high-rise and reinforced concrete buildings ruled out the need for specialized teams. Nevertheless, according to UN statistics, at least US$2.8 million was spent on SAR teams. An alternative solution consists of investing these resources in building the capacity of local or regional SAR teams—the only ones able to be effective within hours—and training local hospitals to dispatch their emergency medical services to the disaster site. Field Hospitals. The limited lifesaving usefulness of foreign field hospitals has been discussed. Again, the lessons learned from the Bam earthquake are clear. The international community spent an estimated US$10.5 million to dispatch approximately 10 mobile hospitals,3 which arrived from two to five days after the impact, long after the last casualty had been evacuated to other Iranian provinces. This delay alone, hard to reduce further, rules out any significant contribution to immediate trauma care and led the hospitals to compete for routine outpatient care with the teams of Iranian volunteers from across the country. A few of the mobile hospitals, better prepared to meet nontrauma needs and to stay much longer than the usual two to three weeks, have been invaluable. No data are available on the number of lives actually saved by mobile hospitals (that is, lives that would not have been saved by local means). Less understood are the negative effects of such hospitals on local health services, which are often marginalized and discredited for their lack of technology and sophistication but which must cope once the external facility leaves. The cost of mobilizing a mobile hospital for a few weeks often exceeds US$1 million, funds that would be more productive in the construction and equipping of a simple but sturdy temporary facility. Such an approach was adopted by the U.S. Army Southern Command in Wiwili, Nicaragua, in the aftermath of Hurricane Mitch. In the case of Bam, Iran, the cost of rebuilding the entire primary and secondary health care facilities and teaching institutions was estimated by the government of Iran to be US$10.75 million, an amount very similar to that expended for the dispatch of field hospitals from the international community. Guidelines for the use of foreign field hospitals are available from WHO and PAHO (2003). In-Kind Donations. Unsolicited donations of inappropriate medical supplies not only are of limited use, but often cause serious logistic, economic, and political problems in the recipient country. Warehousing those supplies and, in many instances, building facilities (incinerators, for example) for the safe disposal of pharmaceutical donations diverts humanitarian funds from more effective uses. Recipient countries collectively share part of the responsibility by not clearly indicating what they do not want to receive and by not speaking out once inappropriate items arrived. Disease Prevention and Control. Postdisaster interventions in surveillance and control of communicable diseases should focus on strengthening existing programs. Benefits will outlive the crisis. Improvised mass immunizations (instead of improved sanitation and public awareness) and vector control by aerial spraying or fogging (instead of breeding-site reduction or waste disposal) are just two examples of wasteful managerial decisions. Shelters. Tent cities should be a last resort. Family-size tents may be expensive and do not last long. Establishing large settlements is easy, but such settlements are difficult to sustain and nearly impossible to terminate. They come with their own sanitation problems and social shortcomings (lack of privacy, loss of family identity, and loss of empowerment). Distributing construction material (or, preferably, cash subsidies) is more cost-effective and tailored to the needs and priorities of end users. Cash Assistance. Developed societies long ago abandoned the distribution of in-kind relief goods and services to their nationals in favor of direct financial assistance in the form of subsidies, grants, or tax relief. The individual is free to determine actual priorities and to seek the most cost-effective source of services (shelter, medical, food, or other). It is therefore surprising that external assistance from these same countries remains focused on the costly delivery of predetermined services or commodities. The most immediate lifesaving needs can be addressed only locally with existing resources and capacity. No cash contribution will meet those immediate needs. Beyond the acute phase, in many countries with market economies, most other services and goods are easily procured by those with financial means, suggesting that income availability is often the single limiting factor in rehabilitation. Undoubtedly, this approach would affect considerably the type (and number) of humanitarian actors by transferring Natural Disaster Mitigation and Relief | 1157 ©2006 The International Bank for Reconstruction and Development / The World Bank 199 power and decision making to the local beneficiaries and relying on local economic forces for delivery to the end user. It may also bring its own set of problems (and abuses), though perhaps that is a small cost, considering the economic and social benefits of the most interested party—the victim—being in charge. Cost-Effectiveness of Prevention and Mitigation The social benefits of making hospitals and water systems more resilient to the effects of natural hazards are recognized but too rarely applied. On the economic side, mitigation also increases the investment capacity in the health sector by preventing losses and the need for reconstruction (PAHO and UN/ISDR 1996; Bitrán 1996). The most compelling case for the cost-effectiveness of mitigation can be made during the planning phase for new installations, when costs of additional structural safety are minimal. Although the social benefits of prevention and risk management are more evident in the health sector than in others, further studies are needed to provide decision makers with quantified parameters of the economic benefits brought about by investment in risk management and disaster reduction. PAHO and UN/ISDR (1996) studies indicate that such increased investment fluctuates between 4 and 8 percent of a hospital’s local construction cost. When the value of services lost is added to the infrastructure loss, the additional investment is reduced to between 2 and 4 percent of direct and indirect losses observed. Even though this is a gross estimate that requires further research in other regions and types of health facilities, the figure is ratified by the estimated cost of reinforcement, which fluctuates but averages between US$2,000 and US$5,000 per bed, compared with the average cost of a new hospital bed of between US$100,000 and US$150,000 (at 1996 prices). Prevention of chemical and radiation accidents can be a highly cost-effective expense that is normally absorbed by the respective industries. Respect for existing norms in the use of radiotherapy and diagnostic equipment and, once such equipment is decommissioned, its proper disposal reduces DALYs from accidents at a modest cost. Mobilization of Resources Funding for preparedness and response programs follows rules and procedures that are distinct from those applicable to development projects. Most donors maintain a specific office or department for humanitarian affairs with a separate budget line. Procedures are also streamlined for quick response to unexpected situations. Processing a request takes a matter of days in emergencies and takes months for preparedness or mitigation projects, but it can take years in typical development projects negotiated with donors or financial institutions. From a ministry of health point of view, competition for disaster resources is with other sectors or humanitarian organizations, not within the sector (as it would be, for instance, with malaria or tuberculosis control projects). Funding for Preparedness. “By strengthening our public health planning for natural disasters and disease outbreaks, we will be in a better position to care for our populations, regardless of the type of hazard that confronts our health departments” (Rottman 2003, 1). This message, addressed to the public health community in the United States, is even more pertinent for developing countries. Most humanitarian offices in more developed countries allocate a modest but increasing proportion of their funds for predisaster capacity building. The capacity of the ministries of health to secure directly nonreimbursable funding depends on the following: • The existence of an established disaster program within the ministry, demonstrating a long-term commitment to health disaster preparedness. • An ongoing dialogue with local representatives of donors and their prior involvement in disaster-related activities or meetings of the health sector. • A realistic projection of concrete activities, taking into consideration the efforts of others, especially NGOs. One- or two-year training or capacity-building projects are more likely to be supported than those of longer duration that have recurrent costs or involve the purchase of equipment (radios, vehicles). • The technical endorsement and support of WHO and other UN agencies. A multisectoral preparedness component is also increasingly included in loans negotiated in the aftermath of disasters. Intended to strengthen the capacity of the civil protection agency, the funding is no substitute for local political commitment to assume recurrent expenses, the only guarantee of sustainability. Resources for Emergency Response. The amount of external resources available for response, financial or material, is influenced by the type of hazard, geopolitical considerations, and the number of deaths (rather than that of survivors in need of assistance). Funding is channeled mostly through humanitarian NGOs, the Red Cross system, or multilateral organizations, rather than through national governments. Consequently, the priority of the health authorities, rather than to seek direct contributions to the ministry, should be to ensure that health needs are properly identified and adequately covered by those agencies benefiting from the donations. Ministries of health often can obtain indirect financial support for their own activities through UN projects. 1158 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio ©2006 The International Bank for Reconstruction and Development / The World Bank 200 Concentrating on several key factors will improve the flow of external resources toward health priorities: within their ministries of health. Some lessons can be learned from this process: • Issuing a rapid and reliable assessment of what is needed and what is not needed for the emergency response, rather than waiting for a detailed assessment of the physical damage. • Focusing on tomorrow’s emergency health problems. External response is unable to address today’s short-lived problems. • Keeping a long-term view. Funding for emergency response is limited to a few months, whereas the health problems caused by the disaster will stay much longer. Projects should offer sustained benefits beyond their conclusion. • Recognizing shortcomings in governance when in contact with the many bilateral fact-finding or assessment missions coming to the disaster site. • The occurrence of a major disaster in the country or its neighbor is the initial catalyst for health authorities to recognize that disasters represent public health risks that must be addressed in an institutionalized manner. • Access to and support from the political level has determined the success or failure in coordinating the external and domestic health response. • A multihazard program covering the entire health sector is most effective. Assigning responsibility for coordination and management among different technical departments according to the type of hazard (chemical or natural, for instance) does not work. • A risk management program should cut across departments (medical care, epidemiology, water supply, sanitation, nutrition, and so forth) of the ministry of health and become sector wide. • The synergy between normal development, preparedness, and disaster response activities should be recognized. Poor development practices increase vulnerability, whereas preparedness improves the attention to daily health challenges. Programs narrowly focused on operational response have generally failed. Funding for Reconstruction. Funding for reconstruction is multisectoral and is often coordinated by an international financing institution (global or regional), together with a consortium of large donor countries. The health sector will compete with other social priorities and the “productive” sectors in an arena where the health burden (measured in DALYs) does not carry the same weight as economic factors. Success will depend on an exhaustive monetary valuation of the health damage, rapid formulation of projects, political support from the country’s highest authorities, and technical support and endorsement of specialized UN agencies and larger NGOs. Funding for Mitigation of Damage. Protecting the national capital investment of the health sector is primarily the responsibility of the country at risk. Development agencies or financial institutions may contribute only marginally to the actual cost of retrofitting installations or improving the design of new facilities. Modest funding for pilot or demonstration prevention programs may be available from both the humanitarian and the development sources of donor countries. Humanitarian offices may support promotion of the concept, development of guidelines or studies on vulnerabilities, and training. The health sector will benefit from close contacts with financial institutions, the ministry of foreign affairs, and other national ministries. Negotiations to ensure that new installations are able to withstand disasters must be initiated at the earliest opportunity, and the corresponding additional costs should be considered in the earliest stages of the project. IMPLEMENTATION OF CONTROL STRATEGIES: LESSONS OF EXPERIENCE AND CHALLENGES FACED All countries in Latin America and the Caribbean have established programs and structures for disaster risk management In Asia, the Asian Disaster Preparedness Center also has documented some interesting experiences (http://www.adpc. ait.ac.th/). THE RESEARCH AND DEVELOPMENT AGENDA Disasters in any one country are relatively infrequent. In addition to being a dangerous temptation for the authorities to postpone preventive actions, this infrequency is an impediment for research and institutional memory.On one hand,the humanitarian culture tends to raise ethical questions on the role of observers at a time when action at all costs is expected. On the other hand, few health academicians wish to embark on projects when control groups and time for advance planning are unavailable. Particularly encouraging are the increased numbers of publications and guidelines by UN organizations and NGOs and the trend toward organizing workshops on lessons learned a few months after a major disaster. These meetings of national experts and officials together with representatives from external actors are invaluable for identifying and sharing operational or institutional successes and failures for the collective benefit of other countries at risk. Epidemiological Research Most of the DALYs attributable to disasters occur immediately at the time of the disaster. Epidemiological research should, Natural Disaster Mitigation and Relief | 1159 ©2006 The International Bank for Reconstruction and Development / The World Bank 201 therefore, complement engineering studies to design better facilities and preparedness measures. After the initial disaster, basic questions need to be answered: How many secondary deaths and disabilities can actually be prevented by improving search and rescue and trauma care? How critical is the time factor in reducing DALY losses and assessing the effectiveness of foreign SAR and field hospitals teams? How can researchers objectively assess the risk of outbreak following disasters? In particular, how can they better differentiate between cases attributable to increased transmission and those resulting from improved surveillance and medical attention provided to the victims? What is needed are data to put to rest unquestioned assumptions and clichés. The alternative is to continue to divert scarce resources away from routine disease control programs and toward costly measures of doubtful effectiveness. Strategic Research Research is required that will compare the effectiveness of preparedness and response strategies and approaches: • With respect to preparedness, how should researchers assess the effectiveness of training and coordination versus that of investing in hardware and stockpiles? For instance, will the accreditation of hospitals based on their safety and readiness improve their disaster performance? • With respect to mitigation, how should limited funding for retrofitting health facilities be allocated? Is nonstructural mitigation a workable alternative in the absence of structural measures? • With respect to response, what is the effect of international assistance in terms of reductions in DALY losses that could not be achieved locally? Is it contributing to strengthening the capacity of the developing countries? What type of humanitarian assistance has proven to be development friendly? • Finally, how should researchers measure the effectiveness of preparedness or mitigation given the unpredictability of disasters? Economic Research Humanitarian response is resistant to concepts of costeffectiveness. Economists should contribute to the comparative study of the immediate and long-term effects of external interventions versus less costly alternatives such as relying on local resources and building local capacity. A cost-benefit analysis of international medical interventions prior to and during a disaster situation is also overdue. Economic assessment of the damage to the health sector remains focused on physical losses and fails to sufficiently consider the broader burden on a society caused by the loss of health services over a sustained period. Refining the existing methodology and developing quantitative indicators to estimate those indirect costs should be a research priority. CONCLUSIONS Natural hazards are not likely to decrease in the foreseeable future. Though geological events may occur independently of any human control, available data suggest that mankind plays a role in global climate. Technological hazards may also increase rapidly as a result of the unregulated development of industries in most countries and possibly the use of weapons-grade hazardous substances against civilian populations. An increase in the number of hazards should not mean that the resulting health burden will also increase. A sustained effort is needed to minimize risk, both by reducing vulnerability through prevention and mitigation and by increasing capacity through preparedness measures. A Strategic Approach The prime objective of a developing country is to develop. Emergencies and disasters have proven to be major obstacles and setbacks in the path toward sustainable development. Conversely, the shortcomings in development programs and institutions reduce the effectiveness of the health response in times of crisis. Development and disaster risk management cannot be addressed separately. Reducing risk is not a luxury reserved for more developed societies; it is a necessity in countries with fragile economies and health systems. It is clearly a public health priority. Disasters, as any other public health problem, need to be addressed on a long-term and institutionalized basis through the establishment in the ministry of health of a program or department for prevention, mitigation, preparedness, and response for all types of disasters. Trends in Latin America suggest that such an approach in the context of sustainable development contributes to narrowing the gap in disaster-related deaths and disabilities (as measured by DALYs) between industrial and developing countries. Disaster risk reduction is not merely a health issue. The economic and political dimensions should not, however, be allowed to overshadow the fundamental fact that disasters are, above all, human tragedies incompatible with the definition of health adopted by the WHO constitution. On one hand, the health sector should adapt and use the methodology of economic valuation of disaster impact as developed by ECLAC; on the other hand, the financial world should also learn to give equal consideration to the health burden (DALYs) in its decision making for development or reconstruction. For this to take place, health and humanitarian actors need to dramatically improve the availability of data. 1160 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio ©2006 The International Bank for Reconstruction and Development / The World Bank 202 Disaster risk reduction is not the exclusive domain of a few experts or officials. It is the collective responsibility of all disciplines and programs in the health sector, as well as a remarkable tool or gateway for collaboration with other sectors. Alone, the ministry of health cannot reduce the health burden or play its coordinating role in the response. Disaster risk reduction is unlikely to produce immediate results. It requires sustained commitment over the years. Learning from Errors Learning from past disasters is difficult. At a national level, the relatively long periods between major disasters result in few decision makers having prior disaster management experience. At an international level, the frequent turnover of relief workers ensures that many of the actors are relatively inexperienced and susceptible to adopting myths and clichés, which are rarely challenged by the media and the academic world. It is time for an international initiative to identify the best practices, and it is time for affected countries and scientists to point out the inadequacies of responses. Humanitarian health interventions, as any other health intervention, should be subject to cost-benefit reviews that compare their benefits in terms of DALY loss reduction to other alternatives, including a possible shift of international emphasis from immediate medical response to preparedness or rehabilitation projects. Local health services are best situated to address the health consequences of disasters. They should be better prepared to do so. A formalized mechanism to transmit and share those lessons learned from past errors and to build the response capacity is required in the health sector. Finally, the greatest potential for saving lives is in reducing the risks and the vulnerability through better infrastructure, land-use management, public awareness, and training. The challenge in risk reduction is to sustain public support and political will in periods of calm. International organizations—WHO in particular—have a unique and critical role to play as advocates for a long-term approach to disaster risk management in the context of sustainable development. ACKNOWLEDGMENTS The field of disaster epidemiology, a concept first introduced in the early 1970s by M. F. Lechat of the University of Louvain in Belgium, is now calling on many disciplines and fields of knowledge. The authors express their gratitude to Caridad Borras for her contribution on radiological disasters and to Jean-Luc Poncelet, Karl Western, Guy Arcuri, Steve Devriendt, and Roberto Jovel for their advice, comments, and suggestions. This chapter relies heavily on the successes and failures of the health sectors in Latin America and the Caribbean, a region where a sustained effort over 25 years, with the support of PAHO, WHO, and donor countries, traced the way to the reduction of risks from extreme events. This chapter owes greatly to a large number of experts and professionals in the health sector of those countries. NOTES 1. In a nonnatural phenomenon, such as the attacks in New York on September 11, 2001, a similar risk has been detected and is perceived as a remnant potential long-term health risk similar to the effect of air contamination from ash from volcanoes. 2. Evidently these direct losses are not easy to determine in longdeveloping events (such as the ones associated with slow processes or climatic variability), because over time there will be overlapping damage, in contrast to the damage that occurs in sudden events such as hurricanes or earthquakes. 3. Data came from reports of the UN Office of Coordinator for Humanitarian Affairs (http://www.reliefweb.org), supplemented by authors’ estimated costs for donors that did not report actual costs. REFERENCES Alexander, D. 1997. “The Study of Natural Disasters 1977–1997: Some Reflections on a Changing Field of Knowledge.” Disasters 21 (4): 284–304. Armenian, H. K., A. K. Melkonian, and A. P. Hovanesian. 1998. “LongTerm Mortality and Morbidity Related to Degree of Damage Following the Earthquake in Armenia.” American Journal of Epidemiology 148 (11): 1077–84. Armenian, H. K., A. Melkonian, E. Noji, and A. P. Hovanesian. 1997. “Deaths and Injuries Due to the Earthquake in Armenia: A Cohort Approach.” International Journal of Epidemiology 26 (4): 806–13. Bernstein, R. S., P. J. Baxter, and A. S. Buist. 1986. “Introduction to the Epidemiological Aspects of Explosive Volcanism.” American Journal of Public Health 76 (Suppl.): 3–9. Bitrán, D. 1996. “Impacto económico de los desastres naturales en la infraestructura de salud.” Paper presented to the PAHO International Conference on Disaster Mitigation in Health Facilities, Mexico City, January. Document LC/MEX/L.2911. ECLAC, Mexico City. Bouma, M. J., and C. Dye. 1997.“Cycles of Malaria Associated with El Niño in Venezuela.” Journal of the American Medical Association 278 (21): 1772–74. Buist, A. S., and R. S. Bernstein. 1986. “Health Effects of Volcanoes: An Approach to Evaluating the Health Effects of an Environmental Hazard.” American Journal of Public Health 76 (Suppl.):1–2. Chang, S. E., and C. Chamberlin. 2004. “Assessing the Role of Lifeline Systems in Community Disaster Resilience.” In Research Progress and Accomplishments 2003–2004, Multidisciplinary Center for Earthquake Engineering Research. Buffalo, NY: 87–94. de Ville de Goyet, C. 2001. “Earthquakes in El Salvador.” Revista Panamericana de Salud Publica 9 (2): 107–13. de Ville de Goyet, C., E. Acosta, P. Sabbat, and E. Pluut. 1996. “SUMA: A Management Tool for Post-Disaster Relief Supplies.” World Health Statistics Quarterly 49 (3–4): 189–94. ECLAC (Economic Commission for Latin America and the Caribbean). 1998. Análisis Costo-Efectividad en la Mitigación de Daños de Desastres Naturales sobre Infraestructura Social. Document LC/MEX/R.643. Mexico City: ECLAC. Natural Disaster Mitigation and Relief | 1161 ©2006 The International Bank for Reconstruction and Development / The World Bank 203 ———. 2003. Handbook for Estimating the Socio-economic and Environmental Effects of Disaster. LC/MEX/G.5. Mexico City: ECLAC. ———. 2002b “Terremotos en El Salvador 2001.” In Crónicas de Desastres No. 11. Washington, DC: PAHO. FEMICA (Federación de Municipios del Itsmo Centro Americano). 2003. “Acueducto Orosi: Una experiencia regional sobre implementación de medidas de prevención y mitigación,” Paper presented at the 10th meeting of the Network for Decentralization and Municipal Development on the topic Local Risk Management: A Challenge for the Development of Municipalities in Central America. Antigua, Guatemala, October 21–24. PAHO and UN/ISDR (Pan American Health Organization and United Nations/International Strategy for Disaster Reduction). 1996. Lecciones aprendidas en América Latina de mitigación de desastres en instalaciones de salud: Aspectos de costo-efectividad. Washington, DC: PAHO. Howard, M. J., J. C. Brillman, and F. M. Burkle. 1996. “Infectious Disease Emergencies in Disasters.” Emergency Medicine Clinics in North America 14 (2): 413–28. IDB (Inter-American Development Bank). 1999. “Hurricane Mitch: Women’s Needs and Contributions.” Document SOC-115, IDB, Washington, DC. IFRC (International Federation of Red Cross and Red Crescent Societies). 2000. World Disasters Report 2000: Focus on Public Health. Geneva: IFRC. ———. 2003. World Disasters Report 2003: Focus on Ethics in Aid. Geneva: IFRC. Kirigia, J. M., L. G. Sambo, W. Aldis, and G. Mwabu. 2004. “Impact of Disaster-Related Mortality on Gross Domestic Product in the WHO African Region.” BMC Emergency Medicine 4 (1): 1. http://www. biomedicalcentral.com/1471-227x/4/1. Malilay, J., M. G. Real, A. Ramirez Vanegas, E. Noji, and T. Sinks. 1996. “Public Health Surveillance after a Volcanic Eruption: Lessons from Cerro Negro, Nicaragua, 1992.” Bulletin of the Pan American Health Organization 30 (3): 218–25. Morgan, O. 2004. “Infectious Disease Risks from Dead Bodies Following Natural Disasters.” Revista Panamericana de Salud Publica 15 (5): 307–12. National Research Council. 1999. From Monsoons to Microbes: Understanding the Ocean’s Role in Human Health. Washington, DC: National Academy Press. PAHO (Pan American Health Organization). 1982. Disaster Reports Number 2: Jamaica, St. Vincent, and Dominica. Washington, DC: PAHO. ———. 1985. “Earthquake in Mexico.” In Disaster Chronicles No. 3. Washington, DC: PAHO. ———. 1998. “El Niño and Its Impact on Health.” Official document CE/122/10. PAHO, Washington, DC. ———. 1999. “Huracanes Georges y Mitch, 1998.” In Cronicas de Desastres No. 7. Washington, DC: PAHO. PAHO and WHO (Pan American Health Organization and World Health Organization). 2000. “Mitigación de desastres en instalaciones de salud: Aspectos estructurales.” In Mitigación de desastres en instalaciones de salud—Material técnico y de capacitacion (CD-ROM). Washington, DC: PAHO and WHO. PAHO (Pan American Health Organization), WHO (World Health Organization), World Bank, and ProVention Consortium. 2004. Guidelines for Vulnerability Reduction in the Design of New Health Facilities. Washington, DC: PAHO and World Bank. Poncelet, J.-L. 1997. “Earthquakes in Latin America: The Role of Cities in Disaster Management.” In Earthquakes and People’s Health: Proceeding of a WHO Symposium, Kobe, Japan, January 27–30. Geneva: WHO. http://www.helid.desastres.net. Rottman, S. 2003. “Director’s Message.” UCLA Center for Public Health and Disasters Newsletter 9 (Fall): 1. UNHCR (United Nations High Commissioner for Refugees). 1998. Handbook for Emergency Situations. Geneva: UNHCR. UN/ISDR (United Nations/International Strategy for Disaster Reduction). 2004. Living with Risk: A Global Review of Disaster Reduction Initiatives. Vol 1. Geneva: UN/ISDR. WHO (World Health Organization). 1992. Assessing the Health Consequences of Major Chemical Incidents—Epidemiological Approaches. Geneva: WHO. ———. 1994. Health Laboratory Facilities in Emergencies and Disaster Situations. Geneva: WHO. http://www.helid.desastres.net. ———. 1996. Health Consequences of the Chernobyl Accident: Results of the IPHECA Pilot Projects and Related National Programmes. Geneva: WHO. ———. 2002. Reducing Risks, Promoting Healthy Life: World Health Report 2002. Geneva: WHO. WHO and PAHO (World Health Organization and Pan American Health Organization). 2003. Guidelines for the Use of Foreign Field Hospitals Following Sudden-Impact Natural Disasters. Washington, DC: WHO and PAHO. http://www.paho.org/English/DD/PED/FieldH.htm. ———. 2002a. “Protección de la salud mental en situaciones de desastres y emergencias.” In Serie de Manuales y Guías sobre Desastres No. 1. Washington, DC: PAHO. 1162 | Disease Control Priorities in Developing Countries | Claude de Ville de Goyet, Ricardo Zapata Marti, and Claudio Osorio ©2006 The International Bank for Reconstruction and Development / The World Bank 204 Chapter 62 Control and Eradication Mark Miller, Scott Barrett, and D. A. Henderson The Controversy: Control or Eradication? We cannot refrain altogether from examining the roots of this controversy if only because the extreme views for and against eradication have exerted and are still exerting a . . . highly detrimental influence on public health practice. —P. Yekutiel, Eradication of Infectious Diseases: A Critical Study Eradication of an infectious disease is an extraordinary goal. Its possibility became apparent as soon as Edward Jenner demonstrated an ability to provide immunity to smallpox. Writing in 1801, Jenner observed that, through broad application of vaccination, “it now becomes too manifest to admit of controversy that the annihilation of the Small Pox, the most dreadful scourge of the human species, must be the result of this practice” (Jenner 1801). Louis Pasteur claimed that it was “within the power of man to eradicate infection from the earth” (Dubos and Dubos 1953). And yet, by and large, public health has proceeded with more modest goals of local and regional disease control. Notable successes have occurred. Indeed, some diseases now thought of as “tropical” were previously endemic in temperate climates. Systematic application of hygiene, sanitation, environmental modification, vector control, and vaccines have led, in many countries, to the interruption of transmission of microbes causing such diseases as cholera, malaria, and yellow fever. Intensive efforts to eliminate breeding sites of the yellow fever mosquito vector, Aedes aegypti, interrupted transmission of this disease in Havana in 1901 and throughout Cuba soon thereafter. Subsequently, yellow fever and malaria were able to be controlled in Panama, thus permitting construction of the Panama Canal. In 1915, the Rockefeller Foundation launched an effort to eradicate the disease worldwide. Transmission appeared to have ceased in the Americas by 1928, but then cases reappeared, and by 1932, it became clear that a nonhuman endemic focus was serving to reinfect areas otherwise free of yellow fever. In the 1930s, F. L. Soper set out to eradicate the Aedes aegypti vector from the Americas. By 1961, Soper reported that he had largely succeeded except for the United States, where the program received little support. By the 1980s, Aedes aegypti had become reestablished in Central and South America. In 1953, Brock Chisholm, the first director-general of the World Health Organization (WHO), tried to persuade the World Health Assembly (WHA) to undertake smallpox eradication, but a number of countries objected on the grounds that eradication was not technically feasible. Instead, in 1955, under the leadership of his successor, Marcolino Candau, WHO began a global effort to eradicate malaria primarily by means of household spraying of DDT. The relatively sophisticated science of malaria control was abandoned in favor of this simplistic technology (Jeffrey 1976). Despite an expenditure of more than US$2 billion, the effort failed. Even while the malaria eradication effort was under way, the Soviet Union, in 1958, proposed to the WHA that smallpox be eradicated. A resolution to this effect was offered in 1959 and passed unanimously. However, the resolution provided little international funding or support. Over the next seven years, disease transmission was interrupted in some 30 countries in Africa, Asia, and South America, but endemic smallpox persisted in the Indian subcontinent, Indonesia, most of SubSaharan Africa, and Brazil. WHO launched an intensified effort in 1967 to eradicate the disease within a decade. This new 1163 ©2006 The International Bank for Reconstruction and Development / The World Bank 205 resolution included an annual budget of US$2.4 million, to be paid according to the WHO scale of assessments. The resolution passed by the narrowest of margins, but a reinvigorated effort was soon under way and paved the way for a historic public health achievement (Henderson 1988). Following an extraordinary worldwide effort, the last case of smallpox was isolated in October 1977, and the disease was certified as being eradicated in 1979, 170 years after Edward Jenner first dreamed of that possibility. Understanding how and why smallpox eradication succeeded is essential to the study of control and eradication. The smallpox success was inspirational, even though the leaders of WHO’s smallpox eradication effort cautioned that, among all the diseases that might be considered candidates for eradication, smallpox was unique (Fenner and others 1988) and that they foresaw no other disease as a candidate for eradication (Henderson 1982). At a meeting convened by the Fogarty International Center of the National Institutes of Health in 1980, scientists, public health officials, and policy makers discussed the merits of eradicating other diseases, with schistosomiasis, dracunculiasis, poliomyelitis, and measles identified as possible candidates (Henderson 1998a). However, no consensus was reached at that time on moving forward with any of those diseases. Poliomyelitis became the next principal target when mass vaccination campaigns, proposed by Albert Sabin (1991), proved remarkably successfully in Cuba and Brazil. In 1985, an American Health Organization coordinated campaign was launched to interrupt poliovirus transmission in the Americas by 1991, and this effort succeeded. Some believed that global eradication might be possible, although others were concerned that the far less developed infrastructure of health, transportation, and communications services in many parts of Asia and Africa would make it an unachievable task. In 1988, the WHA adopted a resolution to eradicate polio, but at that time, a longer-term strategy for ending polio vaccination was neither formulated nor agreed on by the public health and scientific community. The WHA has adopted only one other resolution to eradicate a disease—guinea worm, or dracunculiasis. The eradication of this disease can be achieved by applying simple technologies for providing water that is free of the vector copepod and parasite and for treatment of patients with the disease. This eradication program has made steady progress but has been hampered in part by civil and political unrest and lack of program priority because of low mortality and low incidence in some remaining endemic areas. However, given the environmental restriction of the parasite to rural tropical areas and its relatively low transmissibility, eventual global eradication seems within reach. One other case—that of measles—is worth noting. A number of public health authorities have raised the possibility of eradicating that disease. In the Americas, spurred on by the success of regional cessation of transmission of wild poliovirus, eventual consensus was reached to intensify measles control efforts, primarily through surveillance and periodic pulse application of measles vaccine in national campaigns. As a consequence, transmission of measles virus was temporarily interrupted in the Americas on several occasions but reestablished again by importations (CDC 1998a). Although the U.S. Centers for Disease Control and Prevention (CDC) and WHO have advocated extending measles “elimination” through vaccination campaigns and second-dose opportunities to other regions (Biellik and others 2002; CDC 1998a, 1998b, 1999a, 1999b, 2003d, 2004b, 2004d, 2004f), the intensive control efforts required to break transmission of this highly infectious agent make global eradication unlikely at this time. DEFINITIONS Yekutiel (1980, 5–8) provides an excellent treatise on the concept of eradication, which includes a summary of the multiple definitions that have been formulated (Andrews and Langmuir 1963; Cockburn 1961, 1963; Payne 1963a, 1963b; Spînu and Biberi-Moroianu 1969). A conference devoted to eradication held in Dahlem, Germany, in 1997 (Dowdle and Hopkins 1998) set out to provide precise definitions for control, elimination, eradication, and extinction in a biological, economic, and political context (Dowdle 1998, 1999; Ottesen and others 1998); however, a number of eminent public health officials (Cochi and others 1998; de Quadros 2001; Goodman and others 1998b; Henderson 1998b; Salisbury 1998) challenged these definitions at two subsequent meetings at the CDC (Goodman and others 1998a, 1998b) and the U.S. Institute of Medicine (Knobler, Lederberg, and Pray 2001). Unfortunately, broadly accepted, standard definitions for key concepts pertaining to disease control and eradication do not exist in the literature. Making matters more confusing, certain of the concepts have been given names that are part of our everyday language and so are easily misinterpreted by nonspecialists as meaning something different from the meanings understood by those who are preoccupied with eradication programs. Most unfortunate is the all too casual use of the words elimination and eradication to promote programs that cannot reasonably be expected to achieve the promise implicit in these words. Moreover, the two words themselves are commonly used interchangeably. Control Two concepts are central to this chapter: control and eradication. By control, we mean a public policy intervention that restricts the circulation of an infectious agent beyond the level that would result from spontaneous, individual behaviors to protect against infection (Barrett 2004). 1164 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson ©2006 The International Bank for Reconstruction and Development / The World Bank 206 Although control is a range rather then a level, a particular level of control may be an aim of policy. Because every choice entails consequences, choice of the “optimal” level of control requires economic analysis. Optimal here is defined in relation to the model that gives rise to the result. Control is local and so needs to be looked at from the local perspective. Because one country’s (or region’s) control may affect other countries (regions), a global perspective exists as well. The level of control that is optimal for one country (region) may not be optimal from the perspective of the world as a whole. Thus, a need exists to distinguish between, say, a locally optimal level of control and one that is globally optimal. Finally, control requires ongoing intervention. Sustaining a given level of control requires an annual expenditure. Eradication Eradication differs from control in that it is global. The term denotes the certified total absence of human cases, the absence of a reservoir for the organism in nature, and absolute containment of any infectious source. Eradication permits control interventions to stop or at least to be curtailed significantly. Finally, eradication is binary. Control levels can vary, but a disease is either certified as eradicated or not. Every disease can be controlled, even if only by using simple measures, such as quarantine. The ultimate achievement of control is eradication. But not every disease that can be controlled can be eradicated. Very few diseases, in fact, are potential candidates for eradication. The criteria for the feasibility for eradication as a preference over control are discussed in the section titled “Economic Considerations.” Elimination Control and eradication are the essential concepts, but two other terms bear mention. The first is elimination. Some who are concerned with eradication programs have explicitly defined this term to denote the cessation of transmission of an organism throughout a country or region. In contrast, eradication is defined as a global achievement. Like control, elimination is location-specific and would require ongoing interventions to be sustained in order to prevent reemergence of the disease from microbe importations. Two problems exist with the term elimination. First, it has been used to describe different phenomena, not just that described in the definition given above. For example, some public health officials have promoted programs aimed at “eliminating a disease as a public health threat,” which is interpreted to mean reducing incidence to an “acceptable” level but not necessarily to zero. This usage is very different from the one outlined above and is almost certain to be misunderstood. Second, the definition of the word elimination in common use, as applied to disease control, is indistinguishable from eradica- tion. The 1993 edition of the New Shorter Oxford English Dictionary, for example, defines eliminate as to “remove, get rid of, do away with, cause to exist no longer.” This same dictionary defines eradicate as “pull up or out by the roots, uproot, remove or destroy completely, extirpate, get rid of.” This ambiguity invites misunderstanding among those not intimately involved in an eradication effort. For purposes of clarity, we seldom use the term elimination in this chapter and then only to signify control measures sufficient to interrupt microbe transmission in a specified area. Extinction Finally, the literature sometimes refers to extinction as a possible policy goal. In the context of infectious disease control, the concept is problematic for two reasons. First, proving that an organism has become extinct is impossible. To do so would require demonstrating not only that the organism no longer exists in nature but also that it no longer exists in any controlled environment—a practical impossibility. Second, de novo synthesis of viral agents from published genomes (Cello, Paul, and Wimmer 2002) now put the concept in peril, although much research remains to be done in this area. Extinction, in the context of infectious diseases, may no longer be irreversible. Clearly, policy making will be improved by stating the goal of any particular intervention in precise language. FRAMEWORKS FOR ERADICATION Numerous issues need to be considered in planning expanded control measures that lead, possibly, to regional cessation of transmission or global eradication of disease. These complex issues will be further examined in the chapter. Scientific Considerations Scientific considerations include the nature of potential reservoirs for disease-causing microbes or their vectors, technologies available for interrupting disease transmission, changes in host capabilities to deter infections and disease, and satisfactory containment of organisms in laboratories. Geographic and Environmental Controls. The limit of endemicity for microbes and their associated diseases is determined in part by their ability to exist in nature outside the human host. Both geographic and temporal variations determine the ecological niche of microbes, resulting in variable annual incidence rates throughout the world. This niche limitation is further extended to intermediary vectors and hosts in complex biological systems. Natural environmental barriers also may isolate the habitats of helminths. Infectious agents that are not limited to an environmentally restricted intermediary Control and Eradication | 1165 ©2006 The International Bank for Reconstruction and Development / The World Bank 207 host or those that have longer latent periods, thereby allowing translocation, may have a global pattern of distribution. Examples include the highly transmissible viral agents such as measles, rubella, influenza, and varicella. Although these agents are not geographically constrained, their transmission patterns are directly and indirectly influenced by seasonal environmental factors and population-based immunity. Potential Reservoirs. A microbe and associated disease can not be eradicated if the microbe is capable of persisting and multiplying in a reservoir. Microbes that thrive in nonhuman species may reemerge if control efforts cease, thus leaving human populations susceptible. Similarly, if the infectiousness of a human is long lived or could lead to potential recrudescence, surveillance efforts would have to continue as long as the last individual remained potentially capable of transmitting infection, as would be the case with tuberculosis or hepatitis B infection. Transmissibility. The inherent rate of a microbe’s ability to cause secondary infections is defined by an organism’s reproductive rate in a fully susceptible (R0) and partially susceptible (R) population. The reproductive rate of organisms that infect individuals only once because of durable immunity is inversely proportional to the average age of infection in an endemic area. Agents that cause childhood infections, such as viral respiratory agents, are far more transmissible than helminths and subsequently require more intensive control efforts to interrupt transmission. Natural Resistance to Reinfection. Many natural infections induce long-lived immunity to reinfection. Although the most commonly used vaccines have been available for fewer than 50 years—less than the lifetime of an individual—they, too, are assumed to offer long-lasting immunity. Because eradication depends on reducing susceptible populations in potentially endemic areas, long-lived protection through immunization or natural disease is important to successful programs. Laboratory Containment. Laboratory specimens containing the organism targeted for eradication could serve as reservoirs. Considerable effort may be necessary to ensure their maximum security. That these microbes may be inconspicuous in specimens collected for other purposes poses special challenges. This situation is especially true for the poliomyelitis virus, which may be found in many stool specimens collected for studies completely unrelated to current poliomyelitis eradication efforts. Operational Considerations Optimization of control requires a fundamental appreciation of the biological systems that govern the ecology of microbes and their intermediary and human hosts. The reproductive rate, R, is influenced by many local factors, including population density (of vectors, intermediary hosts, and humans) and other environmentally determined conditions, all highly variable throughout the world. For a disease to be controlled to stop transmission, the intervention-altered reproductive rate must be maintained below 1.0. At the same time, all reservoirs of the responsible microbe must be controlled. Three main components of possible eradication programs are • surveillance, including environmental sampling where appropriate and clinical testing • interventions, including vaccination and chemotherapy or chemoprophylaxis or both • environmental controls and certification of eradication. Each of these components must be undertaken at local, community, national, regional, and global levels. Eradication differs from control in that it is expected to be permanent. Success depends on having adequate surveillance to identify potentially infectious persons and on stopping transmission before infection of a new cohort of susceptible persons arises as a result of births, migration, or the waning effectiveness of prophylactic measures. Disease Surveillance. Effective surveillance requires a sensitive system to detect the presence of microbes within the environment, intermediary hosts, and clinical cases. Surveillance and response systems need to be more efficient than the rate of transmission of the targeted agent. As eradication progresses, the sensitivity of detection systems must be steadily enhanced to detect all existing foci. Nonclinical or latent infections pose formidable barriers to eradication efforts. Operationally, the need for nearperfect sensitivity comes at the expense of lower specificity. Thousands of skin lesions from suspected smallpox patients were tested in reference laboratories during confirmation of smallpox eradication, and tens of thousands of stool specimens are being examined for poliovirus. Highly sensitive systems used to detect measles cases in the Americas began to identify a greater proportion of rubella and parvovirus infections because of the nonspecific surveillance of rash illness. Such findings are important because the identification of other diseases that mimic the targeted disease can lead to a misdirection of resources. However, the ability to detect such similar clinical cases can serve as a proxy measure for the adequacy of surveillance. For example, identification of a minimum incidence of cases of acute flaccid paralysis that is not related to polio has served as an indicator of adequate efforts of case finding for polio. Interventions. Interventions to block transmission of the targeted infectious agent should be easy to deploy and adaptable 1166 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson ©2006 The International Bank for Reconstruction and Development / The World Bank 208 to diverse conditions, given the global goal of eradication. Cost considerations and local acceptance of the required sacrifices (both short and long term) are crucial for success. Interventions may be designed for environmental control of microbes, isolation (quarantine) of clinically infectious individuals to limit their contacts with susceptible persons, treatment of clinical cases to limit the duration of infectiousness, or reduction in the infected pool of individuals through immunoor chemoprophylaxis. Certification. The last tool for eradication is a certification process whereby independent, respected parties certify the absence of disease transmission or the existence of any specific microbe in an uncontrolled reservoir, including laboratories (Breman and Arita 1980). Although certification can be implemented on a regional basis, it must ultimately be done globally. Certification is one of the greatest challenges in any eradication effort, given the exceedingly great difficulty of verifying a negative finding in a reasonably short period of time. When certification is completed, curtailment of control measures should be possible. Strengthening control efforts sufficiently to achieve eradication is a difficult and expensive task. It requires that scaling up of such efforts occur over a wide area—at the community, national, regional, and global levels. Its efficacy depends heavily on the adequacy of local financial and human resources, as well as on a broad range of logistical factors. Economic Considerations Control and eradication programs have many economic dimensions: private versus social net benefits, short-term versus long-term net benefits, and local versus international net benefits. Such interventions also have implications for existing public health programs. Private versus Social Net Benefits. Individuals have private incentives to protect themselves from disease—by means of vaccination, for example. But when individuals protect themselves—when they elect to be vaccinated—they offer a measure of protection to others by helping limit the spread of infection. In brief, the social benefit of vaccination is greater than the private benefit alone. As more people become vaccinated, the marginal private and social benefit of vaccination— that is, the benefit of vaccinating an additional susceptible person—declines. The marginal private benefit is likely to fall because, as more people are vaccinated, the probability of a susceptible person becoming infected falls. The marginal social benefit is likely to fall for the same reason and for one other: as more people become protected, the total number of susceptible persons falls. The marginal social benefit of vaccination falls sharply at the critical level of immunization—the level at which herd immunity is conferred on all susceptible persons. When a population is immunized to this level, a disease ceases to be endemic, and imported infections cannot spark an epidemic. This level is determined by the epidemiology of a disease, but whether it pays to vaccinate to this level depends on the economics, and the economics depend in turn on the social costs and not only the social benefits of vaccination. These costs consist of the direct costs of producing, distributing, and administering a vaccine. The economics depend also on the costs borne by the individuals who are vaccinated, such as those incurred by individuals who experience vaccine complications. The proportionate costs of reaching people who live in remote areas and those who are at special risk, such as migrants and the homeless, increase as the fraction of the population vaccinated increases. The economics of varying levels of disease control depend on the relationship between the marginal social benefits and the marginal social costs of vaccination. As vaccination levels increase, the marginal social benefits of vaccination fall, whereas the marginal social costs rise. Social welfare is maximized where these two relations intersect, which might be called the “optimal” level of vaccination—a level that may or may not achieve cessation of transmission or eradication. Short-Term versus Long-Term Net Benefits. Control programs require ongoing intervention. Sustaining a given level of protection requires that, over time, a certain proportion of new susceptible persons be vaccinated. Eradication differs from control in being permanent. The economics of eradication must therefore take account of long-term benefits as well as short-term costs. The long-term benefits of eradication consist of avoided future infections and vaccination costs—a dividend. To calculate this benefit, one projects future infection and vaccination levels in the absence of eradication, attaches values to these, and discounts them. If this sum exceeds the costs of eradication, then eradication enhances social well-being, and it therefore should be undertaken. In deciding on the benefits of eradication, the cost of future infections and vaccination should ideally be compared with the best alternative to eradication: the level of optimal control (Barrett and Hoel 2003). The costs of eradication must also take into account ongoing surveillance requirements, laboratory containment, and perhaps the maintenance of stockpiles of vaccine in the chance event of disease reemergence. From an economic perspective, attractive candidates for eradication are those diseases that some countries have themselves targeted for interruption of transmission nationally or regionally. Local versus International Net Benefits. Control differs from eradication in another important way. Control refers to Control and Eradication | 1167 ©2006 The International Bank for Reconstruction and Development / The World Bank 209 location-specific interventions. Eradication, by contrast, is global. In economic terms, eradication is a global public good. No country can be excluded from the benefit of eradication, and no country’s consumption of that benefit diminishes the amounts available to other countries. Control, by contrast, supplies only a local public good. Eradication requires a global effort. A disease can be eradicated only if microbe transmission ceases everywhere. This spatial dimension to eradication is of fundamental importance because no world government can implement an eradication policy; the WHA can declare its support for eradication, but WHO does not have the power to enforce the execution of a national program in support of that goal. The outcome experienced by any country depends not only on whether the country itself eliminates the disease within its borders but also on whether all other countries do so. Indeed, eradication is a weakest-link public good. Whether eradication is achieved depends on the level of control adopted by the country that undertakes the least control. In practical terms, any country in which disease is endemic can prevent eradication from being achieved. In 2004, the global polio eradication initiative, after investing more than US$3 billion and involving some 20 million volunteers over a period of 16 years, was placed at risk of failure by the actions of one local administration. In the Kano state of Nigeria, local leaders claimed that the polio vaccine was tainted with the AIDS virus and sterility drugs and declined to participate in a national immunization day program. The European Union then declined to pay for the national program in Nigeria, believing the money would be wasted (Roberts 2004). One consequence was the subsequent spread of polio to nine formerly polio-free countries. Concerted efforts by WHO later persuaded local leaders in Nigeria to rejoin global efforts, but special vaccination programs had to be launched over a population area of more than 300 million persons. This situation dramatically illustrated the vulnerabilities inherent in a weakest-link public good. What are the incentives for states to participate in an eradication effort? To begin, assume that countries are symmetric, meaning that all countries have the same benefits and costs of control. Assume as well that eradication is feasible. Four possible situations then exist (Barrett 2003): • First, the global net benefit of eradication may be negative— the cumulative programmatic costs outweigh the net present value of the cumulative benefits. In this case, elimination would also yield a negative net benefit to every country, and so no country would eliminate the disease. • Second, the global net benefit of eradication may be so large that each country would choose to eliminate the disease even if others did not. In this case, all countries would eliminate the disease, and the disease would therefore be eradicated. In these two cases, no need exists for an international policy. • Third, each country may have an incentive to eliminate a disease only if all other countries have eliminated it. In this case, achieving global eradication requires coordination. Here a role exists for international policy, but all that is required is for each country to be assured that all others will eliminate the disease. • Finally, and noting that the “last” country to eliminate a disease would get just a fraction of the global dividend from eradication, under some circumstances no incentive may exists for this country to eliminate the disease—even if all other countries have done so and even if the entire world would be better off if it did. This case is the most worrisome, because implementation of the efficient outcome would likely require enforcement. All this hypothesizing assumes that countries are symmetric, and of course they are not. Some countries gain less from control and would gain less from eradication than others. Some are unable to implement an elimination program, even if they would very much like it to succeed. In these situations, achieving an eradication goal will require international financing and technical assistance, with the countries that benefit most from eradication compensating the other countries for the costs of eradicating the disease. National and international reproach are often expressed if a country lags in its eradication efforts. International financing has been a key element in all eradication programs. We have thus far looked at eradication from the perspective of only the self-interests of states. But eradication also has implications for development. In particular, eradication has two advantages over control programs. The first is that the rich countries may gain directly if the goal is achieved, giving them a vested interest in ensuring that the goal is achieved. The second is that eradication is permanent, making an investment in eradication financially sustainable. This second advantage is important because financial sustainability has proved to be a key problem for disease control programs in developing countries (Kremer and Miguel 2004). Vertical versus Horizontal Programs. Control and eradication programs cannot be viewed in isolation. All programs have implications for the delivery of comprehensive primary care services. An important question is whether targeted, or socalled vertical, programs draw critical resources away from other health care programs or whether they serve instead to augment competence and capacity. The evidence is mixed. Evidence suggests that disease-specific systems can serve to expand polyvalent services (Aylward and others 1998). Smallpox eradication, for example, gave many national governments the confidence to introduce the Expanded Program on 1168 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson ©2006 The International Bank for Reconstruction and Development / The World Bank 210 Immunization, with the ability to deliver vaccines and micronutrients in routine schedules and through national campaigns. However, other evidence suggests that some vaccination programs have adversely affected primary health services (Steinglass 2001; Taylor, Cutts, and Taylor 1997) and may have even increased costs. Implementation of international initiatives can also expose conflicts of priorities. The polio eradication initiative, for example, has successfully vaccinated children in the poorest of countries against this disease, but in some of these countries it has failed to timely include the coadministration of measles and other common childhood vaccines, which would have had a much greater effect on child mortality. DISEASE-SPECIFIC CASE STUDIES In this section, we apply the reasoning developed previously to provide an empirical analysis of the three most recent eradication programs—smallpox and the two ongoing programs, poliomyelitis and dracunculiasis. Smallpox As noted before, smallpox eradication was achieved in October 1977, 11 years after the intensified program began. Following implementation of a rigorous certification procedure, the WHA declared smallpox eradicated in 1980. Fenner and others (1988) have estimated the annual benefits of smallpox eradication to developing and industrial countries (see table 62.1). These aggregate estimates, obtained by Table 62.1 Benefits and Costs of Smallpox Eradication (Millions of U.S. dollars) Annual amount Beneficiary India 722 United States 150 All developing countries 1,070 All industrial countries 350 Total annual benefit 1,420 Expenditure Total international, on eradication 98 Total national, by endemic countries 200 Combined total, on eradication 298 Benefit-cost ratio International expenditure 483:1 Combined total expenditure 159:1 Source: Adapted from Fenner and others 1988. prorating estimates of the benefits of eradication for India and the United States to all developing and industrial countries, respectively, suggest that developing countries benefited more from smallpox eradication than industrial countries. Qualitatively, a consistent picture emerges: smallpox eradication was not only an extraordinary investment for the world; it was also an investment that benefited every country, rich and poor alike. When the eradication effort began, smallpox was no longer endemic in most industrial countries. Nonetheless, these countries needed to maintain populationwide immunity under the threat of possible imported cases from endemic countries. They would gain from eradication not only through the cessation of vaccination and its associated costs but also by being able to decrease the number of quarantine inspectors at ports of entry and by averting costs of care related to the adverse events from this live vaccine. The still-endemic countries would also save vaccination costs, although most were vaccinating only a comparatively small proportion of their populations. The greater benefit to them was the avoided cost of disease, including the extraordinary death toll. A number of developing countries had accorded smallpox prevention a high priority, as was evidenced by the number that succeeded in interrupting transmission without international assistance. This list includes China, which was not a member of WHO at the time the eradication effort commenced. Indeed, and as shown in table 62.1, the still-endemic countries contributed an estimated two-thirds of the US$298 million cost of eradication. International sources funded the balance. If the latter cost is interpreted as the incremental cost of achieving eradication, the benefit-cost ratio of global smallpox eradication was over 450:1, a singularly high figure. Even including the expenditure by endemic countries, the benefit of eradication exceeded the cost by an unusually large amount. Brilliant (1985) calculated the annual costs of the smallpox eradication campaign for India to be about US$17 million per year, including indirect costs (lost productivity caused by adverse reactions to vaccination) and opportunity costs (health workers being diverted from other programs). These costs were only a fraction of the annual benefits of eradication to India, which, by Brilliant’s calculations, were US$150 million. The benefit estimates by Fenner and others (1988) are much larger, and those of Ramaiah (1976) are smaller, but all three studies draw the same (qualitative) conclusion: smallpox eradication was a good investment for India. Basu, Ježek, and Ward (1979, 312) present estimates identical to those in Ramaiah (1976), but without giving attribution. Originally, India had decided to undertake a smallpox program just one month after the WHA voted to eradicate the disease globally in 1959. The attempt failed, however, largely for administrative reasons (Basu, Ježek, and Ward 1979; Brilliant Control and Eradication | 1169 ©2006 The International Bank for Reconstruction and Development / The World Bank 211 1985; Fenner and others 1988). Essentially, India had an economic incentive to control smallpox on its own (Brilliant 1985, 33) but lacked organizational capacity and an effective strategy for achieving this goal. Note, however, that India had other health priorities, including family planning. According to Brilliant (1985, 33), “for India’s health planners, occupied then by emergencies and competing political demands on scarce resources, the long-term benefits from disease eradication were not a great motivation. Health planners are sensitive to immediate political realities, and the benefits of smallpox eradication would be realized only at some future time when the $3 million annual expenditures for smallpox could be applied to other health problems. In the meantime, however, the cost of putting so many scarce resources into one program rather than into many health needs was high.” Table 62.2 provides estimates of the benefits of smallpox eradication to the United States. The total benefit of eradication to the United States is about the same order of magnitude as India’s, but the breakdown is different. Whereas India benefited mainly from avoided infections, the United States benefited mainly from avoided vaccinations. By the time the eradication program was launched, the United States had already interrupted smallpox transmission, but vaccination was costly, both in economic and human health terms (a small number of people died every year from infections arising from the live vaccine). Defending the nation from imported infections imposed additional costs. In health terms, smallpox eradication saved millions of lives; in economic terms, it yielded a benefit many times greater than the cost. Identifying another investment that has yielded comparable returns and has benefited every country is difficult. One reason that the economics of smallpox eradication were so favorable is that all countries had strong incentives to join in Table 62.2 Benefits of Smallpox Eradication to the United States, 1968 (Millions of U.S. dollars) Amount Direct costs for medical services Vaccination 92.8 Treatment of complications 0.7 Indirect costs, loss of productivity Work losses attributable to vaccination and reactions 41.7 Permanent disability attributable to complications 0.4 Premature death 0.1 Cost of international traffic surveillance and delays in clearance of vessels Total 14.5 150.2 Source: Sencer and Axnick 1973; see also Fenner and others 1988, table 31.2. the eradication of the disease. A huge organizational effort, but only a relatively small incremental cost, was needed to achieve eradication. The specter of global terrorism has recently caused some countries to prepare themselves for a possible smallpox attack by stockpiling vaccine. Although such actions reduce the benefits of eradication, the economics remain favorable. Smallpox, however, was a special case. Many attributes of the disease and the vaccine favored eradication. The vaccine was heat stable and required only a single dose to protect a person for a period of at least 5 to 10 years. Vaccination was easily performed and protected immediately on application. Every individual who became infected exhibited a typical, easily recognized rash, thus permitting accurate surveillance without recourse to laboratory diagnosis. The disease spread slowly so that transmission could readily be stopped by isolating the patient and vaccinating contacts within the area. Poliomyelitis The polio eradication program, launched by the WHA in 1988, has made substantial progress (CDC 2003a, 2003b, 2003c, 2004a, 2004c, 2004e, 2004g, 2005). The incidence of paralytic poliomyelitis in children fell by more than 99 percent, from an estimated 1,000 cases per day worldwide in 1988 to fewer than 4 cases per day in 2003. The number of poliomyelitis-endemic countries also fell, from 125 in 1988 to just 6 by 2003 (Afghanistan, the Arab Republic of Egypt, India, Niger, Nigeria, and Pakistan). This laudable reduction was the result of repetitive vaccination campaigns with easily administered oral polio vaccine to whole regions, to nations, and to large subpopulations. During 2004, however, polio immunization activities in northern Nigeria were halted for an extended period for fear of tainted vaccines, and this permitted the development of epidemics extending throughout the country. The disease spread as well to 10 other African countries and to Saudi Arabia, Yemen, and Indonesia. Transmission has again been reestablished in several African countries (Burkina Faso, Central African Republic, Chad, Côte d’Ivoire, and Sudan). Heroic efforts are being made to control these outbreaks by large-scale immunization, but in countries such as these, where health services are stressed and the health, communication, and transportation infrastructures are weak, disease transmission is difficult to interrupt. Meanwhile, other countries throughout the world that appear to be polio free are continuing their vaccination programs but finding it increasingly difficult to maintain a momentum of interest, effort, and financing. The difficulties of maintaining credible surveillance systems throughout the developing countries were vividly demonstrated by the discovery of polio in Sudan in May 2004, more than three years after the last case had been reported (CDC 2005). In the interim, specimens from 75 to 90 percent of such 1170 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson ©2006 The International Bank for Reconstruction and Development / The World Bank 212 cases were processed in the laboratory, and measures of surveillance for acute flaccid paralysis cases were reported to have been entirely satisfactory. At first, the Sudanese cases were considered to have resulted from importations from Nigeria, and, indeed, some cases were. However, from more detailed laboratory studies, it was determined that type 1 wild virus had been circulating undetected for more than three years and type 3 virus for nearly five years. Clearly, stopping the continuing transmission of wild poliovirus is itself a formidable challenge, the success of which is by no means certain. A problematic discovery since the global eradication program began was the finding that individuals with particular immunologic disorders shed polio vaccine virus for many months to years, thus serving as a reservoir for this virus. The virus, in turn, can revert to a neurovirulent form, which is capable of causing outbreaks of disease (Bellmunt and others 1999). Such individuals may be wholly without symptoms and impossible to identify except through fecal cultures. Moreover, no treatment is known to stop them from shedding virus. They pose an all but insurmountable challenge to the current poliomyelitis eradication effort. The program is further hampered by the tool that has provided so much success—oral poliovirus vaccine (OPV). In resource-poor environments, poliomyelitis is best controlled with the inexpensive, live, and easily administered oral vaccine. The live vaccine is excreted and can infect other susceptible contacts. The ability of OPV to immunize others indirectly makes it an ideal vaccine for achieving high levels of population-based immunity, especially in lower socioeconomic populations that are the most difficult to reach. However, the excreted virus occasionally reverts to a pathologic state, causing not only cases but outbreaks of vaccine-associated paralytic polio, which may not emerge until months or even years after the vaccine has been administered (Kew and others 2004). Unfortunately, the alternative inactivated polio vaccine (IPV) is not immediately an option in many nations, not least because global manufacturing capacity could not begin to meet demand. Other problems include the current cost differential between OPV and IPV, the increased difficulty of administering the vaccine by syringe and needle, and the need to achieve higher coverage rates with IPV because it does not spread from person to person as does OPV. Tragically, if OPV use were discontinued, in the absence of alternative immunity, polioviruses would likely circulate silently (Eichner and Dietz 1996) and reemerge. Preliminary results from a model presented by WHO indicate a greater than 60 percent chance of an outbreak within two years of the possible global cessation of OPV (WHO 2004) because of continuous circulation of undetected live vaccine viruses that can revert. Outbreaks have already been observed in several regions where decreasing use of live vaccine has left pockets of susceptible persons who eventually have been exposed to vaccine- derived pathogenic viruses (Kew and others 2002). Such an outbreak could occur with disastrous speed because the polio virus is far more contagious than that of smallpox. In developing countries, virtually all cases of polio occurred among those under five years of age, older persons having been protected by the natural immunity of earlier infection. Within five years after vaccination ceased, therefore, the population immunity level in the developing countries would be no better than it was before vaccination was introduced. With this is mind, it seems questionable as to whether all health ministers could be persuaded to call for a country-wide cessation of poliomyelitis vaccination itself, given the uncertainties of virus detection in so many remote and inaccessible areas of the world. By definition, eradication implies certifying cessation of virus transmission and the absence of reservoirs so that control interventions can cease. As noted earlier in this chapter, it is only for this reason that eradication yields a dividend. Although the interruption of wild poliomyelitis virus transmission is theoretically feasible, the obstacles to achieving and maintaining this goal are formidable. At this time, it is difficult to foresee a future that does not envisage a continuing vaccination program, perhaps with IPV use in countries that can afford the substantial additional costs entailed and with OPV use in all other countries. The polio eradication initiative, like that for smallpox, has had to rely primarily on voluntary donations provided both to WHO and bilaterally. Playing an especially important role have been the Rotary International Foundation and the Bill & Melinda Gates Foundation. From 1988 to 2004, more than US$3 billion was spent on the effort (WHO 2003). What are the economics of polio eradication? Bart, Foulds, and Patriarca (1996) developed the first global cost-benefit analysis of polio eradication, beginning with the costs incurred since 1986, the year that the Pan American Health Organization launched a regional eradication effort, and extending to 2040. They assumed that eradication would be achieved in 2005, using OPV, and that vaccination would cease after eradication had been certified. Benefits (like costs, discounted at 6 percent) reflect the avoided costs of acute care and avoided vaccination costs after certification. Their analysis showed that the initiative would break even by 2007 and yield a net benefit to the world of more than US$13 billion by 2040—an encouraging result, but it was based on the assumption that all vaccination would stop abruptly in 2005. Khan and Ehreth (2003) developed a similar analysis but provided regional detail. They estimated the costs and medical costs avoided of polio immunization and eradication over the period 1970 to 2050, assuming that vaccination could cease after 2010. As table 62.3 shows, Khan and Ehreth estimated that polio immunization and eradication would entail a negative net cost overall, with Europe and the Americas saving the most and with other regions incurring a positive net cost. Compared Control and Eradication | 1171 ©2006 The International Bank for Reconstruction and Development / The World Bank 213 Table 62.3 Net Costs of Polio Immunization and Eradication (Millions of U.S. dollars) WHO region Africa Americas Eastern Mediterranean Europe Medical care cost savings Immunization costs Net costs Cost/DALY saved 1,100 3,942 2,842 442 76,900 25,460 ⫺51,440 ⫺4,983 1,930 3,512 1,582 426 Continue OPV Low-income countries Middle-income countries 17,249 ⫺21,001 ⫺2,780 Southeast Asia 1,270 6,519 5,249 1,041 Western Pacific 8,670 10,327 1,657 356 128,120 67,009 ⫺61,111 ⫺1,457 Source: Khan and Ehreth 2003. Note: Cost savings, immunization costs, and net costs are present values for 2000 in millions of U.S. dollars, calculated for the period 1970–2050 and discounted at 5 percent. These estimates assume that immunization by OPV can cease after 2010. with other health interventions, this cost to developing countries may still be comparatively cost-effective. However, Khan and Ehreth comment that the cost per disability-adjusted life year (DALY) saved is high for developing countries (see table 62.3). As they explain (Khan and Ehreth 2003, 705), “This implies that without the financial support from developed countries of the world many developing countries would not have opted for polio interventions for implementation. From the developed countries’ point of view, providing support for the polio program is not simply helping the poor and the disadvantaged, it actually represents a good economic investment.” Unfortunately, both of these cost-benefit studies have substantial limitations. First, both show that eradication is economically attractive if one incorporates all costs and benefits from the inception of this program. Because eradication has not yet been achieved, this approach mixes retrospective evaluation and prospective analysis (historical expenditures and benefits are sunk and so are irrelevant to the current situation). Second, benefits and costs are calculated in both studies relative to a world without immunization. A better approach would be to calculate the net benefits of eradication compared with the alternative of an optimal control program. The choice is not between doing nothing and eradication. It is between an optimal level of control and eradication. Finally, both studies assume that vaccination can cease in 2005 or 2010. As explained previously, this possibility is highly unlikely. A more recent analysis by Sangrujee, Cáceres, and Cochi (2004) calculates the costs for 15 years following the goal of certification of eradication in 2005 for three different scenarios: continued use of OPV, OPV cessation with optional use of the killed or inactivated polio vaccine, and OPV cessation with universal IPV. Table 62.4 shows their results. Total Universal IPV 487 4,418 12,196 12,196 12,196 6,409 6,409 6,409 19,969 19,092 23,023 1,120 1,320 1,120 21,089 20,412 24,143 High-income countries Subtotal Stop OPV 1,364 Global response capacity 38,250 World Table 62.4 Postpolio Eradication Costs (Millions of U.S. dollars) Source: Sangrujee, Cáceres, and Cochi 2004. Note: Costs are expressed in present value terms, calculated over the period 2005 to 2020, and discounted at 3 percent. The respective cost to middle- and high-income countries is the same for all three scenarios, reflecting the assumption that the high-income countries will switch to IPV by 2005 and middle-income countries will do so between 2006 and 2008. The scenarios differ only for the low-income countries. In the first scenario, these countries are expected to continue routine immunization using OPV; in the second, immunization ceases in 2011, followed by a system of surveillance and response. In the third scenario, the low-income countries join the others in switching to IPV between 2008 and 2010. Of these three scenarios, the second comes closest to the 2005 post-eradication strategy now advocated by the polio eradication program leadership. Unfortunately, this analysis is also deficient. First, interruption of transmission will not occur before 2006, and certification will take an additional three years. Hence, analysis of post-eradication costs should begin in 2009 at the earliest, with the costs of continuing immunization needing to be borne up until that time. Second, the analysis assumes a capacity to supply IPV that exceeds current estimates. It is not obvious that this scenario is feasible or, if it were, if the costs of scaling up production are adequately reflected in the calculations. Third, and most importantly, table 62.4 indicates that only low-income countries would benefit from polio eradication over this 15-year time scale—and yet the table does not include any estimate of the risk these countries would bear of a possible outbreak. Although this analysis suggests that the discontinued use of OPV promises the greatest return to eradication, this assumes that circulating vaccine-derived polioviruses could be contained if and when they emerged. However, preparing for this possibility would require a far more effective global surveillance system than now exists, maintenance of a laboratory infrastructure, and stockpiles of OPV. In addition, controlling outbreaks with OPV without the risk of viruses reverting to virulence will be exceedingly difficult in the setting of an accelerating proportion of immunologic-naive individuals. The use of OPV in this scenario could very well cause poliomyelitis to 1172 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson ©2006 The International Bank for Reconstruction and Development / The World Bank 214 again become endemic. In any case, the estimated cost of any of the strategies exceeds $20 billion. The economics of polio eradication are thus not as favorable as concluded by either Bart, Foulds, and Patriarca (1996) or Khan and Ehreth (2003). Both studies assume that vaccination can cease without IPV being used as a substitute anywhere, both exclude the costs of maintaining a response capacity, and neither accounts for the real threat of reemergence. Sangrujee, Cáceres, and Cochi (2004) take account of two of these considerations, but their analysis calculates only the costs for 15 years, ignoring both the risk of reemergence and the benefits of eradication. Hence, each study provides only a partial glimpse of the economics of polio eradication and does not adequately address the fundamental difficulty (inability) of stopping vaccination and maintaining eradication. In conclusion, although the economics of polio eradication may have been thought to be favorable by some (Aylward and others 2003), they are far less favorable than were the economics of smallpox eradication, even assuming that polio vaccination could cease. This last omission is especially relevant to the study’s analysis of the eradication program in Sudan. The study projected that, by 1998, infections would cease everywhere except Sudan. (Plainly, this prediction was wrong, although Sudan is the largest problem for the program, mainly because of the ongoing civil war, which has limited accessibility to endemic areas; see Hopkins and others 2002.) It then calculates the net present value of eliminating the disease there. The results are not promising. They show that eradication is attractive only if the disease can be eliminated in Sudan within five years. However, this analysis ignores the dividend that eradication would earn Sudan. It also disregards the most important feature of eradication—that if the disease were certified to have been eliminated from its last stronghold, it would yield a benefit to all potentially vulnerable countries. Thus, the economics of eliminating dracunculiasis from Sudan, if that is where the disease makes its last stand, will be much more attractive than suggested by this analysis. Dracunculiasis Of the several attempts to eradicate diseases, all but one has failed. Even the exception, smallpox, barely succeeded despite the many factors favorable to eradication. Whether any eradication effort will ultimately succeed or fail cannot be known with certainty at the time it is launched. Eradication entails risk. Money spent on eradication may not ultimately pay a dividend. Health risks may also exist. If eradication fails and vaccination levels drop after the eradication goal is abandoned, susceptible persons who were previously shielded from infection may become infected at a later age, when the disease can cause greater harm. The risk also exists that, even if eradication succeeds, the disease may be reintroduced by accidental or deliberate release. The reasons for potential failure of an eradication effort are many. A nonhuman host may not be discovered until the number of infected humans drops to a very low level (as happened with yellow fever). The tools of eradication may be vulnerable to resistance (insecticides and drugs in the case of malaria). Political problems and civil strife may prevent an eradication program from being executed in critical areas where the disease makes its last stand (a problem today for guinea worm). Termination of vaccination may leave populations vulnerable to microbe reintroduction from an unforeseen reservoir or vaccine strain reversion (a risk now facing the poliomyelitis initiative). Another potential reason for failure is the inability to raise the financial resources needed to complete programs that extend beyond expected targets. All eradication programs have experienced serious financial stringencies during the course of their execution. Most eradication programs to date have been launched as visionary, far-reaching efforts but with vastly incomplete Dracunculiasis, or guinea worm disease, is a nematode infection, which is controlled not by vaccination but by education of the affected population, provision of nematode-free water through wells or filtration, and treatment of cases. It is not a global disease but found only in the rural areas of a few very poor tropical countries. This last difference is especially important from an economics perspective. It means that international financing of a guinea worm eradication program needs to rely more heavily on development assistance rather than on the self-interest of donor countries. Thus far, the eradication program has been successful in reducing the number of cases of guinea worm 99 percent from the 1986 level (Carter Center 2004). The geographic range of the disease has also been reduced from 20 to just 12 countries. Although this achievement is important, eradication remains elusive many years beyond 1995, the year that the WHA set for eradication in 1991 (Cairncross, Muller, and Zagaria 2002, 232). Only one cost-benefit study of the guinea worm eradication program has been published (Kim, Tandon, and Ruiz-Tiben 1997), and it is unfortunately flawed in a number of respects. First, as indicated previously, eradication costs should be compared with those associated with an alternative optimal control program. Second, the cost-benefit analysis applies to the period 1987 to 1998 and thus is backward looking. The analysis can reveal whether the money spent previously yielded a benefit in excess of the cost (it did), but it cannot reveal whether eradication was worth pursuing at the time that this study was undertaken. Finally, it takes no account of the investment decision of eradication—the main reason for pursuing the eradication goal in the first place. CONCLUSIONS Control and Eradication | 1173 ©2006 The International Bank for Reconstruction and Development / The World Bank 215 information. Basic epidemiological information and knowledge of the effectiveness and operational constraints of interventions and costs in different settings are often inadequate, and the required monitoring, evaluation, training, and research components of the program may be absent. If a program’s administrators lack a careful, probing analysis of the epidemiology of the various candidate diseases or of the technologies available, and if their comprehension of the potential costs and who would bear them is limited, a program is likely to founder, causing a dispirited staff, confused beneficiaries, and donor fatigue and ambivalence. It is crucial that the eradication methodologies and assumptions in those regions of the world that would be most likely to pose the most significant problems be tested and addressed before launching an eradication program and that evaluation and research continue during the program. Proposals for disease eradication have seldom been brought to the WHA with specific plans, costs, and uncertainties fully laid out. Nor have the expected sources of fiscal support and needed country support been addressed with specific commitments requested of the members. The WHA has only a limited deliberative capacity, and too much cannot be expected of its members in session. However, designated special committees of the WHA can and should be appointed, consisting of both visionary eradicationists and field-experienced public health and social science personnel. The WHA should take up the question of eradication only after the subject has been thoroughly vetted and sufficiently large-scale pilot programs in the most problematic areas have clarified that an adequate understanding of the epidemiology exists and that the appropriate technologies are available. In the past, members have not voted for a specific program for which all the uncertainties have been laid out and the benefits and costs associated with different outcomes have also been calculated. Nor, with one exception, have they voted for a resolution imposing responsibilities, including financing obligations, on individual states. The next time a proposal to eradicate a disease is presented to the WHA, it should be comprehensive. It should demonstrate why the effort is worth taking, even if the final outcome is uncertain; it should bind states, morally if not legally, to fulfill the pledges needed to see the program through to its completion; and it should prepare contingencies should the eradication effort fail. REFERENCES Anderson, R. M., and R. May. 1991. Infectious Diseases of Humans: Dynamics and Control. Oxford, U.K.: Oxford University Press. Andrews, J. M., and A. D. Langmuir. 1963. “The Philosophy of Disease Eradication.” American Journal of Public Health 53: 1–6. Aylward, R. B., A. Acharya, S. England, M. Agocs, and J. Linkins. 2003. “Polio Eradication.” In Global Public Goods for Health: Health Economic and Public Health Perspectives, eds. R. Smith, R. Beaglehole, D. Woodward, and N. Drager, 33–53. Oxford, U.K.: Oxford University Press. Aylward, R. B., J. M. Olive, H. F. Hull, C. A. De Quadros, and B. Melgaard. 1998. “Ensuring Common Principles Lead to Mutual Benefits: Disease Eradication Initiatives and General Health Services.” In The Eradication of Infectious Diseases, eds. W. R. Dowdle and D. R. Hopkins. New York: John Wiley and Sons. Barrett, S. 2003. “Global Disease Eradication.” Journal of the European Economic Association 1: 591–600. ———. 2004. “Eradication vs. Control: The Economics of Global Infectious Disease Policy.” Bulletin of the World Health Organization 82 (9): 683–88. Barrett, S., and M. Hoel. 2003. “Optimal Disease Eradication.” Disease Control Priorities Project Working Paper 22. Bart, K. J., J. Foulds, and P. Patriarca. 1996. “Global Eradication of Poliomyelitis: Benefit-Cost Analysis.” Bulletin of the World Health Organization 74 (1): 35–45. Basu, R. N., Z. Ježek, and N. A. Ward. 1979. The Eradication of Smallpox from India. New Delhi: World Health Organization. Bellmunt, A., G. May, R. Zell, P. Pring-Akerblom, W. Verhagen, and A. Heim. 1999. “Evolution of Poliovirus Type I during 5.5 Years of Prolonged Enteral Replication in an Immunodeficient Patient.” Virology 265: 178–84. Biellik, R., S. Madema, A. Taole, A. Kutsulukuta, E. Allies, R. Eggers, and others. 2002. “First 5 Years of Measles Elimination in Southern Africa: 1996–2000.” Lancet 359 (9317): 1564–68. Breman, J. G., and I. Arita. 1980. “The Confirmation and Maintenance of Smallpox Eradication.” New England Journal of Medicine 303 (22): 1263–73. Brilliant, L. B. 1985. The Management of Smallpox Eradication in India. Ann Arbor: University of Michigan Press. Cairncross, S., R. Muller, and N. Zagaria. 2002. “Dracunculiasis (Guinea Worm Disease) and the Eradication Initiative.” Clinical Microbiology Reviews 15 (2): 223–46. Carter Center. 2004. “Guinea Worm Eradication: Review, June 1, 2003.” Carter Center, Atlanta. http://www.cartercenter.org/printdoc.asp? docID=1785andsubmenu=news. CDC (U.S. Centers for Disease Control and Prevention). 1998a. “Progress toward Elimination of Measles from the Americas.” Morbidity and Mortality Weekly Report 47 (10): 189–93. ———. 1998b. “Progress toward Global Measles Control and Regional Elimination, 1990–1997.” Morbidity and Mortality Weekly Report 47 (48): 104. ACKNOWLEDGMENTS We would like to acknowledge Walter Dowdle and Maria Teresa Valenzuela for their critical review of our manuscript; Joel Breman and Philip Musgrove for providing editorial guidance; John Sentz for his numerous hours devoted to research and editorial assistance; and Cherice Holloway for her preparation of various versions of this manuscript. ———. 1999a. “Global Measles Control and Regional Elimination, 1998–1999.” Morbidity and Mortality Weekly Report 48 (49): 1124. ———. 1999b. “Progress toward Measles Elimination—Southern Africa, 1996–1998.” Morbidity and Mortality Weekly Report 48 (27): 585. ———. 2003a. “Global Progress toward Certifying Polio Eradication and Laboratory Containment of Wild Polioviruses—August 2002–August 2003.” Morbidity and Mortality Weekly Report 52 (47):1158. 1174 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson ©2006 The International Bank for Reconstruction and Development / The World Bank 216 ———. 2003b. “Progress toward Poliomyelitis Eradication—Angola and the Democratic Republic of Congo, January 2002–June 2003.” Morbidity and Mortality Weekly Report 52 (34): 816. Fenner, F., A. J. Hall, and W. R. Dowdle. 1998. “What Is Eradication?” In The Eradication of Infectious Diseases, eds. W. R. Dowdle and D. R. Hopkins, 3–17. New York: John Wiley and Sons. ———. 2003c. “Progress toward Poliomyelitis Eradication—Southern Africa, 2001—March 2003.” Morbidity and Mortality Weekly Report 52 (22): 521. Fenner, F., D. A. Henderson, I. Arita, Z. Ježek, and I. D. Ladnyi. 1988. Smallpox and Its Eradication. Geneva: World Health Organization. ———. 2003d. “Update: Global Measles Control and Mortality Reduction—Worldwide, 1991–2001.” Morbidity and Mortality Weekly Report 52 (20): 471. ———. 2004a. “Brief Report: Global Polio Eradication Initiative Strategic Plan, 2004.” Morbidity and Mortality Weekly Report 53 (05): 107. ———. 2004b. “Measles Mortality Reduction—West Africa, 1996–2002.” Morbidity and Mortality Weekly Report 53 (02): 28. ———. 2004c. “Progress toward Global Eradication of Poliomyelitis, January 2003–April 2004.” Morbidity and Mortality Weekly Report 53 (24): 532. Goodman, R. A., K. L. Foster, F. L. Trowbridge, and J. P. Figueroa, eds. 1998a. “Comments and Discussion Following Work Group Reports.” In “Global Disease Elimination and Eradication as Public Health Strategies,” Bulletin of the World Health Organization 76 (Suppl. 2): 104–8. ———. 1998b. “Global Disease Elimination and Eradication as Public Health Strategies: Proceedings of a Conference. Atlanta, Georgia, February 23–25, 1998.” Bulletin of the World Health Organization 76 (Suppl.) 2: 5–162. Henderson, D. A. 1982. “The Deliberate Extinction of a Species.” Proceedings of the American Philosophical Society 126: 461–71. ———. 2004d. “Progress toward Measles Elimination—Region of the Americas, 2002–2003.” Morbidity and Mortality Weekly Report 53 (14): 304. ———. 1988. “Development of the Global Smallpox Eradication Programme, 1958–1966.” In Smallpox and Its Eradication, eds. F. Fenner, D. A. Henderson, I. Arita, Z. Ježek, and I. D. Ladnyi, 365–419. Geneva: World Health Organization. ———. 2004e. “Progress toward Poliomyelitis Eradication—Nigeria, January 2003–March 2004.” Morbidity and Mortality Weekly Report 53 (16): 343. ———. 1998a. “Eradication: Lessons from the Past.” Bulletin of the World Health Organization 76 (Suppl. 2): 17–21. ———. 2004f. “Progress toward Sustainable Measles Mortality Reduction—South-East Asia Region, 1999–2002.” Morbidity and Mortality Weekly Report 53 (25): 559. ———. 2004g. “Wild Poliovirus Importations—West and Central Africa, January 2003–March 2004.” Morbidity and Mortality Weekly Report 53 (20): 433. ———. 2005. “Progress toward Poliomyelitis Eradication—Poliomyelitis Outbreak in Sudan, 2004.” Morbidity and Mortality Weekly Report 54 (4): 97–99. ———. 1998b. “The Siren Song of Eradication.” Journal of the Royal College of Physicians of London 32 (6): 580–84. Hopkins, D. R., E. Ruiz-Tiben, N. Diallo, P. C. Withers Jr., and J. H. Maguire. 2002. “Dracunculiasis Eradication: And Now, Sudan.” American Journal of Tropical Medicine and Hygiene 67 (4): 415–22. Jeffrey, G. M. 1976. “Malaria Control in the Twentieth Century.” American Journal of Tropical Medicine and Hygiene 25: 361–71. Jenner, E. 1801. The Origin of Vaccine Inoculation. London: Shury. Quoted in Fenner, Hall, and Dowdle 1998. Cello, J., A. V. Paul, and E. Wimmer. 2002. “Chemical Synthesis of Poliovirus cDNA: Generation of Infectious Virus in the Absence of Natural Template.” Science 297 (5583): 1016–18. Kew, O. M., V. Morris-Glasgow, M. Landaverde, C. Burns, J. Shaw, Z. Garib, and others. 2002. “Outbreak of Poliomyelitis in Hispaniola Associated with Circulating Type 1 Vaccine-Derived Poliovirus.” Science 296 (5566): 356–59. Cochi, S., C. de Quadros, W. Dowdle, R. Goodman, P. Ndumbe, D. Salisbury, and others. 1998. “Post-Conference Small Group Report.” In Global Disease Elimination and Eradication as Public Health Strategies, eds. R. A. Goodman, K. L. Foster, F. L. Trowbridge, and J. P. Figueroa. Bulletin of the World Health Organization 76 (Suppl. 2): 113. Kew, O. M., P. F. Wright, V. I. Agol, F. Delpeyroux, H. Shimizu, N. Nathanson, and M. A. Pallansch. 2004. “Circulating Vaccine-Derived Polioviruses: Current State of Knowledge.” Bulletin of the World Health Organization 82 (1): 16–23. Cockburn, A. 1961. “Eradication of Infectious Diseases.” Science 133: 1050–58. Khan, M., and J. Ehreth. 2003. “Costs and Benefits of Polio Eradication: A Long-Run Global Perspective.” Vaccine 21: 702–5. ———. 1963. The Evolution and Eradication of Infectious Diseases. Baltimore: Johns Hopkins University Press. Kim, A., A. Tandon, and E. Ruiz-Tiben. 1997. “Cost-Benefit Analysis of the Global Dracunculiasis Eradication Campaign.” Policy Research Working Paper 1835, World Bank, Washington, DC. de Quadros, C. 2001. “Introduction.” In Considerations for Viral Disease Eradication Lessons Learned and Future Strategies, eds. S. Knobler, J. Lederberg, and L. A. Pray, 22–32. Washington, DC: National Academy Press. Knobler, S., J. Lederberg, and L. A. Pray, eds. 2001. Considerations for Viral Disease Eradication Lessons Learned and Future Strategies. Washington, DC: National Academy Press. Dowdle, W. R. 1998. “The Principles of Disease Elimination and Eradication.” Bulletin of the World Health Organization 76 (Suppl. 2): 22–25. ———. 1999. “The Principles of Disease Elimination and Eradication.” Morbidity and Mortality Weekly Report 48 (Suppl.): 23–27. Dowdle, W. R., and D. R. Hopkins, eds. 1998. The Eradication of Infectious Diseases. New York: John Wiley and Sons. Dubos, R., and J. Dubos. 1953. The White Plague: Tuberculosis, Man and Society. London: Gollancz. Quoted in Fenner, Hall, and Dowdle 1998. Eichner, M., and K. Dietz. 1996. “Eradication of Poliomyelitis: When Can One Be Sure That Polio Virus Transmission Has Been Terminated? American Journal of Epidemiology 143 (8): 816–22. Kremer, M., and E. Miguel. 2004. “The Illusion of Sustainability.” NBER Working Paper W10324, National Bureau of Economic Research, Cambridge, MA. New Shorter Oxford English Dictionary. 1993. Ottesen, E. A., W. R. Dowdle, F. Fenner, K. O. Habermehl, T. J. John, M. A. Koch, and others. 1998. “Group Report: How Is Eradication to Be Defined and What Are the Biological Criteria?” In The Eradication of Infectious Diseases, eds. W. R. Dowdle and D. R. Hopkins. New York: John Wiley and Sons. Payne, A. M.-M. 1963a. “Basic Concepts of Eradication.” American Review of Respiratory Disease 88: 449–55. ———. 1963b. “Disease Eradication as an Economic Factor.” American Journal of Public Health 53: 369–75. Control and Eradication | 1175 ©2006 The International Bank for Reconstruction and Development / The World Bank 217 Ramaiah, T. J. 1976. “Cost-Benefit Analysis of the Intensified Campaign against Smallpox in India.” National Institute of Health Administration and Education Bulletin 9 (3): 169–203. Roberts, L. 2004. “Polio Endgame. Polio: The Final Assault?” Science 303 (5666): 1960–68. Sabin, A. B. 1991. “Measles, Killer of Millions in Developing Countries: Strategy for Rapid Elimination and Continuing Control.” European Journal of Epidemiology 7 (1): 1–22. Salisbury, D. 1998. “Report of the Work Group on Disease Elimination/Eradication and Sustainable Health Development.” In Global Disease Elimination and Eradication as Public Health Strategies, eds. R. A. Goodman, K. L. Foster, F. L. Trowbridge, and J. P. Figueroa. Bulletin of the World Health Organization 76 (Suppl. 2): 72–79. Sangrujee, N., V. M. Cáceres, and S. L. Cochi. 2004. “Cost Analysis of Post-Polio Certification Immunization Policies.” Bulletin of the World Health Organization 82 (1): 9–15. Sencer, D. J., and N. W. Axnick. 1973. “Cost Benefit Analysis.” In International Symposium on Vaccination against Communicable Diseases, Monaco 1973, 22: 37–46. Symposia Series in Immunobiological Standardization. Basel, Switzerland: Karger. Spînu, I., and S. Biberi-Moroianu. 1969. “Theoretical and Practical Problems Concerning the Eradication of Communicable Diseases.” Archives roumaines de pathologie experimentales et de microbiologie 28: 725–42. Steinglass, P. 2001. Thematic Evaluations in 2001 Eradication of Poliomyelitis. Report by the director general, Programme Development Committee of the Executive Board Eight Meeting, December 13, 2001. Document EBPDC8/3. Geneva: World Health Organization. Taylor, C. E., F. Cutts, and M. E. Taylor. 1997. “Ethical Dilemmas in Current Planning for Polio Eradication.” American Journal of Public Health 87 (6): 922–25. WHO (World Health Organization). 2003. Global Polio Eradication Initiative: Estimated External Financial Resource Requirements 2004–2008. Geneva: WHO. ———. 2004. Report of the Strategic Advisory Group of Experts (SAGE). Geneva: WHO. http://www.who.int/vaccines-documents/DocsPDF05/ Sage_Report_2004.pdf. Yekutiel, P. 1980. Eradication of Infectious Diseases: A Critical Study. New York: Karger. 1176 | Disease Control Priorities in Developing Countries | Mark Miller, Scott Barrett, and D. A. Henderson ©2006 The International Bank for Reconstruction and Development / The World Bank 218 Chapter 70 Improving the Quality of Care in Developing Countries John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and Julio Frenk Although the quantity rather than quality of health services has been the focus historically in developing countries, ample evidence suggests that quality of care (or the lack of it) must be at the center of every discussion about better health. The following examples are illustrative: In one study evaluating pediatric care in Papua New Guinea, 69 percent of health center workers reported that they checked for only two of the four examination criteria for pneumonia cases. Only 24 percent of these workers were able to indicate correct treatment for malaria. When clinical encounters were observed at aid posts, providers met minimal examination criteria in only 1 percent of cases (Beracochea and others 1995). In a study in Pakistan, only 56 percent of providers met an acceptable diagnostic standard for viral diarrhea, and only 35 percent met the acceptable standard for treatment (Thaver and others 1998). QUALITY DEFINITION AND POPULATION FRAMEWORK These deficiencies in quality of care represent neither the failure of professional compassion nor necessarily a lack of resources (Institute of Medicine 2001). Rather, they result from gaps in knowledge, inappropriate applications of available technology (Murray and Frenk 2000), or the inability of organizations to change (Berwick 1989). Local health care systems may have failed to align practitioner incentives and objectives, to measure clinical practice, or to link quality improvement to better health outcomes. Increasing evidence, much of it developed since the mid 1990s, shows that quality can be improved rapidly. However, to improve clinical practice—and thus quality of care—quality must be defined and measured, and appropriate steps must be taken (Silimper and others 2002). This chapter highlights approaches to improving clinical practice and quality of care that take place over months instead of years. Indeed, better quality can improve health much more rapidly than can other drivers of health, such as economic growth, educational advancement, or new technology. Definition and Framework Health systems provide health actions—activities to improve or maintain health. These actions take place in the context of and are influenced by political, cultural, social, and institutional factors (shown along the edges of figure 70.1). Demographic and socioeconomic makeup, including genetics and personal resources, affect the health status of individuals seeking care. Access to the health care system is required to obtain the care that maintains or improves health, but simple access is not enough; the system’s capacities must be applied skillfully. Thus, quality means optimizing material inputs and practitioner skill to produce health. As the Institute of Medicine defines it, quality is “the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge” (Institute of Medicine 2001, 244). 1293 ©2006 The International Bank for Reconstruction and Development / The World Bank 219 Institutional Factors Political Factors Health Policy Reforms Demographic and Socioeconomic Factors Health Care Access Structure Process Health Outcomes The Quality of Care Cultural Factors Social Factors Source: Peabody and others 1999. Figure 70.1 Quality-of-Care Framework Elements of Quality. Quality comprises three elements: • Structure refers to stable, material characteristics (infrastructure, tools, technology) and the resources of the organizations that provide care and the financing of care (levels of funding, staffing, payment schemes, incentives). • Process is the interaction between caregivers and patients during which structural inputs from the health care system are transformed into health outcomes. • Outcomes can be measured in terms of health status, deaths, or disability-adjusted life years—a measure that encompasses the morbidity and mortality of patients or groups of patients. Outcomes also include patient satisfaction or patient responsiveness to the health care system (WHO 2000). Structural measures are the easiest to obtain and most commonly used in studies of quality in developing countries. Many evaluations have revealed shortages in medical staff, medications and other important supplies, and facilities, but material measures of structure, perhaps surprisingly, are not causally related to better health outcomes (Donabedian 1980). Although higher technology or a more pleasant environment may be conducive to better-quality care, the evidence indicates only a weak link between such structural elements and better health outcomes (Donabedian 1988). The notable exceptions are cases in which physical improvements either increase access to primary care in very poor settings or increase the volume of a clinical procedure, such as cataract surgery, that is specifically linked to better health outcomes (Javitt, Venkataswamy, and Sommer 1983). At best, however, structure is a blunt approximation of process or outcomes; structural improvements by themselves rarely improve the health of a population. Process, by contrast, can be measured with every visit to a provider. Measuring process is difficult, however, particularly in developing countries. The private nature of the doctorpatient consultation, a lack of measurement criteria, and the absence of reliable measurement tools have limited the ability to assess process (Peabody, Tozija, and others 2004). However, new methods are being developed that can provide valid measurements of clinical practice (Thaver and others 1998). In addition, evidence-based clinical studies have steadily revealed which process measures lead to better health outcomes. This combination of ubiquity, measurability, and linkage to health outcomes makes the measurement of process the preferred way to assess quality. Although good outcomes are the objective of all health actions, outcomes alone are not an efficient way to measure quality for two reasons. The first is the quality conundrum. A patient may receive poor-quality care but may recover fully, or a patient may receive high-quality care for an illness such as cerebral malaria and still not recover. Second, adverse health outcomes are relatively rare and obviously do not occur with every encounter. The classic framework of structure-process-outcome is well established. However, in recent years the concept of quality has been expanded to include specific aims for improvement. For example, the Institute of Medicine’s (2001) landmark report, 1294 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 220 Box 70.1 The Institute of Medicine’s Six Elements of Quality 1. Patient safety. Are the risks of injury minimal for patients in the health system? 2. Effectiveness. Is the care provided scientifically sound and neither underused nor overused? 3. Patient centeredness. Is patient care being provided in a way that is respectful and responsive to a patient’s preferences, needs, and values? Are patient values guiding clinical decisions? 4. Timeliness. Are delays and waiting times minimized? 5. Efficiency. Is waste of equipment, supplies, ideas, and energy minimized? 6. Equity. Is care consistent across gender, ethnic, geographic, and socioeconomic lines? Source: Institute of Medicine 2001. Crossing the Quality Chasm, broadens the concept to include other, more contextual elements to illuminate how process changes can improve care. It focuses on six aims: patient safety, effectiveness, patient centeredness, timeliness, efficiency, and equity (see box 70.1). Quality Assessment Perspectives. We can look at the Institute of Medicine’s aims from two perspectives: patient perception, and technical or professional assessment. Patients’ perceptions of quality depend on their individual characteristics and affect their compliance, follow-up decisions, and long-term lifestyle changes (Zaslavsky and others 2000). Interpersonal relationships, cultural appropriateness, and gender sensitivity—long thought to be luxuries of wealthier countries—are also major determinants of patient access and utilization in developing countries. These findings have led to the inclusion of patient satisfaction and patient responsiveness as outcome measures. Technical assessment concerns whether providers meet normative standards for appropriateness of care or adherence to explicit evidence-based criteria. Although patient perception or satisfaction is important, researchers increasingly rely on objective, evidence-based quality criteria that can be more readily linked to better health outcomes at both the individual and the population levels. Population-Level Considerations. Quality is typically assessed through the interaction between individual doctors and patients. However, emerging evidence shows that the average quality of care given by groups of doctors and other providers is an important determinant of overall community health status. For example, in a cross-sectional analysis in the former Yugoslav Republic of Macedonia, researchers found not only that patients’ heath status was significantly higher in areas where quality was higher but also that the overall self-reported health status of those members of the general population who had not recently received care was higher (Peabody, Tozija, and others 2004). Our quality-of-care framework supports these findings. When process is improved among groups of providers,the aggregate improvement in quality leads to better health outcomes for the entire patient population. In addition, resources can be allocated among clinical interventions based on actual effectiveness and the overall impact of care on the population. For example, cancer chemotherapy may be available and may prolong the lives of cancer patients.However,it may result in fewer lives saved than the expansion of coverage of directly observed treatment shortcourse coverage for tuberculosis patients. QUALITY OF CARE IN DEVELOPING COUNTRIES The process of providing care in developing countries is often poor and varies widely. A large body of evidence from industrial countries consistently shows variations in process, and these findings have transformed how quality of care is perceived (McGlynn and others 2003). A 2002 study found that physicians complied with evidence-based guidelines for at least 80 percent of patients in only 8 of 306 U.S. hospital regions (Wennberg, Fisher, and Skinner 2002). It is important to note that these variations appear to be independent of access to care or cost of care: Neither greater supply nor higher spending resulted in better care or better survival. Studies from developing countries show similar results. For example, care in tertiary and teaching hospitals and care provided by specialists may be better than care for the same cases in primary care facilities and by generalists (Walker, Ashley, and Hayes 1988). One explanation for variation and low-quality care in the developing world is lack of resources. Limited data indicate, however, that high-quality care can be provided even in environments with severely constrained resources. A study in Jamaica, which used a cross-sectional analysis of government-run Improving the Quality of Care in Developing Countries | 1295 ©2006 The International Bank for Reconstruction and Development / The World Bank 221 primary care clinics, showed that better process alone was linked to significantly greater birthweight (Peabody, Gertler, and Liebowitz 1998). A study in Indonesia attributed 60 percent of all perinatal deaths to poor process and only 37 percent to economic constraints (Supratikto and others 2002). Cross-system or cross-national comparisons provide the best examples of the great variation in clinical practice in developing countries. In one seven-country study, researchers directly observing clinical practice found that 75 percent of cases were not adequately diagnosed, treated, or monitored and that inappropriate treatment with antibiotics, fluids, feeding, or oxygen occurred in 61 percent of cases (Nolan and others 2001). Another study compared providers’ knowledge and practice in California and FYR Macedonia, using vignettes to adjust for case-mix severity. Although the quality of the overall or aggregate process was lower in FYR Macedonia, a poor country, the top 5 percent of Macedonian doctors performed as well as or better than the average Californian doctor (Peabody, Tozija, and others 2004). In a study commissioned for this chapter, an international team measured quality in five developing countries (China, El Salvador, India, Mexico, and the Philippines), using the same clinical vignettes at each site. The team evaluated the process for common diseases according to international, evidence-based criteria. Quality varied only slightly among countries. The within-country range of quality of doctors was 10 times as great as the between-country range. Such wide variation strongly suggests that efforts to improve health status must involve policies that change the quality of clinical care. POLICY INTERVENTIONS TO IMPROVE QUALITY The success of quality improvement policies can be measured by their ability to raise the average level of health and reduce variation in quality. Two types of policies are intended to improve quality and thus health outcomes: • those that influence provider behavior by altering the structural conditions of organization and finance or that involve the design and redesign of health care systems • those that directly target provider behavior at the individual or the group level. Within each category, the evidence is examined to see the effect of the policy on the health outcomes of populations. Interventions Affecting Provider Practice by Changing Structural Conditions Although structural components such as materials and staff are not strongly linked to outcomes, other components of structure—organization and finance—can influence process by changing the socioeconomic, legal and administrative, cultural, and information context of the health care system. Legal Mandates, Accreditation, and Administrative Regulations. Legal mandates, accreditation, and administrative regulations affect quality by controlling entry into the practice of health care. These policies include the licensing of professionals and facilities, their accreditation or certification to perform certain procedures, and the formal delineation of functions that various types of health workers can legally perform. Although these policies assume that providers’ prior qualifications are good predictors of actual performance in health care delivery, there is little evidence that such policies have a positive effect on process or outcomes. They are more successful at barring unqualified persons from practicing than at ensuring quality among those who are allowed to practice. A review of health sector regulations in Tanzania and Zimbabwe, for example, revealed that the regulations primarily control entry into the market and ensure a minimum standard of quality (Kumaranayake and others 2000). Hospital accreditation, with its periodic reviews of health facility performance standards, can potentially provide ongoing regulatory pressure for improvement. To date, research has not demonstrated that hospital accreditation programs are linked to improvements in health outcomes. In a randomized controlled trial of a hospital accreditation program in the KwaZulu-Natal province of South Africa, researchers showed a conclusive link between the implementation of the program and improvements in the accreditation standard indicators. However, they were unable to link those indicators to improvements in health outcomes (Salmon and others 2003). Malpractice Litigation to Enforce Legal Mandates To be effective in promoting quality, malpractice litigation must rely on adequate legal and judicial systems, which are deficient in most developing countries. In India, one of the few developing countries with the appropriate legal structure in place, inclusion of the medical sector under the Consumer Protection Act of 1986 allows victims to receive redress for negligent medical practice. Although improvements have resulted, some argue that the system needs greater involvement of professional organizations to be effective (Bhat 1996). Professional Oversight Peer review is as old as professional societies. The power and the influence of such societies vary widely among countries (Heaton 2000). Large provider organizations, such as hospitals or public health institutions, often routinely collect information on provider practices and patient outcomes and use those data to guide, educate, supervise, discipline, or recognize providers. In the Philippines, public health managers used a checklist of 20 observable behaviors against which health workers in remote provinces were rated. The 1296 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 222 performance of providers in facilities where workers were reviewed was significantly better than in comparable facilities that did not adopt the reviews (Loevinsohn, Guerrero, and Gregorio 1995). Others, however, assert that the “quality by inspection” environment engendered by oversight leads to an antagonistic relationship between workers and managers and precludes cooperative problem solving and continuous improvement (Berwick 1989). A qualitative study evaluating supervisor-provider interactions in health care facilities in Zimbabwe found that supervisors were adept at giving technical feedback but were not as proficient at making suggestions for improvement or at working with providers and patients to solve problems (Tavrow, Kim, and Malianga 2002). National and Local Clinical Guidelines In many industrial countries, evidence-based clinical guidelines are used to ensure high-quality care, better health outcomes, and cost-effective treatments. (Examples of institutions supporting this approach are the U.K. National Institute for Clinical Excellence, the U.S. Agency for Healthcare Research and Quality, and the Dutch College of General Practitioners.) Guidelines are typically developed for a clinical disease or symptom. They should be derived from evidence-based criteria resulting from welldesigned clinical investigations or expert opinion. Because they are derived from empirical studies, guidelines in developing countries can, in principle, be identical to those in industrial countries. When resource constraints limit transferability, diagnostic and treatment guidelines may have to be modified. Technologies such as x-ray studies have gained widest acceptance in preventive and primary care services, such as integrated management of childhood illness, where they serve both as clinical standards and as educational guides. Including physicians in the development and review of guidelines has proved particularly effective in the challenging process of implementing guidelines. Sharing Information on Quality Improvement Technology. Worldwide interest in quality has given rise to new professional bodies, scientific publications, and institutions dedicated to sharing ideas and innovations in quality improvement. Organizations such as the Robert Wood Johnson Foundation, the Nuffield Trust, and the Institute for Healthcare Improvement cultivate ideas for improvement, bring people and organizations together to learn from each other, and take action to achieve results. Although the sharing of information on quality health care practices has long been an established part of provider education and training networks, the sharing of information on successful systemwide policies for process improvements could potentially accelerate the scale-up of quality practice. One organization active in developing countries is the Council on Health Research for Development (COHRED), which promotes, facilitates, and evaluates the Essential National Health Research strategy in such countries as Benin, the Arab Republic of Egypt, and Indonesia. COHRED aims to develop a system of effective health research to improve health services, including quality of care. The Quality Assurance project funded by the U.S. Agency for International Development has studied and shared information about quality in the developing world since 1990. Under the Quality Assurance project umbrella, researchers have studied and implemented quality measurement and improvement interventions and have used these case studies to develop a library of tools and articles to promote global quality improvement. Public-Private Provision of Care. In most health care systems, a professional regulatory framework governs the network of civil servants delivering health care. These civil servants operate alongside autonomous, self-governed, private providers— independent for-profit physicians and health clinics and nonprofit nongovernmental organizations (NGOs). Two conclusions arise from the often heated debate about the right balance between public and private services. First, private practitioners provide a significant amount of care in developing countries. Second, though there is no one prescription for striking the right public-private mix, in some cases the public regulatory framework has led to private provision of higher-quality care. The government of Senegal successfully contracted with community-based groups for preventive nutrition services. Eighteen months after nutrition services were implemented, severe malnutrition disappeared among children age 6 to 11 months (Marek and others 1999). The success of the program has led to its expansion nationwide. Targeted Education and Professional Retraining. Continuing medical education is a common approach to improving clinical practice, but it neither changes clinical practice nor advances health outcomes (Davis and others 1995). Newer techniques—targeted education, case-based learning, and interactive and multimodel teaching techniques—have had some success. In Guatemala, distance education targeting diarrhea and cholera case management increased accurate assessment and classification of diarrhea cases by 25 percent. Rehydration did not improve, however, and improvements in counseling were insignificant (Flores, Robles, and Burkhalter 2002). In Tanzania, training staff in the control of acute respiratory infections of young children yielded reductions in under-five mortality within two years (Mtango and Neuvians 1986). Organizational Change. In recent years, organizational change in the health care system has been shown to influence quality of care and to further the six aims of the Institute of Medicine by focusing on the continual design and redesign of Improving the Quality of Care in Developing Countries | 1297 ©2006 The International Bank for Reconstruction and Development / The World Bank 223 systems. The emphasis is on developing organizational and individual capabilities where they most profoundly affect the process of care. Design and redesign interventions assume that simply adding a new resource or a new process in isolation will not improve care because better care is the product of many processes working together. Although change interventions have not been widely used in the developing world because they require large investments to plan and implement, four related models of organizational change have been successful in changing provider practice in developing nations: • Total Quality Management in health care Advances in business management practices to continually design and redesign systems for quality improvement have been effectively adapted for health systems. In Total Quality Management, also known as Continuous Quality Improvement, teams use mutually reinforcing techniques in a cycle of planning, implementing, evaluating, and revising to improve the quality of clinical and administrative processes. These techniques include process mapping, statistical quality control, and structured team activities. In rural Bihar, India, private practitioners who treat sick children were provided with standard case-management information, were given feedback on their performance, and were tracked and monitored over time. This strategy produced significant improvements in practitioners’ case-management skills (Chakraborty, D’Souza, and Northrup 2000). In Malaysia, anesthesia safety has been improved through the implementation of consensus-based protocols that emphasize (a) communication among the operating, recovery, and ward team members; (b) individual feedback; and (c) frequent monitoring to identify areas for improvement (Tan 1999). • Collaborative Improvement Model The early success of Total Quality Management techniques has given rise to a related model, the Collaborative Improvement Model. It addresses broad and complex systemic processes within health care systems and has facilitated the scale-up of quality improvements. This model, designed to continuously improve organizational and individual performance, comprises four elements: definition of an aim, measurement, innovation, and testing to see whether the innovation meets the original aim. This approach strikes a pragmatic balance between the need for action and the need to be scientifically grounded. It has been used with success in Peru and the Russian Federation. In Peru, the collaborative improvement model was used by multidisciplinary teams in 41 clinics to design changes aimed at achieving world-class tuberculosis care. The preliminary results have led to impressive changes in the process of care, but it is too early to determine whether they have been effective in improving quality (Berwick 2004). • Plan-Do-Study-Act cycle The Plan-Do-Study-Act (PDSA) cycle calls for action-oriented learning in quality improve- ment. Team members using the PDSA model design a quality-improvement intervention (plan), implement it on a small scale (do), evaluate the results (study), and implement or alter the intervention accordingly (act). Often multiple PDSA cycles are necessary before the appropriate improvement method can be identified. All improvement techniques that involve the design and redesign of systems use some form of the PDSA cycle. Successful scale-up of a PDSA prototype is possible with careful leadership oversight. A team of investigators in Russia’s Tula province developed a series of successful interventions for adults who have poorly controlled hypertension. The interventions, which were started in 20 clinics, were expanded to 500 clinics within 18 months. The scale-up resulted in a sevenfold increase in patients receiving hypertension management at the primary care level and an 85 percent reduction in admissions for hypertension. In Tver province, the same group addressed problems related to prenatal care. They began with 5 hospitals and scaled up to cover all 42 hospitals and all maternity clinics in the province. The result was a 99 percent reduction in newborns with hypothermia and a reduction in pregnancy-induced hypertension from 44 percent to 6 percent (Berwick 2004). Although the experience of researchers implementing interventions that are based on system redesign in the developing world has been largely positive, it is not clear whether the resources and leadership exist to bring these interventions to scale through country or regional policies. Further evidence is needed concerning the real-world feasibility and cost-effectiveness of system redesign. • Internal enabling environment Creating the right environment for change involves leadership and leadership training, clinicians empowered to make quality improvement decisions, and resources for quality improvement planning activities (Silimper and others 2002). The internal enabling environment in Costa Rica promoted strong leadership that led to the adoption of structural adjustment loans in the early stages of health sector reforms. The loans were used to maintain such public health programs as mother and child nutrition, even though public spending dropped and prices increased dramatically (Peabody 1996). An enabling environment can also be created by teams of individuals, each representing different stakeholder groups (physicians, nurses, staff members, patients, and so forth) or simply by a strong leader with an interest in teamwork and the resources to support a discrete quality improvement function for team members. Interventions Directly Affecting Provider Practice Practitioners are often forced to provide care in uncertain settings. Technical limitations may reduce the ability to diagnose 1298 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 224 or predict outcomes, or they may have only probabilistic knowledge about the efficacy of their proposed treatment for a particular patient. The nature of clinical practice is often solitary, and physicians have few available ways to gauge their clinical acumen and skills. Performance-based feedback, however, can reward high-quality care and increase knowledge about appropriate actions. If the feedback mechanism is effective, it can also serve as the basis for establishing systemwide incentives for improving quality of care. Training with Peer Review Feedback. In Mexico City, physician retraining on treatment of diarrhea, combined with the concurrent creation of a peer-review structure, decreased the use of antibiotics and increased the use of oral rehydration therapy (ORT). These improvements continued to be seen in a follow-up evaluation 18 months later (Gutierrez and others 1994). The approach has been effectively expanded to prescribing practices for rhinopharyngitis among primary care physicians, using an interactive training workshop and a managerial peer-review committee (Perez-Cuevas and others 1996). Performance-Based Remuneration. A potentially powerful instrument for accelerating quality improvements involves making payments directly to providers who meet quality standards that are based on process indicators associated with favorable patient outcomes. Systems that tie performance to remuneration use relatively small incentives—equivalent to 3 to 10 percent of the provider’s total compensation. Performance-based remuneration has been successfully used in the United States to compensate both private and public providers (McBride, Neiman, and Johnson 2000). Examples of performance-based incentives come from developing countries too. The Nicaraguan Ministry of Health has implemented a pilot program in six hospitals that offers an incentive bonus (a maximum average of 17 percent of hospital revenue) for facilities that achieve performance targets that include quality measures (Jack 2003). In Haiti, a performancebased payment scheme was set up for NGOs that provided services to the population. The scheme resulted in all three participating NGOs reaching target immunization coverage rates (Eichler, Auxila, and Pollock 2001). Thus, payment for specified and observable performance (in terms of provider effort, client coverage, or health impact on the population) can be usefully applied to NGOs and private providers. The specific features of performance-based remuneration are crucial. A study evaluating the South African government’s experience in contracting with private organizations to operate district hospitals found no cost savings—in fact, the government was spending more than if it provided the services itself. The contracting may have failed because remuneration was not based on specific process or outcome measures. Instead, the contractor’s obligation, the methods of monitoring performance, and the sanctions for nonperformance were only minimally specified (Broomberg, Masobe, and Mills 1997). High Volume of Care. Evidence exists that a high volume of care by individuals or institutions leads to better health outcomes (Habib and others 2004). Physician experience (learning) and practice (repetition) lead to fewer complications, less resource use, and better quality for a variety of procedures, such as cataract surgery and laparoscopy (Brian and Taylor 2001). More complex procedures, including endarterectomy, cancer surgery,and coronary bypass surgery,have shown similar effects. Volume effects leading to better health outcomes are not confined to surgical procedures (Zgibor and Orchard 2004). Facilities specializing in the care of chronic diseases such as diabetes, myocardial infarction, and heart failure are also associated with better outcomes. Debate exists over how much of the volume effect is due to specialist care. The benefits of highvolume care persist, however, even after controlling for referral and case-mix biases. When carefully trained nonphysicians are substituted for physicians, volume effects persist but can be accomplished at significantly lower costs. In one study, nurse practitioners and physician’s assistants were able to provide high-quality care for common outpatient conditions such as hypertension, diabetes, asthma, otitis media, pharyngitis, and back pain at substantially lower costs than that of physicians (Douglas and others 2004). Performance-Based Professional Recognition. Providers work in a community of peers in which professional status, prestige, and recognition are often as valuable as material rewards. Nonmonetary incentives, such as public recognition or disclosure, administrative privileges, and awards from professional organizations, can promote improvements in quality. Uganda, for example, implemented the Yellow Star Program as part of a broader health services improvement project. This program evaluated health facilities on a quarterly basis, using 35 indicators of technical and interpersonal quality, and awarded a large yellow star to facilities that scored 100 percent in two consecutive quarters. The star was then prominently displayed outside the facility. The Mexican Ministry of Health has implemented a strategy that combines the accreditation and the training strategies discussed earlier with nonmonetary incentives. The National Crusade for Quality in Health Care introduces quality-oriented incentives to health facilities and medical schools. It also includes public recognition in an effort to encourage learning and to change practice. The National Crusade has already generated measurable improvements in the responsiveness of state-level health systems (Secretaría de Salud de Mexico 2003). Both types of policies examined in this section are associated with better quality and better health outcomes—lower Improving the Quality of Care in Developing Countries | 1299 ©2006 The International Bank for Reconstruction and Development / The World Bank 225 premature mortality and avoidable morbidity, increased patient satisfaction, and more health-seeking behaviors. When effective, these policies result in increased coverage rates, better prescribing patterns, and increased adherence to clinical guidelines. They can spell the difference between an individual’s survival or death, between an individual benefiting from the encounter with the health sector or being harmed by it, and between an individual and society rising from poverty or sinking deeper into it. MEASURING QUALITY Improving quality requires that we measure it accurately. The successful outcomes discussed in the previous section rely on the links between policy and changes in clinical practice. Such links, however, can be created and demonstrated only when valid and reliable measures of process are easily understood, inexpensive to obtain, resistant to manipulation, and related to better health outcomes. Measuring Structure Material measures of structure abound. Numerous facilitybased surveys in developing countries have cataloged capital equipment and staffing levels, and financial reports track budgets and expenditures (but rarely production costs). Facility inventories of drugs and supplies are generally available; service utilization figures are routinely reported to national-level authorities. Such measurements, however, are often beside the point. Even when material structural deficiencies are corrected, they are not reliably linked to changes in health outcomes. Measuring the organization and financing of health care is more difficult. Although descriptions of the organization and financing of health systems exist, objective functional assessments of systems (such as patient flows, the patient referral system, or details of the relative pricing of services) are less often available. or obtaining payments, for example) and thus lack crucial clinical details. One prospective study showed that charts identified only 70 percent of items performed during the clinical encounter (Luck and others 2000). In a related analysis, 6.4 percent of the items recorded in the chart were false and had never really occurred. Where resources and infrastructure are sufficient, the electronic medical record (EMR) is becoming a priority for health systems worldwide. EMR technology promotes uniformity, legibility, and communication, which can lead to guideline use and reduce prescription errors. It also holds the promise of managing populations rather than individuals by aggregating patients into groups. However, the EMR has not always lived up to its potential. In many countries, some impressive successes have occurred—as have spectacular failures, costing billions of dollars (McConnell 2004). The great heterogeneity in recordkeeping practices, problems with medical records (both paper and electronic), and costs of trained medical abstractors have led to a search for other reliable ways to measure quality. Direct Observation and Recording of Visits. Direct observation and recording of visits is a commonly used approach in developing countries (Nolan and others 2001). Ethically, the provider and the patient must be informed of the observation or recording, which introduces participation bias because provider behavior may change as a result of being evaluated. In addition, trained observers are costly, and variation between observers is difficult to remedy. Technical advances have mitigated longstanding difficulties in measuring process. Five approaches and their strengths and weaknesses merit consideration: chart abstraction, direct observation and recording of visits, administrative data, standardized patients, and clinical vignettes. Administrative Data. Administrative data, collected for purposes of managing the delivery of care, are available in all but the poorest settings. A data collection system, once established, is ubiquitous and can provide information on charges and many cost inputs. Administrative data, however, lack sufficient clinical detail to be useful in evaluating process. In a 2003 study, an incorrect diagnosis was recorded in the data 30 percent of the time (although the diagnosis was made correctly). Overall, these data reflected the actual clinical diagnosis only 57 percent of the time (Peabody, Luck, Jain, and others 2004). As information systems advance, accuracy problems may be mitigated, although the lack of adequate clinical detail will continue to limit the use of administrative data. Chart Abstraction. Chart abstraction, or review of the medical record, has long been used to measure technical quality. Such familiar quality evaluations as clinical audits, physician report cards, and profiles are based on chart abstraction. The core strength of the medical record is that it is ubiquitous and can generally be obtained after each encounter. Chart reviews, however, suffer from problems of legibility when notes are handwritten. Often they are generated for reasons other than recording the actual events of the clinical visit (legal protection Standardized Patients. Standardized patients can be a gold standard for process measurement (Luck and Peabody 2002). Trained to simulate illness, standardized patients present themselves unannounced into a clinical setting to providers who have previously given their consent to participate in the study. At the conclusion of the visit, the standardized patient reports on the technical and interpersonal elements of process. Standardized patients are reliable over a range of conditions and provide valid measurements that accurately capture Measuring Process 1300 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 226 variation in clinical practice among providers over time. However, they are expensive and useful only for adult conditions and only those conditions that can be simulated. Thus, they are not practical for routinely evaluating quality. Clinical Vignettes. Clinical vignettes were developed explicitly for measuring quality within a group of providers and evaluating quality at the population level. Vignettes are responsive to variation in quality, and providers readily accept them if they are given anonymously (Peabody, Luck, Glassman, and others 2004). More than 20 vignettes have been used in 13 countries around the world. They can be administered on paper, by computer, or over the Internet. Providers are typically presented with several cases. When process is being measured for many providers, each provider is presented with the same case or set of cases, thus eliminating the need for case-mix adjustment. The provider completing the vignette is asked to take a history, do an examination, order the necessary tests, make a diagnosis, and specify a treatment plan. The questions are open ended and include interactive responses that simulate the visit and evaluate the physician’s knowledge. In two separate, prospective validation studies among randomly selected providers, vignettes consistently demonstrated greater predictive validity of process than did the abstracted medical record. Vignettes have been validated against the gold standard of standardized patient visits, and they reflect actual clinical practice, not just physicians’ knowledge. Vignettes have several other advantages. Because exactly the same case can be given to many providers, vignettes are useful for comparison studies. They are also useful for pre- and postevaluations of policy interventions designed to improve quality. Finally, they are inexpensive to administer and straightforward to score, making them particularly useful in developing countries. ECONOMIC BENEFITS AND COSTS OF QUALITY CARE Policy interventions can lead to higher-quality process of care and can rapidly improve a population’s health outcomes, but is quality improvement cost-effective? This section shows that it is. We compare the economic benefits of better quality of care at the individual and population levels with the costs of implementing quality improvement interventions. We then discuss why these interventions not only increase individual and social welfare but also are cost-effective in the long run. Individual Economic Benefits Individuals benefit from better quality of care because they are physically, emotionally, and mentally healthier. These benefits can be quantified subjectively by self-report, objectively by physiological assessments (such as blood pressure), and monetarily by measuring income. Other things being equal, a healthy individual generates more income than one who is often sick. This benefit goes beyond the period of illness. Research on early childhood development has shown that higher-quality prenatal and postnatal care not only decreases mortality but also improves subsequent school performance, which is critical to future labor productivity (Van der Gaag 2000). The monetary benefits of better individual health can be assessed by examining the individuals’ expected income in the context of a life cycle model. Expected income depends on the risk of death at various points in time and the corresponding opportunities for educational attainment. This scenario can be simulated by improving quality and then estimating how much the higher quality lowers mortality and increases education attainment, both of which increase an individual’s future income (see figure 70.2). Social Macroeconomic Benefits Societies that have healthier populations also have higher levels of human capital and a greater capacity to generate wealth. Higher quality of care for the individual increases society’s human capital by reducing both the number of premature deaths (thus increasing the labor force) and the amount of temporary or permanent disability (thus improving worker productivity). Providers and insurers also benefit from lower costs by avoiding unnecessary or inappropriate care. Thus, society benefits from both better health and lower public expenditures for treatment, which can then be reallocated to other productive uses. Interventions that improve quality have an especially high social value when they have large positive externalities (for instance, when better process reduces the incidence of a communicable disease). Sometimes, however, society benefits but some stakeholders do not. For example, physicians who provide better preventive care may experience less demand for their curative services and associated resources. Several attempts have been made to estimate the correlation between health outcomes and long-term economic growth. The high prevalence of such diseases as malaria has been linked in some studies to a slowing of economic growth by one to two percentage points per year. These studies were severely limited by the number of countries and by the many unobserved factors excluded from the models (Sachs 2001). These limitations suggest another way to estimate the benefits of higher quality on health outcomes and long-term economic growth. Because diagnostic accuracy and treatment of malaria can be improved with better-quality care, improving quality should increase national income through reductions in mortality rates. Indeed, cross-country data suggest that a one-year increase in life expectancy is associated with an increase in the gross domestic product (GDP) growth rate of 1 to 4 percentage points (Bloom, Canning, and Sevilla 2001). Our own simulations show Improving the Quality of Care in Developing Countries | 1301 ©2006 The International Bank for Reconstruction and Development / The World Bank 227 Increase in present value expected income (percent) 120 100 80 60 40 9 7 20 5 Effect of 0 nine policy intervention scenarios on quality 3 3 8 1 13 18 23 Unemployment rate (percent) Source: Authors’ calculations. Note: These results model the effect on income of a policy intervention that leads to higher quality. The effect is determined as an increase in the present value of income for varying rates of unemployment. Higher quality is based on nine different scenarios of quality improvement. For each successive scenario, infant mortality rates are reduced by 1 percent and educational attainment is increased by 5 percent. The baseline possibilities are for the Islamic Republic of Iran. Figure 70.2 Economic Benefits of a Quality Intervention That Reduces Child Mortality Rates and Leads to Higher Educational Attainment Accumulated gains in GDP during a 50-year period relative to current GDP 0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0 5.0 4.0 3.0 2.0 0.01 0.015 0.02 0.025 1.0 0.03 0.035 Yearly decline in child mortality rates (percent) 0.04 Change in GDP growth rate from one year increase in life expectancy Source: Authors’ calculations. Figure 70.3 Gains in GDP Resulting from Reductions in Child Mortality Rates that quality improvements can result in as much as a 5 percent annual reduction in child mortality rates, which can generate, over 50 years, economic gains equivalent to 18 percent of current GDP (see figure 70.3). Similar results would be obtained if the effect of better quality on morbidity and disability were simulated. Economic Costs Policies that improve the quality of care have both direct and indirect costs. Direct costs relate to the human and physical resources needed to implement the intervention. Indirect costs come from more subtle changes, including alterations in the quantity of health services provided, in provider demand for 1302 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 228 various inputs (such as equipment and medication), in the market prices of heath care, in government health budgets, and ultimately in the macroeconomy. For interventions at the local level, such as training doctors in a particular region, it is usually sufficient to measure direct costs. Although the level of detail required can be overwhelming when the interventions are complex, the calculations are usually straightforward. The costs of local interventions depend on local prices of such inputs as labor, transportation, training kits, food, space rental, and accommodations. The cost of training providers in the appropriate treatment of childhood illnesses ranges from a low of US$1 to a high of US$430 (Santoso, Suryawati, and Prawaitasari 1996). The direct and indirect costs of interventions at the central or local government level are harder to quantify. Expanding training programs to all public providers, enforcing standards for private and public providers, changing payment systems, and developing policies to protect consumers against malpractice are macro-level interventions that have direct program-level costs. They affect the economy as a whole by changing the allocation of public resources and the relative prices of goods and services. Macroevaluations of health policy interventions are seldom conducted, even though systemwide interventions are likely to have the highest effect on quality and health-related benefits. Cost-Effectiveness of Improved Process Two interventions that vividly illustrate the cost-effectiveness of improvements in clinical practice and outcomes have been chosen: detection and treatment of acute respiratory illnesses and appropriate drug use and treatment for diarrhea. Better Treatment of Pneumonia in Children. Part of the high mortality from childhood pneumonia in the developing world can be explained by poor-quality care, which is defined as the inability either to accurately diagnose or to treat the disease. Our prototype intervention has two cost components: the cost of implementing an educational activity for providers and the cost of treating nonsevere and severe childhood pneumonia. The former component is based on a study and uses conservative high-end cost estimations (Kelley and others 2001); the latter is the midpoint from another study (Stansfield and Shepard 1990). The number of lives saved depends on the effect of the intervention—that is, the change in the percentage of cases diagnosed and treated; the prevalence rate of both types of pneumonia; the population covered by each provider; the case-fatality ratio; and the effectiveness of the treatment. Both the case-fatality ratio and the effectiveness ratio were fixed at middle values suggested by earlier work (Stansfield and Shepard 1990). For the other parameters, a large range of variation was considered, producing 450 scenarios. Finally, six impact levels were considered, which were based on two previous studies (Chakraborty, D’Souza, and Northrup 2000; Mtango and Neuvians 1986). The analysis showed that, under average conditions, improving quality of care for conditions of acute respiratory illness can be very cost-effective. When the baseline quality is low and the disease prevalence is high, an intervention that raises quality has a cost-effectiveness ratio of US$132 to US$800 per life saved; if the policy intervention is ineffective or the prevalence of pneumonia is low, the average cost of saving a life could be more than US$2,000. When 60 percent of cases are already appropriately diagnosed and treated, the costeffectiveness ratio rises to US$5,000 per life saved.1 Better Treatment of Diarrhea. Diarrhea remains one of the leading causes of childhood morbidity and mortality in the developing world. The diarrhea incidence rate among children in resource-constrained countries can reach six to seven episodes per year (Thapar and Sanderson 2004). ORT is the accepted standard of care for acute diarrhea. Unfortunately, a large proportion of cases are still treated with nonrehydration medication, including antibiotics and antidiarrheals. Improved diagnosis of dehydration and reduced use of unnecessary medications, however, lead to better outcomes. Various interventions can make sizable changes in the diagnostic and prescribing patterns of providers. Verbal case review, combined with a package of additional intervention referred to as INFECTOM (Information, Feedback, Contracting with Providers to Adhere to Practice Guidelines, and Ongoing Monitoring), increased the proportion of cases treated correctly from 16 percent to 48 percent (Bloom, Canning, and Sevilla 2001). One study reports that small group, face-to-face interventions reduced antimicrobial prescriptions by 16 percent and antidiarrheal prescriptions by 7 percent among a group of providers treating acute diarrhea in Indonesia (Santoso, Suryawati, and Prawaitasari 1996). The same study showed that formal seminars reduced antimicrobial use by 10 percent and antidiarrheal use by 7 percent. On the basis of these studies, an average cost per intervention was used, ranging from US$25 to US$125. The savings from switching to a less costly treatment (instead of antibiotics, for example) were subtracted from the direct costs that are related to implementing the training activity. Because other savings, such as those related to a lower use of inpatient services, were ignored, the estimates are conservative. Savings could be greater: Two years after an ORT unit was established at the Kamuza Central Hospital in Malawi, 50 percent fewer children with diarrhea were admitted to the pediatric ward, and those admitted required 56 percent less intravenous fluid for rehydration (Martines, Phillips, and Feachem 1990). Again, the number of lives saved depends on the disease prevalence; the effect of the policy on treatment quality; the Improving the Quality of Care in Developing Countries | 1303 ©2006 The International Bank for Reconstruction and Development / The World Bank 229 population covered by each provider; the average case-fatality ratio, which was set at 6 per 1,000 on the basis of Snyder and Merson (1982); and the effectiveness of the treatment. For the latter parameter, reductions in mortality rates following ORT treatment of 40 to 60 percent and reductions in effectiveness ratios of 5 to 100 percent have been reported (Shepard, Brenzel, and Nemeth 1986). Accordingly, the effectiveness ratio was set at 80 percent. As before, alternative values for the other parameters were adopted, generating 450 scenarios. Educational interventions to improve the quality of care for treatment of diarrheal diseases are also highly cost-effective. In general, the cost of saving a life through educational interventions is less than US$500 and could be as low as US$14. Scenarios with high cost per life saved (more than US$6,000) are when prevalence rates are low or when implementation costs for quality-related interventions are high. Although the data available to estimate the costs and benefits of health outcomes and process are limited, these simulations, combined with published reports of successful policy interventions, clearly show the cost-effectiveness of interventions that improve health outcomes through better quality of care. However, reliable measures of quality are necessary to design and evaluate these interventions. Such studies should be complemented by cost-benefit and cost-effectiveness analyses. Sometimes, in public health emergencies, for example, control groups may not be practical or ethical, in which case real-time operations research is an acceptable substitute. In the area of research topics, top priority should be given to quality monitoring and assurance strategies to gain an understanding of exactly what the health system is contributing to society and at what cost. Quantifying the associated costs of different variants of quality monitoring and assurance strategies should also be a high-priority item on the quality research agenda. The second priority should be to increase the evidence base regarding the effects on provider behavior of public policies concerning quality of care and whether they lead to better health outcomes. We need to learn more about the long-term effects of different contracting and remuneration policies on providers’ practices and the consequent results of such policies for health outcomes. Finally, we need to understand how contracting and remuneration policies affect problems unique to the developing world, such as the use of doctor substitutes and the migration of skilled providers to wealthier countries. CONCLUSION RESEARCH AGENDA ON QUALITY Most of the issues discussed throughout this chapter represent important topics for research. Establishing a research agenda requires prioritizing both the type of research and the topics to be studied. Quality-of-care research must also strike a balance between relevance to decision making and excellence in scientific rigor (Frenk 1992). Observational studies are needed to document the extent and correlates of quality at various levels: individual providers, institutional providers, health care systems, and whole populations. Apart from offering much-needed basic descriptions (especially in developing countries), these studies can test specific indicators of the dimensions of quality and can compare the measurement approaches discussed earlier. Intervention studies introduce planned changes into health care settings and assess their consequences. It is fundamentally important that intervention studies compare one provider group or policy alternative with another. In addition, control groups must be used so that any observed change can be attributed to the intervention itself rather than to another source of variation. The external validity of studies is often undermined by the choice of highly specific sites, making it difficult to generalize the findings and to build a body of sound evidence. If randomized trials cannot be conducted, the preferred option is quasi-experimental studies with clear control groups and longitudinal designs (Peabody and others 1999). Good quality means that providers are able to manage an individual’s or a population’s health care by timely, skillful application of medical technology in a culturally sensitive manner within the available resource constraints. Eliminating poor quality involves not only giving better care but also eliminating underprovision of essential clinical services (systemwide microscopy for diagnosing tuberculosis, for example); stopping overuse of some care (prenatal ultrasonography or unnecessary injections, for example); and ending misuse of unneeded services (such as unnecessary hysterectomies or antibiotics for viral infections). A sadly unique feature of quality is that poor quality can obviate all the implied benefits of good access and effective treatment. At its best, poor quality is wasteful—a tragedy in severely resource-constrained health care systems. At its worst, it causes actual harm. Despite the urgency of improving health in developing countries, quality of care has been largely ignored. Both providers and patients agree this must change, but how can this goal be reached? From the information marshaled for this chapter, we can draw five conclusions: • Better quality leads to better health outcomes in developing countries. • Process, the proximate determinant of health outcomes, can be measured in valid and reliable ways, such as clinical vignettes and electronic medical records. 1304 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 230 • Measured in the above ways, the process of care in developing countries is poor. • The process of care can be improved in the short term. • Policies affecting structural conditions, including the actual process of care or the continual design and redesign of the health care system, have been shown to be effective in developing countries. We believe that two broad strategies would help to rapidly improve health care quality in developing countries: • encouraging explicit comparative research on outcomes and process • disseminating empirical findings on quality variation. Encouraging Explicit Comparative Research on Outcomes and Process Comparisons highlighting different outcomes can be compelling. For example, when 30-day mortality rates for coronary artery bypass surgery at various facilities were disclosed in the United States, care started to shift from many low-volume hospitals to high-volume hospitals (Chassin 2002). In developing countries, comparisons show that the insured are more likely to have cesarean sections than are the uninsured (Barros and others 1991). Although critics of comparative analysis are justified in saying that systems and populations vary, such criticism misses an important point: Differences in outcome highlight possibilities that help in the search for the underlying causes of poor quality. Although poor quality may have many causes, one of them is almost always poor clinical practice, which can be remedied. We also favor a league or summary table approach to making comparisons. In this approach, the providers being compared agree on criteria before prospective assessments are done. The data for the comparison should be of the highest quality; the league tables themselves should be easy to interpret; and the findings should be rapidly available (Devers, Pham, and Liu 2004). The league table itself should be set up at the regional, national, and international levels so that a variety of benchmarks are available. Implementing quality comparisons will greatly facilitate the process of policy evaluation and cost-benefit analysis and help indicate directions for future research. Access to accurate, consistent quality-of-care data will compel external funders, such as the World Bank, to build quality assurance into their lending and development programs. As major health programs such as the Global Fund to Fight AIDS, Tuberculosis, and Malaria are scaled up around the world, mechanisms to measure and improve care quality will grow more important. Disseminating Empirical Findings on Quality Variation Public dissemination of information on quality, particularly in low-literacy countries, does not seem to create the individual- based choice market that many have envisaged. Instead, it motivates managers and providers to undertake changes that improve the delivery of care (Schneider and Lieberman 2001). Outside pressure—perceived or real—appears to extend the quality debate beyond traditional boundaries, allowing for innovative collaborations and “out of the box” thinking (Devers, Pham, and Liu 2004). Nongovernmental and private organizations involved in health care delivery should also be required to report basic quality measures, perhaps as a condition for funding, thus ensuring that similar pressure to improve quality is exerted outside the public sector. Public dissemination can create shock waves when poor quality is “discovered,” leading to popular demand to increase quality. For example, findings of widespread medical errors in the United States, estimated to have resulted in as many as 98,000 deaths per year, launched the medical safety revolution (Institue of Medicine 2001). Dissemination among physicians and surgeons by means of report cards and ratings has been effective at changing clinical practice. One advantage of dissemination among providers is that the results can be more refined and technical than ratings meant for wider audiences. Dissemination is the responsibility of public research and public initiative. Because dissemination is inherently controversial, it requires public financing—even more than other public goods (Jamison, Frenk, and Knaul 1998). Ultimately, improving quality is about value. In health care, price is not a reliable proxy for quality and cannot be used as a guide. Because patients and consumers cannot directly observe quality, their ability to demand high-quality services is limited, and they are often left to settle for a market that has suboptimal equilibrium and poor quality of care. In addition, providers often lack knowledge of optimal treatments and technologies and thus are not aware of how they can produce higher-quality care. Because the provider-patient interaction is so private and personal, quality of care is hard to observe and to measure. New measurement tools, however—such as clinical vignettes and the electronic medical record—are being developed and improved. As research links care with outcomes and cost inputs, we can expect to have more accurate and reliable data about clinical practice for use in making quality assessments. Investments in quality, however, must be judged critically as well. When we invest in quality, an investment can be beneficial but can come at a cost. So while quality goes up, value can go up or down—or costs can go up while quality actually goes down or stays the same, thus pushing the value of care down and undermining other efforts to improve quality. Finally, as we showed for acute respiratory illness and diarrhea, quality can go up and costs go down, thus increasing overall value. Examples of this optimal outcome must be actively sought out and reported, because the success of a given investment cannot be known in advance (Berwick 2004). Improving the Quality of Care in Developing Countries | 1305 ©2006 The International Bank for Reconstruction and Development / The World Bank 231 Improving health status does not have to rely solely on macroeconomic growth or other long-term development indicators. Health outcomes can be rapidly improved in the short term by ensuring the appropriateness of the circumstances or setting under which the health care encounter occurs (structural improvement) or by increasing the likelihood that health care providers behave in ways most beneficial to patients under the prevailing circumstances (process improvement). However, this improvement will not occur spontaneously or routinely, despite the best intentions of beneficiaries, providers, and governments. Quality improvement tools and technologies and information on successful quality improvement policies must be consistently shared among developing countries to build local capacity. Funding and incentives must also be consistent with high quality. Finally, the political will to ensure that quality becomes a top priority on the health reform agenda must be sustained. Davis, D. A., M. A. Thomson, A. D. Oxman, and R. B. Haynes. 1995. “Changing Physician Performance: A Systematic Review of the Effect of Continuing Medical Education Strategies.” Journal of the American Medical Association 274 (9): 700–5. Devers, K. J., H. H. Pham, and G. Liu. 2004. “What Is Driving Hospitals’ Patient-Safety Efforts?” Health Affairs 23 (2): 103–16. Donabedian, A. 1980. The Definition of Quality and Approaches to Its Assessment. Ann Arbor, MI: Health Administration Press. ———. 1988. “The Quality of Care: How Can It Be Assessed?” Journal of the American Medical Association 260 (12): 1743–48. Douglas W. R., D. H. Howard, E. R. Becker, E. K. Adams, and M. H. Roberts. 2004. “Use of Midlevel Practitioners to Achieve Labor Cost Savings in the Primary Care Practice of an MCO.” Health Care Economics 39 (3): 607–25. Eichler, R., P. Auxila, and J. Pollock. 2001. “Performance-Based Payment to Improve the Impact of Health Services: Evidence from Haiti.” World Bank Institute Online Journal. Flores, R., J. Robles, and B. R. Burkhalter. 2002. “Distance Education with Tutoring Improves Diarrhea Case Management in Guatemala.” International Journal of Quality in Health Care 14 (Suppl. 1): 47–56. Frenk, J. 1992. “Balancing Relevance and Excellence: Organizational Responses to Link Research with Decision Making.” Social Science and Medicine 35 (11): 1397–404. NOTE 1. As a reference, if the average expected value of a life is close to US$30,000, even the highest cost-effectiveness ratio found in the analysis (US$12,000 per life saved), would imply a cost-benefit ratio below 50 percent, assuming an initial average wage of US$1,000 growing at 2 percent per year, a 5 percent discount rate, and unchanged mortality rates. Gutierrez, G., H. Guiscafre, M. Bronfman, J. Walsh, H. Martinez, and O. Munoz. 1994. “Changing Physician Prescribing Patterns: Evaluation of an Educational Strategy for Acute Diarrhea in Mexico City.” Medical Care 32 (5): 436–46. Habib, M., K. Mandal, C. V. Bunce, and S. G. Fraser. 2004. “The Relation of Volume with Outcome in Pachoemulsification Surgery.” British Journal of Ophthalmology 88: 643–46. REFERENCES Heaton, C. 2000. “External Peer Review in Europe: An Overview from the ExPeRT Project. External Peer Review Techniques.” International Journal of Quality in Health Care 12 (3): 177–82. Barros, F. C., J. P. Vaughan, C. G. Victora, and S. R. Huttly. 1991. “Epidemic of Caesarean Sections in Brazil.” Lancet 338 (8760): 167–69. Institute of Medicine. 2001. Crossing the Quality Chasm. Washington, DC: National Academy Press. Beracochea, E., R. Dickson, P. Freemand, and J. Thomason. 1995. “Case Management Quality Assessment in Rural Areas of Papua New Guinea.” Tropical Doctor 25 (2): 69–74. Jack, W. 2003. “Contracting for Health Services: An Evaluation of Recent Reforms in Nicaragua.” Health Policy and Planning 18 (2): 195–204. Berwick, D. 1989. “Continuous Improvement as an Ideal in Health Care.” New England Journal of Medicine 320 (1): 53–56. ———. 2004. “Lessons from Developing Nations on Improving Health Care.” British Medical Journal 328 (7448): 1124–29. Bhat, R. 1996. “Regulating the Private Health Care Sector: The Case of the Indian Consumer Protection Act.” Health Policy and Planning 11 (3): 265–79. Jamison, D. T., J. Frenk, and F. Knaul. 1998. “International Collective Action in Health: Objectives, Functions, and Rationale.” Lancet 351 (9101): 514–17. Javitt, J., G. Venkataswamy, and A. Sommer. 1983. “The Economic and Social Aspects of Restoring Sight.” In 24th International Congress of Ophthalmology, ed. P. Henkind, New York: J. B. Lippincott. Bloom, D. E., D. Canning, and J. Sevilla. 2001. “The Effect of Health on Economic Growth: Theory and Evidence.” NBER Working Paper 8587, National Bureau of Economic Research, Cambridge, MA. Kelley, E., C. Geslin, S. Djibrina, and M. Boucar. 2001. “Improving Performance with Clinical Standards: The Impact of Feedback on Compliance with the Integrated Management of Childhood Illness Algorithm in Niger, West Africa.” International Journal of Health Planning Management 16 (3): 195–205. Brian, G., and H. Taylor. 2001. “Restoring Sight to the Millions: The Aravind Way.” Bulletin of the World Health Organization 79 (3): 270–71. Kumaranayake, L., P. Mujinja, C. Hongoro, and R. Mpembeni. 2000. “How Do Countries Regulate the Health Sector? Evidence from Tanzania and Zimbabwe.” Health Policy and Planning 15 (4): 357–67. Broomberg, J., P. Masobe, and A. Mills. 1997. “To Purchase or to Provide? The Relative Efficiency of Contracting-Out versus Direct Public Provision of Hospital Services in South Africa.” In Private Health Providers in Developing Countries: Serving the Public Interest?, ed. S. Bennett, B. McPake, and A. Mills. London: Zed Press. Loevinsohn, B. P., E. T. Guerrero, and S. P. Gregorio. 1995. “Improving Primary Health Care through Systematic Supervision: A Controlled Field Trial.” Health Policy and Planning 10 (2): 144–53. Chakraborty, S., S. A. D’Souza, and R. S. Northrup. 2000. “Improving Private Practitioner Care of Sick Children: Testing New Approaches in Rural Bihar.” Health Policy and Planning 15 (4): 400–7. Chassin, M. R. 2002. “Achieving and Sustaining Improved Quality: Lessons from New York State and Cardiac Surgery.” Health Affairs 21 (4): 40. Luck, J., and J. Peabody. 2002. “Using Standardised Patients to Measure Physicians’ Practice: Validation Study Using Audio Recordings.” British Medical Journal 325 (7366): 679. Luck, J., J. Peabody, T. R. Dresselhaus, M. Lee, and P. Glassman. 2000. “How Well Does Chart Abstraction Measure Quality? A Prospective Comparison of Standardized Patients with the Medical Record.” American Journal of Medicine 108 (8): 642–49. 1306 | Disease Control Priorities in Developing Countries | John W. Peabody, Mario M. Taguiwalo, David A. Robalino, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 232 Marek, T., I. Diallo, B. Ndiaye, and J. Rakotosalama. 1999. “Successful Contracting of Prevention Services: Fighting Malnutrition in Senegal and Madagascar.” Health Policy and Planning 14 (4): 382–89. Santoso, B., S. Suryawati, and J. E. Prawaitasari. 1996. “Small Group Intervention vs. Formal Seminar for Improving Appropriate Drug Use.” Social Science and Medicine 42 (8): 1163–68. Martines, J., M. Phillips, and R. G. Feacham. 1993. “Diarrheal Diseases.” In Disease Control Priorities in Developing Countries, ed. D. Jamison, W. H. Mosley, A. R. Measham, and J. L. Bobadilla, New York: Oxford University Press. Schneider, E. C., and T. Lieberman. 2001. “Publicly Disclosed Information about the Quality of Health Care: Response of the U.S. Public.” Quality in Health Care 10 (2): 96–103. McBride, A. B., S. Neiman, and J. Johnson. 2000. “Responsibility-Centered Management: A 10-Year Nursing Assessment.” Journal of Professional Nursing 16 (4): 201–9. McConnell, H. 2004. “International Efforts in Implementing National Health Information Infrastructure and Electronic Health Records.” World Hospitals and Health Services 40 (1): 33–37, 39–40, 50–52. McGlynn, E. A., S. M. Asch, J. Adams, J. Keesey, J. Hicks, A. DeCristofaro, and E. A. Kerr. 2003. “The Quality of Health Care Delivered to Adults in the United States.” New England Journal of Medicine 348 (26): 2635–45. Mtango, F. D., and D. Neuvians. 1986. “Acute Respiratory Infections in Children under Five Years: Control Project in Bagamoyo District, Tanzania.” Transactions of the Royal Society of Tropical Medicine and Hygiene 80 (6): 851–58. Murray, C. J., and J. Frenk. 2000. “A Framework for Assessing the Performance of Health Systems.” Bulletin of the World Health Organization 78 (6): 717–31. Nolan, T., P. Angos, A. J. Cunha, L. Muhe, S. Qazi, E. A. Simoes, and others. 2001. “Quality of Hospital Care for Seriously Ill Children in LessDeveloped Countries.” Lancet 357 (9250): 106–10. Peabody, J. 1996. “Economic Reform and Health Sector Policy: Lessons from Structural Adjustment Programs.” Social Science and Medicine 43 (5): 823–35. Peabody, J., P. Gertler, and A. Liebowitz. 1998. “The Policy Implications of Better Structure and Process on Birth Outcomes in Jamaica.” Health Policy 43 (1): 1–13. Peabody, J., J. Luck, P. Glassman, S. Jain, J. Hansen, M. Spell, and M. Lee. 2004. “Measuring the Quality of Physician Practice by Using Clinical Vignettes: A Prospective Validation Study.” Annals of Internal Medicine 141 (10): 771–80. Peabody, J., J. Luck, S. Jain, D. Bertenthal, and P. Glassman. 2004. “Assessing the Accuracy of Administrative Data in Health Information Systems.” Medical Care 42 (11): 1066–72. Peabody, J., M. Rahman, P. J. Gertler, J. Mann, D. O. Farley, and G. M. Carter. 1999. Policy and Health: Implications for Development in Asia. Cambridge, UK: Cambridge University Press. Peabody, J., F. Tozija, J. Muñoz, R. J. Mordyke, and J. Luck. 2004. “Using Vignettes to Compare the Quality of Care Variation in Economically Divergent Countries.” Health Services Research 39: 1937–56. Perez-Cuevas, R., H. Guiscafre, O. Munoz, H. Reyes, P. Tome, V. Libreros, and G. Gutierrez. 1996. “Improving Physician Prescribing Patterns to Treat Rhinopharyngitis: Intervention Strategies in Two Health Systems of Mexico.” Social Science and Medicine 42 (8): 1185–94. Sachs, J. D. 2001. Macroeconomics and Health: Investing in Health for Economic Development. Report of the Commission on Macroeconomics and Health. Geneva: World Health Organization. Salmon, J., J. Heavens, C. Lombard, and P. Tavrow. 2003. “The Impact of Accreditation on the Quality of Hospital Care: KwaZulu-Natal Province, Republic of South Africa.” Operations Research Results 2 (17). Published for the U.S. Agency for International Development (USAID) by the Quality Assurance Project, University Research Co., Bethesda, MD. Secretaría de Salud de Mexico. 2003. Salud: México 2002. Mexico City: Secretaría de Salud. Shepard, D. S., L. E. Brenzel, and K. T. Nemeth. 1986. “Cost-Effectiveness of Oral Rehydration Therapy for Diarrhoeal Diseases.” PHN Technical Note 86-26, Population, Health, and Nutrition Development, World Bank, Washington, DC. Silimper, D. R., L. M. Franco, T. Veldhuyzen van Zanten, and C. MacAulay. 2002. “A Framework for Institutionalizing Quality Assurance.” International Journal for Quality in Health Care 14 (Suppl. 1): 67–73. Snyder, J. D., and M. H. Merson. 1982. “The Magnitude of the Global Problem of Acute Diarrhoeal Disease: A Review of Active Surveillance Data.” Bulletin of the World Health Organization 60 (4): 605–13. Stansfield, S. K., and D. S. Shepard. 1990. “Acute Respiratory Infection.” In Disease Control Priorities in Developing Countries, ed. D. Jamison, W. H. Mosley, A. R. Measham, and J. L. Bobadilla. New York: Oxford University Press. Supratikto, G., M. E. Wirth, E. Achadi, S. Cohen, and C. Ronsmans. 2002. “A District-Based Audit of the Causes and Circumstances of Maternal Deaths in South Kalimantan, Indonesia.” Bulletin of the World Health Organization 80 (3): 228–34. Tan, S. K. P. 1999. “Safety with Anaesthesia—The Paradigm of Continuous Improvement.” Medical Journal of Malaysia 54 (1): 1–3. Tavrow, P., Y.-M. Kim, and L. Malianga. 2002. “Measuring the Quality of Supervisor-Provider Interactions in Health Care Facilities in Zimbabwe.” International Journal of Quality in Health Care 14 (Suppl. 1): 57–66. Thapar, N., and I. Sanderson. 2004. “Diarrhoea in Children: an Interface between Developing and Developed Countries.” Lancet 363 (9409): 641–53. Thaver, I. H., T. Harpham, B. McPake, and T. Garner. 1998. “Private Practitioners in the Slums of Karachi: What Quality of Care Do They Offer?” Social Science and Medicine 46 (11): 1441–49. Van der Gaag, J. 2000. “From Child Development to Human Development.” In From Early Child Development to Human Development, ed. M. E. Young, Washington, DC: World Bank. Walker, G. J., D. E. Ashley, and R. J. Hayes. 1988. “The Quality of Care Is Related to Death Rates: Hospital Inpatient Management of Infants with Acute Gastroenteritis in Jamaica.” American Journal of Public Health 78 (2): 149–52. Wennberg, J. E., E. S. Fisher, and J. S. Skinner. 2002. “Geography and the Debate over Medicare Reform.” Health Affairs (Millwood) Suppl. Web Exclusives: W96–114. WHO (World Health Organization). 2000. World Health Report 2000 Health Systems: Improving Performance. Geneva: WHO. Zaslavsky, A. M., J. N. Hochheimer, E. C. Schneider, P. D. Cleary, J. J. Seidman, E. A. McGlynn, and others. 2000. “Impact of Sociodemographic Case Mix on the HEDIS Measures of Health Plan Quality.” Medical Care 38 (10): 981–92. Zgibor, J. C., and T. J. Orchard. 2004. “Specialist and Generalist Care for Type 1 Diabetes Mellitus—Differential Impact on Processes and Outcomes.” Disease Management and Health Outcomes 12 (4): 229–38. Improving the Quality of Care in Developing Countries | 1307 ©2006 The International Bank for Reconstruction and Development / The World Bank 233 ©2006 The International Bank for Reconstruction and Development / The World Bank 234 Chapter 71 Health Workers: Building and Motivating the Workforce Charles Hongoro and Charles Normand Policy on human resources for health should support health policy objectives and be a means for achieving policy goals. The implication of such a focus is that health systems development should start by identifying the tasks that must be carried out and the skills needed to perform them. Meeting policy goals depends on being able to recruit, train, and retain staff with the necessary bundles of skills. Traditionally, skills are defined by membership of a profession, especially medicine, nursing, midwifery, and the allied health professions. Low- and middleincome countries (LMICs), often from necessity, have widened the range of health care workers to meet the service needs, with some people trained in extremely basic skills and others receiving enhanced training, such as nurses trained in emergency obstetrics. What is meant by a doctor or nurse also varies. Even though structures and institutions vary widely, some problems are common to most LMICs. First, persuading doctors to work in remote rural areas is difficult, and they typically do not remain long in such posts. Second, emigration of doctors and nurses is extensive. Third, it is common for doctors to work in both the public and the private sectors (referred to as dual practice), sometimes harming public services. Dual practice may encourage doctors to skimp on their public health efforts, to pilfer supplies, and to induce demand for their private services (Bir and Eggleston 2003). Many health sector human resource (HR) problems are predictable from a simple labor market perspective, given the combinations of incentives confronting health care workers and the constraints policy makers face. Experience in LMICs shows how problems have arisen and what policies have succeeded. Economics predicts that employers will employ workers as long as the additional value of their services is at least as great as the cost of employing them, and workers will work if the rewards are of greater value than those accruing to other uses of their time. If key professionals are in short supply, higher salaries will be needed to attract them. Workers will invest in training if they value higher future incomes and more interesting work above the costs of income lost during training and of fees paid for training programs. This chapter focuses on how health systems might build and improve HR capacity. Appropriate HR capacity is critical for the effective implementation of disease control interventions. Salaries account for 50 to 80 percent of health sectors’ recurrent costs (Bach 2000). Table 71.1 shows the number of physicians and nurses per 100,000 population in selected countries. The number of health workers is related to the level of development because of the tight resource constraints facing LMICs and because of supply constraints, often exacerbated by migration of skilled workers (Awases, Gbary, and Chatora 2003) and prevalence of AIDS. In Africa, where the disease burden is high and increasing rapidly, the number of health workers is particularly low. Most African countries export health professionals to highincome countries. A study in six African countries showed that most health workers intend to migrate for higher salaries. In Ghana, 70 percent of 1995 medical graduates had emigrated by 1999 (Awase, Gbary, and Chatora 2003). Pay differentials provide strong incentives to migrate. For example, a junior doctor in the United Kingdom averages a monthly salary of US$3,029 and a registered nurse averages US$1,500, compared with US$300 1309 ©2006 The International Bank for Reconstruction and Development / The World Bank 235 Table 71.1 Numbers of Physicians and Nurses, Selected Countries and Country Groups, 1998 Country Angola Bangladesh Physicians per 100,000 population Nurses per 100,000 population Physiciannurse ratio 5 100 0.05 19 11 1.7 HEALTH CARE PROVISION AND ASSOCIATED HUMAN RESOURCE NEEDS Studies on developing services to meet the Millennium Development Goals emphasize the importance of making health workers with the appropriate skills available and motivating them (Jha and Mills 2002). The problems include lack of technical skills, low motivation, and poor support networks (Kurowski and others 2003). This chapter, therefore, focuses on HR planning, training and professional development, incentives for workers to accept and stay in posts and to deliver services, and alternatives to conventional professional groups. Boliviaa 29 14 2.1 Botswana 20 100 0.2 Brazil 136 44 3.1 Burkina Faso ⬍3 20 0.15 Central African Republic ⬍3 ⬍10 0.3 20 50 0.4 106 94 1.1 Incentives and Motivation 4 5 0.8 57 34 1.7 7 67 0.1 The labor market model outlined earlier provides a framework for analyzing the role of incentives. A health worker will accept a job if the benefits of doing so outweigh the opportunity cost. Improving recruitment and retention requires either offering higher rewards that make alternative employment less attractive or making qualifications less “portable”—that is, less likely to be recognized in other countries. The development of new health professions in many countries is a way of reducing the portability of qualifications, thereby reducing the opportunity cost of jobs at home. Another advantage is that training can be more specific to local health system needs, but ensuring quality and safety are important issues. Health workers will choose to train and increase their skills if the rewards of doing so exceed the cost. In general, the supply of skilled professionals rises as rewards increase, because more will seek training, more will return to the workforce, and fewer will move to other jobs or other countries. Because health workers value both financial and nonfinancial rewards, they will work for lower salaries if other job characteristics are attractive. The causes of health HR problems in developing countries are complex, and attempts to address them must reflect this complexity. Table 71.2 suggests a framework for exploring links between factors at individual, organizational, and health system levels. The framework is inspired by a systems approach, which gives prominence to the roles of and relationships between different component parts in influencing the whole. The individual health worker level serves as a starting point for exploring the determinants of health worker behavior and performance (Kyaddondo and White 2003). Performance here means productivity and quality of services. Individuals respond to individual concerns through coping strategies, such as informal and dual practices, with associated consequences. There are multiple links between individual health worker behavior and organizational and systemic factors. Organizational and system arrangements define Equatorial Guinea India Nepal Pakistan Papua New Guinea b Peru 10 7 1.4 South Africa 20 100 0.2 Sri Lanka 37 103 0.4 Low-income countries 73 132 0.6 Middle-income countries 142 278 0.5 High-income countries 286 750 0.4 Global average 146 334 0.4 Global median 114 233 0.5 Sources: PAHO 2003; Support for Analysis and Research in Africa 2003. a. 1999 data. b. 1996 data. per month for a Ugandan medical officer and US$180 for a registered nurse. Scaling up service provision using current provision models would require large increases in resources and could require a change in strategy by development partners toward supporting recurrent costs (Jha and Mills 2002). The labor market model indicates that higher salaries would be needed to attract additional staff members, so funding would have to increase more than in proportion to the number of staff members employed. Many LMICs pay health workers on civil service scales, which they control to contain overall government spending. This practice further widens the gap between salaries for professionals at home and abroad. Although improved economic performance and increased development assistance may allow some increases in health spending, in most LMICs it is not plausible that such increases would be sufficient to make the necessary skills available without a range of strategies, including better regulation, stronger incentives, and initiatives to make key skills available at lower cost. 1310 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand ©2006 The International Bank for Reconstruction and Development / The World Bank 236 Table 71.2 Framework for Diagnosing HR Issues in the Health Sector Level Issue Individual health worker Internal capacity Knowledge, skills, competencies, attitude Remuneration Salary and perquisites, payment methods Work environment Immediate environment See organizational-level issues Distant environment See systemic-level issues Productivity Outputs or outcomes per given unit of time or per individual or group of individuals Responses to organizational and systemic constraints Low motivation, morale, productivity, and quality of services Informal practices, for example, unofficial fees, dual practice, misuse of public resources, and unethical practices Emergence of unskilled informal practitioners Organizational level Presence or absence of an appropriate physical and operational context—that is, availability of materials and facilities to deliver services of acceptable quality, workload management, and organizational norms and practices (organizational culture, management and leadership styles) Organizational autonomy across key strategic and functional issues Systemic level Level of bureaucracy and decentralization in the health system, and funding and regulatory arrangements Policy context—for example, pro-market policies and diversification of the health care market in the spirit of a publicprivate mix Socioeconomic and political context Medico-legal policy and enforcement Source: Authors. Note: This chapter uses a narrow definition of health system that considers socioeconomic and political issues as part of a wider context. the incentive context for health workers and influence both organizational and individual performance. Therefore, the configuration of the health system must create incentives for appropriate supply and deployment of health workers. HR development experts tend to focus more on problems encountered in the lower tiers of this framework. Political pressure for short-term solutions partly explains why many countries do not address HR problems comprehensively. The wider context can also be important. Good governance at the national level is necessary to make policy interventions at the health system level or below effective. Financial Incentives Most of the comparatively scarce evidence on the relative importance of financial and other incentives for health workers at the individual level comes from developed countries. Two findings emerge from recruitment and turnover studies. First, at extremely low salaries, financial incentives are particularly important (Normand and Thompson 2000). Second, at least half of the variation in turnover can be attributed to financial incentives (Gray and Phillips 1996). These findings leave considerable scope for improving retention using organizational changes, but such changes will be only partially successful if much better financial rewards are available elsewhere. International migration has increased as restrictions on moves to high-income countries have been eased (Bach 2000). Many developed countries have shortages of health professionals and actively recruit from low-income countries, thereby raising the opportunity cost of remaining at home. Health Care Systems’ Responses to Health Worker Issues Health sector reforms have been widespread in recent years, often with international support. These reforms tended to focus more on structures and financing and less on resource issues (Martineau and Buchan 2000). Other government reforms aimed mainly at improving efficiency and reducing the cost of government administration have often had large effects on the health workforce (Adams and Hicks 2000; Corkery 2000). Some changes have attempted to introduce better incentives, such as performance-related pay and renewable contracts, and to remove underperformers and ghost workers. Evidence on the effects of these reforms suggests that more emphasis should have been placed on designing incentives to improve performance and retention and on moving further away from workforce quotas and norms. Using the three levels of analysis, the following sections consider policies, management, and incentives and how they can help match skills to needs. HEALTH CARE STAFF Workforce planning should be dynamic and should link policy goals to staff members’ skills and numbers and to performanceenhancing incentives. Workforce Planning to Meet Policy Goals Several factors make workforce planning in health particularly difficult, including changing needs as service models change, long training time for some professions, and lack of direct government control over the number of professionals being trained, for example, because of the growth of private medical Health Workers: Building and Motivating the Workforce | 1311 ©2006 The International Bank for Reconstruction and Development / The World Bank 237 schools, such as in Bangladesh, or because of people going overseas for training (although Singapore has addressed this problem by restricting the colleges that the government recognizes for registering doctors). The greatest difficulty comes from the unpredictable loss of skilled staff members to private health sector jobs, jobs abroad, and jobs outside health. Thus, a close link exists between HR planning and incentives and regulation. In Ghana, nurse training has often been the only available form of tertiary education for women, and many of those who are trained do not practice. A key to more rational workforce planning is better coordination between health planning and planning for training and education. Powerful interest groups can oppose the expansion of training. Training establishments often oppose change because it may disrupt existing arrangements and threaten current staff members. The development of new professional groups faces particular resistance from existing professional groups, which, quite correctly, perceive the new groups as posing a threat to their interests. For example, some dentists in South Africa expressed concern over training of dental technicians who carry out a wide range of preventive and restorative dentistry at lower fees (Matomela 2004). Models for HR needs are easy to devise, but determining the appropriate model parameters is difficult. For example, health planners must estimate the length of a nursing career— potentially up to 45 years but often much less, especially if nurses are willing to work outside nursing (Phillips and others 1994). Good data are needed on dropout rates from training. HR sections of health ministries are usually poorly resourced, have low status, and work with poor-quality data, and this situation must be changed if planning is to improve. Basic Skills Training and Continuing Skill Development Whereas the quality of basic training of health professionals varies widely in LMICs, the provision of continuing education and development is almost universally inadequate. Hence, skill levels of staff members fall over time. Evidence indicates that good-quality continuing professional development is a positive incentive and helps to retain staff members. Requirements to undertake continuing education can be made a condition of continued professional registration and can thereby provide some guarantee of competence. Good basic education includes development of both professional skills and learning skills. Basic training and continuing development should be planned together. In many cases, the large investment in basic training is lost because of lack of maintenance, so that shifting some resources to updating and renewing skills is efficient. A further challenge is to align the content of training to the skills professionals need. Many training programs in LMICs provide skills oriented toward service needs in developed countries, although there have been attempts to change this balance. For example, more than 25 percent of Malawi’s curriculum for medical students focuses on community health. Given that any training program can cover only a portion of relevant knowledge, focusing on locally relevant topics is increasingly important. Educational reforms in many medical schools in Africa and elsewhere are based on the community-based educational model (Jinadu, Olofeitime, and Oribador 2002). Numbers and Types of Health Professionals Categories of workers result from combinations of previous and current needs, national traditions, interest group pressures, and historical accidents. Doctors, nurses, and some paramedical professions have wide international recognition but vary in definition. Professional traditions and professional bodies bring some safeguards for quality and safety, and at best, professionals champion the needs of patients. Membership of a recognized profession can bring desirable independence from management. However, internationally recognized qualifications make it easy for professionals to migrate to countries offering higher incomes and better careers. Most developing countries have new categories of staff that do not match internationally recognized professions (Buchan and Dal Poz 2003). Examples include nurses with extended training and roles and people working at subnurse levels with training of a few weeks to three years. Bangladesh has family welfare visitors, health assistants, and medical assistants who might elsewhere be classified as nurses or auxiliary nurses; in Uganda, clinical officers have three years of training and work as subdoctors; and nursing aids in Uganda have three months of training. Training is for specific roles without the generic training in conventional professions. Typically, such employees are mobile nationally, but they do not transfer easily across countries. In the labor market model, employers want to employ staff members if their contribution to service provision is of greater value than the cost of their employment. Because those with portable qualifications can work in other countries, salary levels needed to retain workers reflect that possibility. Theory suggests that staff members will develop new skills if such an investment of their time and money produces significantly increased salary or benefits. Many countries cannot fulfill their requirements for health workers, but normally this difficulty reflects salaries that are too low to attract staff. However, raising salaries may make employment of the full complement of staff members unaffordable. Staffing norms serve little useful role if the salaries needed to fill the posts are unaffordable. Decisions about how many people should be employed and in what capacities should be based on the contributions those employees will make and the costs of employing them. Staffing norms can be useful for 1312 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand ©2006 The International Bank for Reconstruction and Development / The World Bank 238 planning, but they require careful analysis of affordability of care, the skills needed, and the way to provide those skills most efficiently. Several countries have of necessity turned to new models of provision using staff skilled in the delivery of key elements of high-priority services, such as immunization and emergency obstetric care. Safety and Effectiveness of New Health Professions Research on the new professions is limited, and much of the material is anecdotal (Buchan and Dal Poz 2003). A growing literature from developed countries indicates that nurses can be safe and effective in place of doctors in primary care (Venning and others 2000). The fear is that the absence of a formal profession and the lack of internationally recognized training could damage quality and safety. This issue is important, but even if new professionals are less safe than doctors, they may be much safer than the absence of a service such as emergency obstetric care. In some countries, new professions play a major part in the provision of services. A good example is Malawi, where clinical officers with extensive training (but much less than that of doctors) are a major resource, carrying out surgical procedures and administering anesthetics as well as providing medical care. In some countries, regulations govern such extended roles (McAuliffe and Henry 1995). Fenton, Whitty, and Reynolds’s (2003) study of emergency cesarean sections carried out by clinical officers in Malawi found that the overall maternal death rate was 1.3 percent, which is high, but much lower than if services had not existed. Perinatal deaths were 13.6 percent. None of the anesthetists was medically qualified, but outcomes were better when these practitioners had received anesthetics training (maternal deaths were 0.9 percent compared with 2.4 percent). The researchers found no significant difference in outcome between medically qualified surgeons and those trained as clinical officers. Care should be taken in interpreting the results of one study, but it does suggest that well-trained clinical officers can safely substitute for doctors in providing some important procedures. Human Resource Policy and New Staff Groups New staff groups are increasingly providing essential services in LMICs. In Zimbabwe, a new cadre called primary health care nurses, whose qualifications are lower than general nurses, was introduced in 2003 to curb external migration by nurses (Chimbari 2003). At the system level, such a development requires regulation and standard setting; at the service provision level, appropriate supervision and management is needed; and at the individual level, incentives and training need to be considered. Employing fully qualified doctors and nurses might be the safest option, but failing to provide services because of staffing constraints is unlikely to be the next best option. HEALTH WORKER INCENTIVES The World Health Report 2000 defines incentives for health workers as “all the rewards and punishments that providers face as a consequence of the organizations in which they work, the institutions under which they operate, and the specific interventions they provide” (WHO 2000, p. 61). Health workers face a hierarchy of incentives or disincentives generated by the work they do, the way they are paid, and the organizational and system context in which they work. Incentives are generally designed to accomplish the following: • • • • • to encourage providers to furnish specific services to encourage cost containment to support staff recruitment and retention to enhance the productivity and quality of services to allow for effective management. Responses of providers to incentives depend on context and on the stage of their career. Incentives that induce productivity vary with experience, stage in a career path, and changes in providers’ social responsibilities. Ideally, incentive structures should recognize the evolutionary nature of work expectations. Typically, incentives vary by type of employer: nongovernmental organization, public, or private. Public sector incentives tend to be the weakest because resource constraints and bureaucratic rules on civil servant employment constrain the use of both financial and nonfinancial incentives. Typology of Incentives Extrinsic incentives can be individual and organizational, monetary and nonmonetary (table 71.3). Discussions of provider behavior in LMICs have focused mainly on financial incentives, partly because of their low income levels compared with industrial countries. The challenge is to establish an optimal mix of financial and nonfinancial incentives that generate the desired behavior of health workers. Experience from vertical programs for priority diseases or services—for example, poliomyelitis, malaria, family planning, and sexually transmitted diseases—provide evidence about different incentives. Programs often offered staff members better pay and incentive packages than those other public health workers received (Beith and others 2001). The exact effects of stronger incentives are unknown, but these programs generally succeeded, as evidenced by the eradication of leprosy, the near eradication of poliomyelitis in many countries, and the large drop in average fertility in developing countries in the 1990s. Successful vertical programs used combinations of incentives, including better salaries, field and transportation allowances, streamlined management, specialized training, better facilities and material resources, and results-oriented Health Workers: Building and Motivating the Workforce | 1313 ©2006 The International Bank for Reconstruction and Development / The World Bank 239 Table 71.3 Typology of Incentives Individual incentives Organizational incentives Environmental incentives Financial Internal Amenities Salary Autonomy Transportation Pensions Accountability Job for spouse Illness, health, accident, and life insurance Market exposure School for children Travel and transport allowances Child care allowance Rural location allowance Heat allowance Financial responsibility External Governance Public finance policy Regulatory mechanisms Retention and professional allowances Subsidized meals, clothing, and accommodation Nonfinancial of health system capacity, which depends in part on the alignment of health workers’ objectives with policy and with system goals. Aligning health worker and system objectives is difficult. The aim is to have satisfied health workers who are motivated to work harder (Hicks and Adams 2001). Evidence is limited, but financial and nonfinancial incentives are mutually reinforcing, and changing the culture of the health system to make goals more readily understood and shared can make financial incentives more powerful. Such change in the organization of health care can be politically sensitive because it can give health sector workers advantages over other public employees. Incentives may have conflicting effects. For example, decentralization might create the autonomy needed for effective management, but without transparent management and career structures and job security, providers might view such a change as a threat (Kyaddondo and White 2003). Getting the balance right requires understanding the socioeconomic and political circumstances and may be helped by using participatory approaches to policy making and implementation. Vacation days Flexible working hours Context Access to training and education Sabbatical and study leave Planned career breaks Occupational health Functional and professional autonomy Technical support and feedback systems Transparent reward systems Valued by the organization Source: Adapted from Zurn 2003. Note: These are mostly extrinsic incentives. management to support improved health worker productivity and program performance. Goals were clearly specified, were understood and shared by the staff, and were often linked to incentives. The choice of vertical structures also reflects the perceived difficulties of using existing health systems, with their excessive bureaucracy, underfunding, and lack of capacity to implement integrated disease control. Vertical programs must eventually be reintegrated into the system. The HIV/AIDS pandemic is a good example of a disease that might require targeted interventions until the capacity of health systems in LMICs improves to a level that allows the disease to be managed like other diseases. The success in integrating vertical programs depends on the parallel development Context is defined here from an individual or an organizational provider’s perspective. It constitutes what Adams and Hicks (2000) refer to as external incentives—that is, methods used by health systems to control the activities of health organizations or funders. The power of incentives depends on context. Health systems in developing countries have varying cultural and economic histories that shape providers’ expectations and responses to incentives. Financial incentives are strong when health workers’ incomes are low, as in most developing countries. Nevertheless, examples of strong nonfinancial incentives exist in countries such as Thailand, where family ties and kinship affect health workers’ decisions on where to work. Such nonfinancial incentives affect the size of the financial incentives needed to change where people choose to work. History and experience determine a country’s working culture and norms. In developing countries, most health systems are large bureaucracies whose management is driven centrally by guidelines, standards, and reporting systems. Incentives in such systems work against innovation, risk taking, and improved efficiency. A possible approach is to introduce changes that are based on the ideas of so-called new public management. New public management replaces line management with contracts or agreements between funders and policy makers on the one hand and providers on the other. Providers are given more managerial autonomy and are controlled by means of contracts and regulation. This approach can more easily embody new financial incentives, and autonomous providers can develop cultures that are more innovating. Such 1314 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand ©2006 The International Bank for Reconstruction and Development / The World Bank 240 a radical change in managerial context can, in principle, make other incentives easier to use. Other dimensions of context are the regulatory framework and its enforcement. Most developing countries have regulations governing the activities of the health sector. These regulations tend to be outdated or poorly enforced (Bloom, Han, and Li 2001). The main reason for regulatory ineffectiveness is low institutional capacity and widespread corruption. The symptoms of regulatory failure are widespread informal activities, dual practice, malpractice and medical negligence, and the presence of unqualified drug sellers (for example, in Bangladesh and Tanzania) and practitioners (as in India) (Bhat 1996; Killingsworth and others 1999; McPake and others 1999). Where the regulatory system is dysfunctional, providers tend to pursue their individual interests, often in private practice, to the detriment of organizational and system performance. Effective incentive systems that are based on performance require regulation and governance structures that minimize the common problems of patronage and corruption (Rasheed 1995). Health system organization factors include governance and the degree of decentralization. Links exist between working culture and norms and the structural aspects of health system organization. The locus of control and decision making play an important part in health worker behavior. In theory, designing incentive schemes that are responsive to health workers’ needs is much easier in a decentralized system. This theory is based on the belief that subnational units are better placed to make effective decisions on funding, regulating, and organizing frontline activities than are centralized units. However, experience in developing countries shows that lack of capacity at subnational levels has constrained decentralization, sometimes leading to unintended effects such as wrong priorities (Bloom, Han, and Li 2001). Any move toward decentralization requires investment in new management skills and capacities. Incentives in Practice Many countries have attempted to reform their economies and health sectors to improve general economic and health system performance. For example, Cambodia, the Arab Republic of Egypt, Uganda, and Zambia have attempted civil service reforms (Corkery 2000). These reforms include attempts to reduce the size of the civil service to lower costs and to improve productivity using incentives such as formal employment contracts and performance-based pay and promotion. Such reforms have been largely unsuccessful in developing countries because of the political difficulties in reducing the size of the civil service. Structural and organizational changes are typically unpopular with labor unions, especially if union members perceive them as threatening their well-being. Experience also underscores the difficulties of aligning system and organizational objectives with individual providers’ objectives (Martineau and Buchan 2000). The effect of incentives can be assessed in terms of their objectives (Adams and Hicks 2000). Table 71.4 summarizes incentive packages used in selected countries. The results shown should be interpreted with caution, because of problems of attribution and poor data. Adams and Hicks (2000) argue that economic incentives in payment mechanisms for physicians conform to economic logic, but little is known about the response of other categories of health workers to such incentives. Experience in Thailand illustrates the labor market model outlined earlier. In general, public doctors prefer to practice in urban areas, where conditions are usually more attractive and opportunities for private practice are better. Thailand pays public doctors who work in rural and remote areas significantly more than those working in urban areas, and this incentive has persuaded some to move (Wibulpolprasert and Pengpaiboon 2003). The government also added nonfinancial incentives, such as changing physicians’ employment status from civil servants to contracted public employees, providing housing, and introducing a system of peer review and recognition. These initiatives were coupled with significant environmental changes, including sustained rural development. In most developing countries, providers in rural areas are paid less than those in cities, and it is hard to recruit and retain health workers in rural areas. China provides another example of how changes in the environment—for example, the introduction of pro-market policies—can change provider behavior, in this case from relying on government salaries alone to the use of “red packages” (Bloom, Han, and Li 2001). These red packages were gifts that were traditionally exchanged as an expression of mutual appreciation, but they have now evolved into informal cash payments from patients to health workers. Health systems have a spectrum of workers with different skills and expectations, and incentives for one group can have negative effects on others (Adams and Hicks 2000). Policy makers must strike a balance between competing interests of professional groups and system goals. The unionization of labor and the growth of professional associations or councils can give health workers considerable bargaining power. Solving one problem can create others. This situation often occurs when governments respond to the grievances of the most vocal professional groups, usually doctors, and neglect other groups. This piecemeal approach has caused HR crises, such as strikes and go-slows. Although health workers are normally somewhat motivated to pursue health policy goals, their own interests can conflict with those goals. Providing higher salaries to health workers, by increasing costs, can reduce access to services by some social groups (Bloom, Han, and Li 2001). Compensation Provider payment systems transfer resources from payers (governments, insurers, and patients) to providers (Maceira 1998) Health Workers: Building and Motivating the Workforce | 1315 ©2006 The International Bank for Reconstruction and Development / The World Bank 241 Table 71.4 Incentive Packages for Health Workers, Selected Countries Complementary measures Objectives Incentives Constraints Results Recruiting and retaining staff in the country Pay competitive salaries Fiscal policies that increase the after-tax marginal value of salaries Budget limitations Helped retain physicians in Bahrain Service standards and controls to prevent reduced work effort in the public system Work effort that may be concentrated in private practice, leading to a deterioration of quality in public practice Considered successful in Bahrain Tolerate informal payments Not applicable Informal charges that limit access and may impede reforms that involve formal user fees and exemptions Resulted in widespread use of informal payments in Eastern and Central Europe, Sub-Saharan Africa, and some East Asian and Pacific countries (Balabanova and McKee 2003; Chakraborty and others 2002; Thompson and Witter 2000) Provide higher salaries or location allowances (Wibulpolprasert and Pengpaiboon 2003) Decentralized administration Overall staff shortages Freedom to allocate institutional revenues or savings from operational efficiency to fund incentives Budget limitations Premium payments for working in rural areas found successful in Thailand (Wibulpolprasert and Pengpaiboon 2003) Include seniority awards in pay scales Allow after-hours private practice in public institutions Recruiting and retaining staff in rural areas Base remuneration on workload Improved infrastructure and staff competence Low public service salaries Policies to reduce salaries as a share of operating costs Professional and lifestyle disadvantages In some countries, resulted in deterioration of public systems where providers also engage in independent private practice (McPake and others 1999) Smaller potential for earnings from private practice than in urban areas Conflicting financial incentives (for example, loss of housing allowance in Bangladesh) Risks posed by internal conflicts and civil wars (for example, Colombia and Uganda) Require service in defined areas as condition of licensing or specialty training Consistent application of policies on transfers and tenure Provide opportunities for government-sponsored further education Loss of confidence if health workers perceive the selection process as arbitrary Providers’ concerns that a temporary posting may become indefinite Aided retention of professionals in Ghana and Zimbabwe (Chimbari 2003) Provide housing and goodquality educational opportunities for health workers’ families Adequate salary Budget limitations Found successful for nurses but not doctors in Nepal Recruit trainees from rural areas Emphasis on public health and family practice in training curricula Traditionally, overrepresentation of urban area students in student populations Found successful in Thailand 1316 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand ©2006 The International Bank for Reconstruction and Development / The World Bank 242 Table 71.4 Continued Objectives Incentives Enhancing the quality and availability of primary care Provide training and promotion opportunities for nurses and medical auxiliaries Complementary measures Clear job descriptions and criteria for promotion Train multifunctional health workers Pay health workers more if they do not practice privately Results Opposition by professional associations to expanded roles for multifunction health workers in Nepal Resulted in successful retraining of health assistants and other health workers in rural areas in Nepal to make them eligible for promotion Limited training capacity in Uganda Mobilize women volunteers from communities, traditional birth assistants, and local leaders Encouraging teaching and research and reducing the internal brain drain Constraints Resulted in regrading of state-certified nurses to state-registered nurses in Zimbabwe (Chimbari 2003; Pannarunothai, Boonpadung, and Kittidilokkul 2001) Allowances perhaps uncompetitive with private practice earnings None In Nepal, found successful in basic medical sciences but resulted in massive resignations in clinical departments Uncommon incentive, although a few countries (for example, Thailand) do pay professional allowances or nonpractice allowances Improving the quality of care Specify clinical guidelines in provider contracts License institutions and professionals based on defined standards Pass laws requiring the registration of drugs and other potentially dangerous substances Leadership role by professional organizations Vested interests of professional associations Uncommon in developing countries Inclusion in the curricula of medical schools Weak peer review systems Some success recorded in Cambodia’s contracting experiment Tradition of professional self-regulation Regulatory capture and a culture of self-protection Acceptance of civil and legal authority Low capacity to enforce laws and regulations Low consumerism and weak advocacy Reduced number of hospitals and unqualified doctors in Estonia Resulted in limited success according to evidence from most developing countries (Bhat 1996) Source: Adapted from Adams and Hicks 2000. and can be structured to provide financial incentives. Most studies focus on payment mechanisms for doctors and their effect on productivity, costs, and quality of services (Bitran and Yip 1998). Table 71.5 summarizes common payment mechanisms and the desired incentives. The evidence shows that the operation of payment mechanisms is sensitive to the payment structure and how it is implemented (Berman and others 1997; Bitran and Yip 1998; Chomitz and others 1998). Payment systems are more successful when built on existing traditions and culture (that is, when they take into account gift systems or, indeed, levels of corruption). It is normally best to use a combination of payment methods. For instance, if there is a shortage of public providers, they might be paid a basic salary for normal working hours and fees for service for after- hours work. This method creates incentives for providers to do extra work and increase throughput, but providers may divert patients to after-hours services, and the method’s feasibility depends in part on monitoring and governance standards. The challenge is to find payment combinations that motivate providers to provide desired volume and quality of services while containing costs. Empirical Evidence on Payment Methods Evidence of provider payment systems that have successfully aligned system and provider incentives is still limited (Bitran and Yip 1998). Interesting findings come from small-scale experiments such as Cambodia’s New Deal (box 71.1). Health Health Workers: Building and Motivating the Workforce | 1317 ©2006 The International Bank for Reconstruction and Development / The World Bank 243 Table 71.5 Major Payment Mechanisms Payment mechanism Key incentives for providers Fees for service Increase the number of cases seen Increase service intensity Provide more expensive services Case payment (for example, diagnosis-related groups) Increase the number of cases seen Daily charge Increase the number of bed days through longer stays or more cases Flat rate (bonus payment) Provide specific bonus services and neglect other services Capitation Attract more patients to register while minimizing the number of contacts with each and minimizing service intensity Salary Reduce the number of patients and the number of services provided Global budget Reduce the number of patients and the number of services provided Introducing financial incentives for health workers is costly. Policy makers in governments and development partners need to ensure that adequate funding is available and sustainable. Resources are also needed to improve working environments and system capacities. Both financial incentives and other incentives are important, but services are likely to improve only if financial incentives are strengthened. Decrease service intensity Provide less expensive services Source: Bennett, McPake, and Mills 1997. workers’ salaries were considered by many to be below the minimum required for a decent life, and workload is increasing because of HIV/AIDS. The Cambodian experiment attempted to align individual health workers’ and system goals through performance-based bonus payments and a set of internal regulations. Regulations can alter the working and organizational culture in a way that allows individual-based incentives to work. There were problems in enforcing penalties for violating regulations. Failure to enforce regulations may lead providers to lose confidence in the system. Countries with limited administrative and institutional capacity should use simple payment mechanisms that are enforceable within their capacity constraints (Barnum, Kutzin, and Saxexian 1995). A lesson from the experiment is that the context matters, and any strategy for offering incentives to workers must be embedded in traditions and cultural practices. In a competitive environment, contracts are a useful tool for aligning health workers’ behavior with organizational and system objectives. In the Cambodian example, contracts between the purchaser and district-level facilities—and between districtlevel facilities and management committees—were an attempt to establish accountability structures that specify targeted activities. More interesting was the attempt to transfer some management risk and responsibilities to individual health workers using subcontracts that permitted management committees to monitor their activities and pay them accordingly, though whether the contracts were well specified is not clear, and the administrative and transaction costs are unknown. The use of contracts requires management and monitoring capacity. Group Incentives Health workers typically work in teams. This system weakens financial incentives because the efforts of individuals may have little influence on overall performance. Indeed, individual incentives can worsen team cooperation. For example, if promotion is competitive and depends on measures of individual productivity, this approach can be a disadvantage for those who work for system goals in cooperative ways. Designing effective group incentives is difficult. Paying group bonuses for achieving a given level of output can work only if individual team members feel adequately rewarded for their efforts and if there is no perceived free-rider problem. Most of the limited evidence on group incentives is for developed countries and shows that much depends on the production process and the organization of the teams (Ratto, Propper, and Burgess 2002). Group financial incentives tend to be weak, and using other approaches such as team building, better sharing of information, and improved working conditions is probably better. Influence of System Capacities and Sustainability Issues on Incentives The theoretical merits and demerits of different incentives are well understood, but system capacities and financial constraints may limit their applicability. Few developing countries have health systems that are capable of effectively implementing and operating some of the payment systems shown in table 71.5. The overall funding for the health sector may be too low to pay providers more. Also, the skills and expertise needed to design and implement contract- and case-based payment methods may be inadequate, and the country may lack the information technology needed to capture relevant data to support such contract- or case-based payment methods. Most health workers in developing countries are civil servants, and the particular needs of health workers may be lost in a general public service. Some countries are considering delinking health workers from public service commissions and setting up independent health commissions to run the health sector. In Zambia, however, delinking failed because of a lack of capacity at both the national and the local levels to implement the necessary HR changes (Martineau and Buchan 2000). Evidence from Trinidad and Tobago suggests that insufficient government commitment impeded the transfer of staff members 1318 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand ©2006 The International Bank for Reconstruction and Development / The World Bank 244 Box 71.1 Cambodia’s New Deal Experiment: The First Year The New Deal experiment in Sotnikum district, Siem Reap province, was launched in 2000 by the Ministry of Health, Médicins Sans Frontiéres, and the United Nations Children’s Fund. It is an example of a concerted attempt to break the vicious circle of underpayment of health staff members and underuse of public health services by tackling the problem of low official income. The New Deal was developed following wide consultations and consensus building, and locally credible management structures were established to monitor and enforce the new framework. Staff motivation was a major problem among health workers, as manifested by high levels of absenteeism from work, low time input at work (an average of one to two hours a day), and poor quality of services. Informal charges, drug thefts, and dual practice by public health workers were common, largely because of their low public salaries: government staff received US$10 to US$12 per month, compared with a minimum of US$100 required for a basic standard of living. At the same time, because of informal charges and extensive use of unregulated private services, households spent more than US$30 per inhabitant per year on health services, equivalent to 11 percent of total household expenditure. The New Deal was seen as a vehicle for improving services by enhancing personal income, and its overall objectives were (a) to improve access to quality health care, (b) to build up the health system, and (c) to act as a catalyst for changes in national health policy. The principle underlying the improvements in the personal income of public health workers was that they would better comply with internal regulations governing (a) job descriptions and working hours; (b) payment of informal fees; (c) misappropriation of drugs, materials, and funds; and (d) diversion of patients to private practice. The district (referral) hospital, health centers, and operational office were each managed by an elected management committee, and individual contracts were signed between staff members and management committees. The contracts stipulated that a bonus would be paid in exchange for strict adherence to internal regulations. The benchmark bonus level was set at an average of US$60 to US$90 per person per month. The management committee was responsible for enforcing the new framework of accountability. Official fees were also introduced on the assumption that the population would agree to pay for better public service. The district got its funding from government appropriations, user fees, and external subsidies from various sources; however, given the overall lack of funding available for the scheme, Médicins Sans Frontiéres and the United Nations Children’s Fund had to provide an initial injection of funds to support the bonus system. At the hospital level, individuals signed contracts with the management committees, and compliance with internal regulations improved. The staff was generally present, fees were transparent, emergencies were attended at night, patients received drugs, and informal payments were not demanded. Use of health services increased significantly after the arrangement had been introduced. The number of documented violations was limited, though problems were encountered in sanctioning penalties. Staff members started receiving bonuses that gradually grew beyond the negotiated maximum, creating a hospital debt crisis by midyear, compounded by understaffing and underemployment problems, which meant that most staff members worked overtime. Nevertheless, the quality of services improved significantly, and per capita family expenditures on health fell 40 percent. Source: Soeters and Griffiths 2003. from the public service, leading to disillusionment among workers and effective opposition from unions (England 2000). In countries with thriving private sectors, devising strong incentives for public sector workers is difficult. For instance, in Uganda, the private not-for-profit sector used to have better working conditions and pay than the public sector and consequently had better staffing levels. The government had to increase public sector salaries significantly in the 1990s to attract health workers back. The use of fees for service in the private sector when public health workers are paid a salary is likely to encourage private practice among public workers. Thus, the effects of methods and levels of payments are influenced by what is happening in the private sector. Optimal Combination of Health Worker Compensation and Incentives Although the optimal mix of provider compensation depends on context and policy objectives, some general policy guidelines on the design of payment methods to achieve organizational Health Workers: Building and Motivating the Workforce | 1319 ©2006 The International Bank for Reconstruction and Development / The World Bank 245 and system goals are available. Linking compensation to performance makes intuitive sense, but care is needed in working out the details. Health workers respond to both financial and nonfinancial incentives, but the extent of the effect varies, and the two can interact. For new payment systems to work well, health workers must be governed by effective managerial authority. Because new payment systems aim to encourage particular behaviors and hold providers accountable, clear responsibility must be delineated within provider organizations. This delineation may be easier to achieve if the management of providers has some autonomy. Evidence from developing countries that have attempted to introduce managerial autonomy and corporatization of health service institutions, such as public hospitals and medical stores, indicates that delinking health workers from government control is politically sensitive. Nevertheless, such organizational or system changes are desirable if new payment methods are to create the right incentives and achieve the desired changes. Part of the context for incentive systems is what type of disease control activities are best provided through markets or hierarchies. Traditionally, the public sector has been dominant. The economic arguments for government involvement are well understood, but delivery of services within the framework of government policy objectives can be by private (both for-profit and not-for-profit) providers. Thus, the private sector is increasingly involved in the social marketing of condoms and bednets, franchising, and contracting (Bennett, McPake, and Mills 1997). From an economic viewpoint the only issues are the cost, quality, and sustainability of such arrangements. Emerging evidence on private sector involvement in health services suggests that the private sector is willing to participate in nonclinical disease control activities if the incentive structure is right. Private not-for-profit providers, such as hospitals and clinics associated with churches, have traditionally complemented government health care activities, especially in poor and peripheral populations (Gilson and others 1997). In recent years, Bangladesh has experimented with contracting nongovernmental organizations to provide primary care services in urban areas. Lessons from this experience are still emerging and indicate that, despite many early mistakes, this form of provision can be innovative and can help make a break from bureaucratic traditions. Such contracting depends on having contracting skills in both parties to the contract. A good understanding of context and incentives is also crucial. In summary, incentive or payment packages should attempt to link payment with individual or group performance and should be assisted by supportive organizational and system changes if the desired provider behavior is to be achieved. No single best combination of payment methods exists. ADVICE FOR GOVERNMENTS Governments in developing countries face huge challenges in strengthening their health systems, especially their HR capacity, if cost-effective disease control interventions are to achieve their desired results. Strengthening their systems will entail developing self-sustaining systems for the supply, use, and retention of health workers. The following considerations are important in relation to putting effective policies and incentive structures in place: • Countries should explicitly link the planned number of each category of staff members to health policy goals and set priorities, taking overall resources into account when planning HR needs. • Countries should recognize that the salaries necessary to recruit and retain staff members will depend on the opportunities such workers have for other employment within the country and abroad, and planned numbers in each category should be based on this reality. • Countries should understand both that qualifications that are recognized internationally are likely to attract higher salaries and that such qualifications may only be partially suited to the needs of essential health services in LMICs. They should focus on developing the most important skills by training new types of health workers, taking into account evidence that use of such health workers can be safe when properly trained. Many countries will be unable to prevent the loss of professionals with portable qualifications, because salaries offered will be far below those available elsewhere. • Countries’ training policies should take into account the decline in skills over time and the need to allocate scarce resources between basic training and continuing staff development. • Countries should adapt and not imitate compensation and incentive structures, given the evidence that effective incentive structures depend on local conditions and traditions as well as on universal principles. • Policy makers should remember that the availability and cost of suitably qualified human resources will affect feasibility and cost-effectiveness of disease control interventions. • When developing vertical disease control programs, program managers must avoid introducing powerful incentives that damage existing services by drawing away key personnel. • Policy makers should identify potentially harmful, unintended consequences when designing regulation and incentive systems. For example, if doctors are allowed to practice in both public and private services, the effects of private practice on incentives in public practice tend to be negative unless carefully monitored. • Countries should recognize that the use of incentives to improve performance normally requires good regulatory frameworks and skilled managerial resources. 1320 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand ©2006 The International Bank for Reconstruction and Development / The World Bank 246 RESEARCH AND DEVELOPMENT AGENDA New staff categories are emerging in many LMICs, and these workers are an important part of the workforce. Such staff categories are likely to increase, given migration and the high cost of employing people with portable qualifications, but little research is available on the appropriateness and safety of the new sets of skills, and little is known about the range of new professions, the content of and approach to training, the extent of professional supervision, and the outcomes of treatment. Sharing experience of such staff categories would be valuable. Priorities, therefore, include a study to map the different new staff groups in health systems in LMICs and to classify their tasks, roles, and training, and studies to compare the outcomes of conventional and new staff groups. In addition to gaining a better understanding of the patterns, roles, and performance of new staff groups, data are needed on the length of time such workers remain in their posts, the extent to which their new qualifications are portable, and their migration patterns. Information is also lacking on how best to provide professional supervision for these new staff groups and how to encourage such employees to be professional in their work. Limited evidence is available on the relationship of different health care compensation methods to individual and organizational behavior in developing countries. The following are possible research areas (and some practical steps) that might help fill information gaps and further understanding of the role of health worker compensation and incentives in disease control in developing countries: • Databases. A useful step would be to set up HR databases for developing countries as the Pan American Health Organization has done for its region. • Literature review. A review of unpublished materials on countries’ experiences with using different payment and compensation mechanisms at national or subnational levels would also be useful. Failed experiments are seldom published, but they provide useful lessons. • HR supply. Traditional HR planning models are no longer effective in handling health system dynamics in developing countries. More research is required to develop HR models in health that include the effects of HIV/AIDS, migration, scaling-up of existing interventions, new technology, and reforms. The underlying question should be how HR supply mechanisms can meet health systems’ needs in terms of numbers, knowledge, skills mix, and competencies. • Demand and utilization. Getting the size of the health workforce right is important in its own right, but that alone is insufficient for improving health workers’ motivation and productivity. Research needs to focus on how to improve the motivation and performance of health workers in resource-constrained environments and on what is needed to retain professionals in such settings. We know little about how health care workers make decisions about a range of incentives and disincentives generated by organizations and the systems in which they work. For example, what does it take to convince doctors and nurses to work in rural and remote parts of a country? To what extent are financial and nonfinancial incentives important in attracting people into training as health workers, deploying them to needy areas, motivating them, and retaining them in the system? To a significant extent, current problems in improving access to care, in widening the range of effective services that are provided, and in improving the quality of care depend on better matches of skills to needs, better motivation of staff, and clearer understanding of how improved structures and incentives will work. Perhaps as important is that much of the debate focuses on developments within traditional patterns of staffing of services, but new patterns are increasingly emerging, and the extent of evaluative research is inadequate for drawing strong conclusions on how such developments can alleviate the constraints facing health systems. The development of incentive systems should be coupled with the development of organizational and institutional capacity that supports sustainable HR development in general. REFERENCES Adams, O., and V. Hicks. 2000. “Pay and Non-pay Incentives, Performance, and Motivation.” Paper prepared for the World Health Organization’s December 2000 Global Health Workforce Strategy Group, World Health Organization, Department of Organization of Health Services Delivery, Geneva. Awases, M., A. Gbary, and R. Chatora. 2003. Migration of Health Professionals in Six Countries: A Synthesis Report. Brazzaville: World Health Organization, Regional Office for Africa. Bach, S. 2000. “Human Resources and New Approaches to Public Sector Management: Improving Human Resource Management (HRM) Capacity.” Paper prepared for the World Health Organization’s December 2000 Global Health Workforce Strategy Group, World Health Organization, Department of Organization of Health Services Delivery, Geneva. Balabanova, D., and M. McKee. 2003. “Understanding Informal Payments for Health Care: The Example of Bulgaria.” Health Policy 62 (3): 243–73. Barnum, H., J. Kutzin, and H. Saxexian. 1995. “Incentives and Provider Payment Methods.” International Journal of Health Planning and Management 10 (1): 23–45. Beith, A., R. Eichler, J. Sanderson, and D. Weil. 2001. “Can Incentives and Enablers Improve Performance of Tuberculosis Control Programs? Analytical Framework, Catalogue of Experiences and Literature Review.” Unpublished paper, Management Sciences for Health, Rational Pharmaceutical Management Project Plus, and Stop Tuberculosis Partnership, November, 2004. Bennett, S., P. McPake, and A. Mills, eds. 1997. Private Health Providers in Developing Countries: Serving the Public Interest. London: Zed Books. Health Workers: Building and Motivating the Workforce | 1321 ©2006 The International Bank for Reconstruction and Development / The World Bank 247 Berman, P. A., A. K. Nandakumar, J. J. Frere, H. Salah, M. El-Edawy, S. El-Saharty, and N. Nassar. 1997. A Reform Strategy for Primary Care in Egypt. Technical Report 9. Bethesda, MD: Partnership for Health Research, Abt Associates. Kyaddondo, D., and S. R. White. 2003. “Working in a Decentralized System: A Threat to Health Workers’ Respect and Survival in Uganda.” International Journal of Health Planning and Management 18 (4): 329–42. Bhat, R. 1996. “Regulation of the Private Health Sector in India.” International Journal of Health Planning and Management 11 (3): 253–74. Maceira, D. 1998. “Provider Payment Mechanisms in Health Care: Incentives, Outcomes, and Organizational Impact in Developing Countries.” Major Applied Research 2, Working Paper 2, Partnership for Health Reform, Abt Associates, Bethesda, MD. Bir, A., and K. Eggleston. 2003. “Physician Dual Practice: Access Enhancement or Demand Inducement.” Working Paper, Tufts University, Department of Economics, Medford, MA. Bitran, R., and W. C. Yip. 1998. “A Review of Provider Payment Reform in Selected Countries in Asia and Latin America.” Major Applied Research 2, Working Paper 1, Partnership for Health Reform, Abt Associates, Bethesda, MD. Martineau, T., and J. Buchan. 2000. Three Diverse Case Studies on the Importance of Human Resources to Successful Health System Reforms. Washington, DC: American Public Health Association. Matomela, N. 2004. “Department Outlines Amended Dental Act.” BuaNews Pretoria, October 7. Bloom, G., L. Han, and X. Li. 2001. “How Health Workers Earn a Living in China.” Human Resources for Development Journal 5 (1–3): 25–38. McAuliffe, M. S., and B. Henry. 1995. Nurse Anaesthesia Worldwide: An Analysis of Practice, Education, and Legislation. Geneva: World Health Organization. Buchan, J. M. D., and M. R. Dal Poz. 2003. “Role Definition, Skill Mix, Multi-Skilling, and ‘New Workers.’” In Towards a Global Workforce Strategy: Studies in Health Services Organization and Policy, vol. 21, ed. P. Ferriho and M. Dal Poz, 275–300. Antwerp, Belgium: ITG Press. McPake, B., D. Asiimwe, F. Mwesigye, A. Turinde, M. Ofumbi, L. Ortenblad, and P. Streefland. 1999. “Survival Strategies of Public Health Workers in Uganda: Implications for Quality and Accessibility of Care.” Social Science and Medicine 49 (7): 849–65. Chakraborty, S., R. Gatti, J. Klugman, and G. Gray-Molina. 2002. “When Is ‘Free’ Not So Free? Informal Payments for Basic Health Services in Bolivia.” Unpublished paper, World Bank, Washington, DC. Normand, C., and C. Thompson. 2000. “Review of the Primary Care Rehabilitation Project in Azerbaijan.” Report prepared for the United Nations Children’s Fund. Chimbari, M. J. 2003. “A Report on Health Care Providers in Zimbabwe.” Disease Control Priorities Project Working Paper. PAHO (Pan American Health Organization). 2003. Bangladesh Health Labor Market Study. Washington, DC: PAHO. Corkery, J. 2000. Public Service Reforms and Their Impact on Health Sector Personnel in Uganda. Geneva: International Labour Organization and World Health Organization. Pannarunothai, S., D. Boonpadung, and S. Kittidilokkul. 2001. “Paying Health Personnel in the Government Sector by Fee-for-Service: A Challenge to Productivity and Quality, and a Moral Hazard.” Human Resources for Health Development (electronic journal) 1 (2) England, R. 2000. “Health Sector Reform Experiences: Lessons for Belize from Trinidad and Tobago?” Issues Note, Institute for Health Sector Development, London. Fenton, P. M., C. J. Whitty, and F. Reynolds. 2003. “Caesarean Section in Malawi: Prospective Study of Early Maternal and Perinatal Mortality.” British Medical Journal 327 (7415): 587. Gilson, L., J. Adusei, D. Arhin, C. Hongoro, P. Mujinja, and K. Sagoe. 1997. “Should Governments Contract out Clinical Health Services to Church Hospitals?” In Private Health Provider in Developing Countries: Serving the Public Interest, ed. S. Bennett, B. McPake, and A. Mills, 276–302. London: Zed Books. Gray, A., and V. L. Phillips. 1996. “Labour Turnover in the British National Health Service: A Local Labour Market Analysis.” Health Policy 36 (3): 273–89. Hicks, C., and O. Adams. 2001.“Pay and Non-pay Incentives, Performance, and Motivation.” Paper prepared for the World Health Organization’s December 2001 Global Health Workforce Strategy Group, World Health Organization, Geneva. Jha, P., and A. Mills. 2002. Improving Health Outcomes of the Poor: Report on Working Group 5 of the Commission of Macroeconomics and Health. Geneva: World Health Organization. Jinadu, M. K., E. O. Olofeitime, and P. Oribador. 2002. “Evaluation of an Innovative Approach to Community-Based Medical Undergraduate Education in Nigeria.” Education for Health 15 (2): 139–48. Killingsworth, J. R., N. Hossain, Y. Hedrick-Wong, S. D. Thomas, A. Rahman, and T. Begum. 1999. “Unofficial Fees in Bangladesh: Price, Equity, and Organizational Issues.” Health Policy and Planning 14 (2): 152–63. Kurowski, C., K. Wyss, S. Abdulla, N. Yémadji, and A. Mills. 2003. “Human Resources for Health: Requirements and Availability in the Context of Scaling-Up Priority Interventions in Low-Income Countries—Case Studies from Tanzania and Chad.” London School of Hygiene and Tropical Medicine, London. Phillips, V., A. M. Gray, D. Hermans, and C. Normand. 1994. “Health and Social Service Manpower in the U.K.: A Review of the Research 1986–1992.” Public Health and Policy Department Publication 7, London School of Hygiene and Tropical Medicine, London. Rasheed, S. 1995. “Ethics and Accountability in the African Civil Service.” DPMN Bulletin 3 (1): 12–14. Ratto, M., C. Propper, and S. Burgess. 2002. “Using Financial Incentives to Promote Teamwork in Health Care.” Journal of Health Services Research Policy 7 (2): 69–70. Soeters, S., and S. Griffiths. 2003. “Improving Government Services through Contract Management: A Case from Cambodia.” Health Policy and Planning 18 (1): 74–83. Support for Analysis and Research in Africa. 2003. “The Health Sector Human Resources Crisis: An Issues Paper.” Academy for Educational Development, Washington, DC. Thompson, R., and S. Witter. 2000. “Informal Payments in Transitional Economies: Implications for Health Sector Reform.” International Journal of Health Planning and Management 15 (3): 169–87. Venning, P., A. Durie, M. Roland, C. Roberts, and B. Leese. 2000. “Randomised Controlled Trial Comparing Cost-Effectiveness of General Practitioners and Nurse Practitioners in Primary Care.” British Medical Journal 320 (7241): 1048–53. WHO (World Health Organization). 2000. The World Health Report 2000—Health Systems: Improving Performance. Geneva: WHO. Wibulpolprasert, S., and P. Pengpaiboon. 2003. “Integrated Strategies to Tackle the Inequitable Distribution of Doctors in Thailand: Four Decades of Experience.” Human Resources for Health 1: 12. Zurn, P. 2003. “Incentives for Human Resource Management.” Paper presented at the Workshop on Human Resource for Health Development: The Joint Learning Initiative, Veyrier-du-Lac, France, May 8–10. 1322 | Disease Control Priorities in Developing Countries | Charles Hongoro and Charles Normand ©2006 The International Bank for Reconstruction and Development / The World Bank 248 Chapter 73 Strategic Management of Clinical Services Alexander S. Preker, Martin McKee, Andrew Mitchell, and Suwit Wilbulpolprasert Financial resources alone are insufficient for individuals to benefit from the opportunities presented by modern health care systems. Some countries have achieved much better levels of health than would be expected given their financial resources (Mehotra 2000); many examples of poor-quality care in countries at all levels of development reflect not only scarce resources but also inadequate management of what resources are available (see chapter 70). Many inputs must come together at the appropriate time and in the appropriate place to achieve maximum health gain. These inputs include human resources (in particular, trained staff); physical resources (such as pharmaceuticals and technology); and intellectual resources (in the form of evidence and the ability to apply it appropriately). This congruence requires that the production, distribution, and combination of these resources be actively managed and that the relations among the various elements that contribute to health care be optimized. The challenge of bringing these diverse inputs together is becoming increasingly complex. Until the 1950s, providing basic care at low cost to large populations was relatively straightforward, given political will and sufficient resources. Relatively few effective drugs were available; even fewer drugs were effective in managing chronic disease. The available technology was limited to simple x-ray machines and chemistry tests that required few skills to administer. Consequently, scaling up the delivery of basic health care was relatively easy. The situation in the former Soviet Union illustrates this state of affairs. Beginning in the 1930s, the Soviet Union implemented a vast system to provide basic health care where almost none had existed. The very simple care available was sufficient to produce significant reductions in maternal and childhood mortality rates. What such a system could not do, however, was respond effectively to the possibilities opened up by the explosion in diagnostic and therapeutic knowledge that began in the mid 1960s with the availability of new and easily tolerated treatments for many common chronic disorders. As individuals became able to survive with their chronic diseases, they aged and acquired other conditions, many of which could now be treated effectively, constantly increasing the complexity of the health care required. The inability to manage this increased complexity resulted in persistently high mortality rates from treatable conditions at a time when corresponding mortality rates were falling in Western countries (Andreev and others 2003). This situation has certain parallels with that faced by many low- and middle-income countries today. Through a variety of mechanisms, a political commitment to the health sector is manifest in the increased availability of funding, such as through the Global Fund for AIDS, Tuberculosis, and Malaria (http://www.theglobalfund.org/en/). Much discussion has focused on one of the elements of health care that these initiatives will support: the supply of drugs that target the microbiological agents responsible for these three diseases. Yet improved outcomes will be achieved only if such agents are linked to the many other elements required to diagnose and treat these patients. Most obviously, the supply of drugs such as antiretroviral agents must be coupled with those used to treat the opportunistic infections that exacerbate AIDS. Care for the acute episode of illness should be linked to general support for patients and family members as well as to activities designed to prevent further spread of the disease. Furthermore, as drugs to combat infections become more widely available, it is 1339 ©2006 The International Bank for Reconstruction and Development / The World Bank 249 probable—unless highly developed prevention systems have been put in place—that drug resistance will increase; this outcome has been evidenced with tuberculosis in those parts of the world where treatment has been available but poorly managed, such as the former Soviet Union and Peru (Farmer, Reichman, and Iseman 1999). The resulting resistant infections are much more complicated and expensive to treat. The rise in antibiotic resistance provides one of the most graphic examples of the consequence of the failure to manage the delivery of health care (see chapter 55). Yet even where the financial resources and the political will exist to deliver effective health care, many health care systems contain numerous constraints to success (Hanson and others 2003): • At the first level, that of the community or household, there may be inadequate demand for services or physical, financial, or social obstacles to their use. This situation calls for action to increase access and affordability, including health care financing reforms (see chapter 12). It also requires policies to ensure that services are culturally appropriate, that they address the particular needs of underserved populations, and that they provide dignity and privacy. Moreover, services should be physically accessible, both in terms of distance from population settlements and in terms of their construction—that is, facilities must be responsive to the needs of persons with disabilities. • At the second level, the delivery of health care, there may be a shortage of resources, such as staff members, drugs, and equipment. However, to bring these resources together would require actions at the third level that anticipate future needs, as well as actions that ensure that the needed drugs and equipment are purchased at the best price possible, are subject to appropriate quality controls, and are distributed where needed. • The third level includes health sector policy and strategic management. Effective action may be constrained by weak systems of management that are unable to take into account the changing health needs of the population and the changing demands on health care providers. Management weaknesses include inadequate pharmaceutical regulation and supply, ineffective training of health professionals, inability to engage with civil society, and a failure to put in place incentive systems to facilitate effective health care. Constraints at this level may originate outside the country, as governments are faced with demands by donors to follow paths that either undermine their policy goals or remove the flexibility needed to achieve them. Constraints acting at this level also arise when policies in other areas affect the health sector, such as when a weak, overly bureaucratic, and unreformed civil service system implements obsolete regulations; when there are inadequacies in infrastructure, such as poor communication and transportation links; or when there are weaknesses in the banking system. • The fourth level refers to the environmental and contextual constraints on effective policies. The delivery of effective care may be affected by the physical environment, including climate and population dispersion. However, an equally important constraint is weak governance working within unsupportive policy frameworks, which may be compromised further by corruption, weak rule of law, political instability, weak public accountability, and lack of a free press. For example, de Soto (2000) has shown that, in many middle-income countries, it is almost impossible even to create a simple garment repair business because of a failure of legislative reform, in particular a lack of clearly defined property rights. As a result, much of the economic activity in those countries is informal or even marginally illegal, a response that is of particular concern in health care, given the scope for unlicensed and incompetent providers to endanger the public. This framework underscores the importance of coordinating action at multiple levels. Health services can operate effectively only if policies are in place at the community level to ensure that those in need have access to services, and only if policies are in place at higher levels to ensure that the resources are available to provide those services. This analysis of different levels demonstrates the importance of taking a systemwide approach to the management of health services. However, because of limited space, this chapter focuses primarily on the third level, that of strategic management. It first examines the nature of management in general and the specificities of management in the health system. It then explores some issues that arise when managing health services in different settings. It concludes with an exploration of some of the strategies used in low- and middle-income countries to optimize the delivery of care, using a framework developed by Oliveira-Cruz, Hanson, and Mills (2003). WHAT IS MANAGEMENT? One of the earliest definitions of management was that of French mining engineer Henri Fayol. Writing at the beginning of the 20th century, he stated, “To manage is to forecast and plan, to organise, to command, to coordinate, and to control” (Fayol 1949). Put simply, managing is about assessing probable future scenarios, deciding how best to respond to them, bringing together the resources needed for that response, and deploying them as effectively as possible. Until relatively recently, most management research was concerned with industrial production, for which outputs could be measured relatively easily. Relatively less attention was given to management of 1340 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 250 Box 73.1 Specificities of Health Care Organizations Health care services differ from many other organizations in many ways: • Defining and measuring outputs is difficult. • The work involved is more variable and more complex than in many other organizations. • Much of the work is of an emergency nature and cannot easily be deferred. • The consequences of error can be severe. • Activities by different groups of staff members are highly interdependent, requiring a high level of coordination. • The work involves a high degree of specialization. • Workers are highly professional, with a primary loyalty to the profession rather than to the organization. • There is limited scope for effective organizational or managerial control over clinicians, the group most responsible for generating work and expenditure. • Dual lines of responsibility often create problems of coordination, accountability, and confusion of roles. Source: Shortell and Kaluzny 1983. service industries in general and health care services in particular. As Shortell and Kaluzny (1983) noted, health care services are different from many other organizations. Of course, many of the specificities shown in box 73.1 are differences of degree, with health services sharing many features with other service organizations. Yet important differences exist. Managing Health Care Services During the 1970s, health care services in many countries faced growing criticism for their perceived failure to articulate explicit goals or to develop the means to achieve them (Enthoven 1985; Griffiths 1984). This failure was contrasted with the perceived success of the private sector, which was seen as more capable of innovation and more responsive to demand. These developments gave rise to what has been termed new public management (Hood 1991), which is characterized by the following: • greater role for professional management in the public sector • closer scrutiny of the work of professionals, involving performance measurement and target setting • link between resource allocation and measurable outputs • “unbundling” of previously integrated units, with contracting for previously integrated services • shift to competition as a key to reducing costs and an emphasis on a private-sector management style • careful use of resources to drive down the cost of labor and other inputs, where possible. Recognizing the many reasons for market failure in health care (Arrow 1963), new public management builds on several concepts that arise from new institutional economics. Included in these is contestability (Baumol, Panzar, and Willig 1982), in which the benefits that competition is thought to bring can arise even when competition is absent, thus ensuring that the barriers to market entry are sufficiently low to allow other providers to emerge. User choice is given priority over most other goals, including equity. The enthusiasm for new public management was largely ideological, reflecting the contemporary rejection of an expanded role for the state; the extent to which this model was actually able to achieve what was claimed for it remains highly contested (Le Grand and Bartlett 1993; Stewart 1998). In particular, critics drew attention to the high transaction costs involved (Evans 1997) and the lack of evidence that competition can actually bring about the intended improvements in quality of care (Maynard 1998). One feature of the new public management is its emphasis on general management, with managers possessing skills and expertise that can be applied to any sector. These managerial attributes are considered to be of greater importance than technical or professional knowledge. As a consequence, in some countries, the balance of power has begun to shift away from health professionals and toward general managers. In many places, the initial enthusiasm has given way to disillusionment and subsequently to a more balanced view that, though the precise solution will reflect the particular circumstances of the health care system, what is needed is a partnership between these two groups. In some countries, this development will mean that managers must assume a greater role in relation to the delivery of clinical care. Such an expanded role will extend from their traditional responsibilities, such as financial controls, hotel services, and payroll management, to active participation in Strategic Management of Clinical Services | 1341 ©2006 The International Bank for Reconstruction and Development / The World Bank 251 setting and monitoring standards for care delivery, linked to a responsibility for ensuring that the resources needed for care delivery are available. In other countries, this role may involve stepping back a little. In an analysis of the British National Health Service, in which the degree of managerial control over the delivery of health care has proceeded further than in many other industrial countries, Harrison and Pollitt (1994) note how clinical decision making is increasingly driven by diagnostic and treatment protocols. Although often developed locally, working arrangements are increasingly specified, with the introduction of timetabled job plans for medical specialists and much greater measurement of outcomes. Yet Harrison and Pollitt argue that the growth of managerial control over professional activity is likely to be constrained by the increasing involvement of professionals in management, even if they do not fully adopt the managerial agenda. A further constraint is the persisting ability of professionals, because of their specialized knowledge, to resist managerial control and the related unwillingness of lay managers to extend their control into certain areas in which they do not feel competent. Managing for Improved Quality of Care An increasing volume of research in industrial countries has focused specifically on managerial and organizational responses to evidence that health systems often deliver suboptimal care (Institute of Medicine 2001). Quality of care is addressed in more detail in chapter 70. However, some of the key messages from this research are relevant here. One message is that change must take place at all levels of the system. In this context, Ferlie and Shortell (2001) identify four such levels: the individual, the group or team, the organization, and the larger system or environment. They note the growing evidence that strategies that focus on individuals alone are unlikely to be successful, whereas those that are embedded within broader organizational change are more likely to be effective (Davis and others 1995). A second key message is the importance of teamwork, with evidence that well-functioning, multiprofessional teams provide better quality care (Aiken, Sochalski, and Lake 1997). However, change may be inhibited by barriers at the level of the organization, including lack of a consistent focus on quality, inadequate information, lack of physician involvement, and inadequate managerial support (Shortell, Bennett, and Byck 1998). MANAGING CLINICAL SERVICES IN DIFFERENT SETTINGS Clinical services are provided in a variety of settings, from the patient’s home to ambulatory care facilities and hospitals providing inpatient care. They include those services that involve direct contact between a patient and a health care professional, as well as indirect contact, such as when a pathologist provides a diagnosis on a biopsy or blood sample. Reflecting the focus of much existing research, this section is structured in terms of different settings of care: ambulatory care, hospitals, and community care. Unfortunately, rather less research transcends these often artificial and arbitrary divisions to look at the more important issue of the patient’s journey through the health care system, given that one of the greatest managerial challenges facing those delivering clinical services is how to ensure that the journey is efficiently navigated (McKee and Nolte 2004). Furthermore, available research that focuses on health facilities is often difficult to generalize because of the different meanings attributed to common terms such as hospital, health center, or more prosaically, hospital bed. For example, a major teaching hospital in a capital city, such as the Kenyatta National Hospital in Nairobi (http://www.kenyattanationalhospital.org/ services.html), which offers invasive cardiology, renal transplants, and radiotherapy, is very different from a rural hospital with perhaps 100 beds and a single operating theater that provides only the most basic surgical and obstetric care. Hospitals Although hospitals are rarely the first places of contact between patients and health systems, and although hospitals do not provide the greatest share of health care, it is appropriate to begin with them because they often account for the largest share of public health sector expenditure (OECD 2003). They are also particularly difficult to manage for several reasons: • One reason is the diversity of tasks that a hospital must undertake (Healy and McKee 2002b). Many hospitals fulfill roles that go beyond the delivery of patient care to provide training and research, support to community-based facilities, and even local employment or civic identity symbols. • A second reason is the blurring of boundaries between hospitals and the rest of the health care system, which has occurred as a result of the emergence of many innovative models of care that cross the boundary between secondary and either primary or social care. A related issue is the shift taking place in many countries to managing patients through a complex combination of short inpatient stays and visits as an outpatient to specialist clinics and diagnostic facilities (McKee and Healy 2001). This approach is vastly more complicated to manage than the traditional model in which patients were admitted to wards from which they were taken to undergo investigations and treatment at a time convenient for the specialist concerned, a process managed by senior nurses. The new model requires new health worker roles, which might be termed case 1342 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 252 managers. These case managers help patients to navigate the system. • A third reason is the contrast between the rapidly changing demands on hospitals and the structural rigidities of hospitals themselves (McKee and Healy 2000). The original justification for creating hospitals as institutions was the need to concentrate equipment such as operating theaters, x-ray machines, and laboratories, and expertise such as medical specialists. Yet changes in the nature of health care are raising questions about how hospitals of the future should be configured. Many laboratory functions are being replaced by testing kits that can be used at the bedside, diagnostic equipment such as ultrasound scanners is being used in primary care, and a new generation of primary care workers are acquiring greatly augmented skills. In this rapidly changing environment, managers may be faced with aging hospital buildings that may lack sufficient electrical sockets for the greatly increased amount of equipment now available, or managers may have staff members with deeply ingrained ways of working who pose a particularly acute managerial problem. These issues can be seen in the Kenyatta National Hospital in Nairobi, where a new managerial approach was developed but faced problems because of an unclear understanding of the kind of services to be provided, weak managerial capacity, and a lack of focus in targeting services (Collins and others 1999). In Zambia, financial management and accountability improved when the hiring of hospital staff was delinked from the national civil service, yet the process has been derailed on a wider scale because of trade unions’ resistance to the changes (Hanson and others 2002). Ambulatory Care Ambulatory care, delivered on an outpatient basis, is the commonest form of contact between patients and health care providers. Although it often receives relatively little attention from policy makers compared with the more resourceintensive inpatient hospital care, ambulatory care contributes substantially to health care system performance (Berman 2000). Good management of ambulatory services is essential because these services are often the entry and exit points for consumers; however, these services can be difficult to manage effectively (Waghorn and McKee 2000). Effective coordination of ambulatory and inpatient services is needed to ensure that patients are cared for in the most appropriate settings, thereby reducing inefficiencies such as the overuse of hospitals for nonemergency care. Such coordination often involves developing shared protocols for referral and management. In Benin and Guinea, for example, diagnostic and treatment decision trees were developed collaboratively with the local staff, leading to more efficient use of resources (Levy-Bruhl and others 1997). In Zambia and Tanzania, strengthening of management capacities in primary care facilities through a team-based approach to decision making that linked planning to budgeted action plans led to improved client perceptions of facilities and to a marked increase in utilization (Few and Harpham 2003). However, the challenges of management in the ambulatory care sector are great in many countries; this sector is often highly fragmented, with extensive and largely unregulated private provision and few levers to exert pressure for change. Community and Social Care A particular challenge is how best to link long-term management of medical conditions with community and social care in those cases in which an effective response to health needs spans the interface. Management of chronic physical or mental illness in the elderly, for example, can fall under the responsibility of home care and volunteer agencies, day centers, day hospitals, rehabilitation hospitals, and long-term care institutions, as well as community-based health teams (Bergman, Beland, and Perrault 2002). A systematic review of community-based care for elderly people in industrial countries concluded that such schemes can favorably affect rates of institutionalization and costs. However, comprehensive approaches involving program restructuring are often necessary, and cost-effectiveness depends on the characteristics of the health and social care systems. The review’s authors identified as a critical challenge the expansion of those programs considered to be successful (Johri, Beland, and Bergman 2003). Low-income countries face particular obstacles because they often lack effective alternatives to hospital care. As a result, patients are frequently cared for by their families but with little support, or they are consigned to large, poorly equipped, and poorly staffed institutions. This situation has stimulated the development of models of “community care,” in which health care providers work with communities to deliver services. In the area of mental health, for example, the World Health Organization (WHO) has developed models of care that cover a range of care settings, including community centers and outreach services and residential homes, backed up by access to hospital outpatient and emergency care (WHO 2001). Similarly, the complexities of caring for people with disabilities in low-income countries have led to internationally developed guidelines that advocate a shared role for heath care providers and local communities (Helander 2000). Accordingly, effective coordination of those services clearly depends in large part on effective management. Shifting from hospital-focused care to community care introduces many managerial challenges. One element of an effective response should be to heighten the autonomy of patients in managing their diseases, but this Strategic Management of Clinical Services | 1343 ©2006 The International Bank for Reconstruction and Development / The World Bank 253 response requires attitudinal changes among providers, who must commit to a real shift of power to patients, supported by effective information systems and safeguards for vulnerable patients (Litwin and Lightman 1996). Health care systems are generally poor at addressing longterm illnesses, especially when those illnesses require integrated care spanning primary, secondary, and community providers (McKee and Nolte 2004; WHO 2001). The often low status accorded to these conditions in the hierarchy of priorities, coupled with fragmentation between health and social sectors (WHO 2002), will require greater commitment to managerial reforms that can improve the delivery of appropriate services. THE SPECIFICS: WHAT WORKS? This section turns to those policies that are designed to enhance the resources available to deliver health care and to combine them in ways that optimize the potential benefits. It looks, in turn, at the different elements required to deliver effective care: human resources, physical resources, intellectual resources, and the organizational or social resources that bind them together. The section begins with the most important resources for health care systems: the people who provide care. Developing Human Resources A key element in the delivery of effective health care is how to provide staff members with the appropriate combination of skills to do their jobs effectively. Increasing Skills. In their review undertaken to inform the Commission on Macroeconomics and Health, Oliveira-Cruz, Hanson, and Mills (2003) identified 13 studies that assessed the effects of training to enhance skills. Though the results were mixed, training programs were overall more likely to have positive rather than negative effects. Several studies focused on communication and counseling skills, which often lead to improved client satisfaction. A study from Zambia showed that training must be linked to other resources; although training was associated with improved transmission of information, there was no decline in the number of complaints from clients who remained unhappy about long waits and short contact time (Faxelid and others 1997). Changing Skill Mix. The division of tasks among different health care workers reflects many considerations, but evidence about who would be best at doing these tasks is rarely considered. There may be regulations restricting tasks to one professional group, such as the right to prescribe, or there may be cultural norms, which while unwritten have just as great an effect. Underlying these factors is a set of issues that includes a difference in the power of different professions, itself often a reflection of gender relationships in society, with a predominantly male medical profession controlling a predominantly female nursing profession. However, increasing evidence suggests that traditional demarcations do not support the optimal ways to provide care, and there is considerable scope for changing the mix of skills involved in delivering many aspects of health care. This topic has recently been reviewed systematically by Sibbald, Shen, and McBride (2004), who have developed a taxonomy of the types of change in skill mixes that are possible (table 73.1). Their review shows that many tasks undertaken by one professional group can yield comparable and often better results when performed by another group. In particular, they show how nurse-led clinics often achieve better outcomes than traditional doctor-led service (Connor, Wright, and Fegan 2002; Stromberg and others 2003; Vrijhoef, Diederiks, and Spreeuwenberg 2000; Vrijhoef and others 2001, 2003). Table 73.1 A Taxonomy of Changes in Skill Mix in Health Care Changing roles Enhancement Increasing the depth of a job by extending the role or skills of a particular group of workers Substitution Expanding the breadth of a job, in particular by working across professional divides or exchanging one type of worker for another Delegation Moving a task down a traditional unidisciplinary ladder Innovation Creating new jobs by introducing a new type of worker Changing the interface between services Transfer Moving the provision of a service from one health care setting to another (for example, substituting community for hospital care) Relocation Shifting the venue from which a service is provided from one health care sector to another without changing the people who provide it (for example, running a hospital clinic in a primary care facility) Liaison Using specialists in one health care sector to educate and support staff members working in another (for example, hospital outreach facilitators in primary care) Source: Sibbald, Shen, and McBride 2004. 1344 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 254 Although Sibbald, Shen, and McBride focus their review on experience in industrial countries, by challenging many deeply held beliefs they indicate what could be done in other settings around the world, after taking into account local circumstances such as the skills and expertise of those involved, as well as any salient regulatory or training issues. Strengthening Management In their review of constraints to health service delivery, Oliveira-Cruz, Hanson, and Mills (2003) identified 10 studies that evaluated the effect of management strengthening. The activities in those studies included the following: • workshops for identifying and prioritizing managerial programs • introduction of regular planning and evaluation cycles • quality assurance methods • establishment of routine communication systems • training activities. They concluded that the results were generally positive, with more rational use of funds; greater availability of funds as a consequence of better planning; improved coordination and integration of programs; improved methods of working; better staff morale; enhanced data collection, reporting, and use; and increased community participation. WHO has developed an approach to strengthening management that has been successful in a variety of settings (Cassels and Janovsky 1995). It is important to identify where specific managerial skills are lacking and to explore different ways of obtaining them, whether through training, recruitment, or links with related organizations. For example, improved financial management in district health teams in Ghana was made possible by integrating staff members from local government accounts offices (Kanlisi 1991); a similar initiative was successful in The Gambia (Conn, Jenkins, and Touray 1996). However, a word of caution is required. Although a management strengthening exercise undertaken in Tanzania was successful when implemented at the local level, it failed when scaled up because the same degree of involvement by the originating team was no longer possible (Barnett and Ndeki 1992). needs (England 2000; Preker, Harding, and Travis 2000). Capital charging—requiring managers to explicitly account for the value of physical assets out of funding allocation or contract revenues—has been developed as a response, successfully heightening public sector management of capital investments in the United Kingdom and New Zealand (Heald and Scott 1996). Capital charging has been proposed as a strategy to stimulate better capital management in developing countries as well. For example, in Malaysia a corporatized hospital has been required to reimburse invested capital through dividends, with the Malaysian government recouping one-third of its original investment within five years (Hussein and Al-Junid 2003). Similarly, the Kenyatta National Hospital in Nairobi was obliged to account for all accruals (for example, property and depreciation) when it was given greater autonomy. Though changes in accounting management have experienced some shortcomings, improvements have been seen in financial transparency, timeliness of reporting, donor satisfaction, and revenue collection (Collins and others 1999). Within the public sector, changes in line management have facilitated the incorporation of more explicit infrastructure concerns into the planning process. The central authority in Hong Kong (China) has made capital acquisition decisions jointly with hospitals during annual planning processes (Yip and Hsiao 2003). The introduction of business planning to district-level planning in Turkmenistan heightened accountability for maintaining physical infrastructure: use of a global budgeting model (that is, increased autonomy in line management as well as performance monitoring) led to reduced resource allocation to personnel and a greater than fivefold increase in maintenance expenditures (Ensor and Amannyazova 2000). Explicitly managing capital investments in both the short and the long term may facilitate efficient resource allocation. Although capital charging is a relatively straightforward technical solution, capital investment can be particularly susceptible to political derailment (Anell and Barnum 1998). In the hospital sector, for instance, many transition economy countries have had difficulty downsizing infrastructure because those with decision rights to manage capital (that is, local governments) are different from those who have incentives to do so, such as hospital managers (Jakab and Preker 2003). Managing Physical Resources Managing infrastructure and other capital assets such as hospitals and health centers requires investment planning in both the short term (for example, maintenance) and the long term (for example, new acquisitions). Historically, however, costs associated with capital consumption and maintenance have not been met through operating budgets, resulting in few incentives for public sector health planners to efficiently manage infrastructure or to respond to market demand and consumer Strengthening Drug Procurement, Regulation, and Distribution Managing pharmaceutical resources is crucial for ensuring access to essential drugs and promoting their rational use (Mossialos, Mrazek, and Walley 2004). WHO defines the goals of rational use of drugs as delivering medications effectively— appropriate to patients’ clinical needs and at dose levels and durations appropriate to their individual requirements—and Strategic Management of Clinical Services | 1345 ©2006 The International Bank for Reconstruction and Development / The World Bank 255 at an affordable cost (WHO 1985). The public sector plays a key role in providing the framework for rational use of drugs (Quick 1997) through measures ranging from drug regulation to clinical practice guidelines. National drug policies (NDPs) can be effective in regulating private and public sector provision of essential medicines. The Lao People’s Democratic Republic’s NDP has been important in improving private pharmacy service quality (Stenson, Tomson, and Syhakhang 1997). In Burkina Faso, an NDP has enhanced the performance of rural pharmacies (Krause and others 1998). At the local and facility levels, increasing accountability can also lead to a more rational use of drugs. A simulation exercise in Tunisia that required physicians to relate pharmaceutical budgeting to involvement in the procurement process improved prescribing practices by containing costs while increasing the use of essential drugs (Garraoui, Le Feuvre, and Ledoux 1999). Enhanced management information systems, with corresponding supervision, monitoring, and top-level support, have improved contraceptive management in several countries (Kinzett and Bates 2000). The introduction of standard treatment guidelines and formularies has reduced overprescribing in several countries, and educational materials for consumers in Cameroon increased compliance with antibiotic regimens (Nabiswa, Makokha, and Godfrey 1993). A comprehensive review of interventions used in SubSaharan Africa, where health systems are plagued by shortages of supplies, high costs, large-scale use of proprietary drugs, waste, and theft, provided considerable evidence to suggest what works in those countries (Foster 1991). Successful interventions included the following: • selection and precise quantification of drug needs—in particular, the creation of essential drug lists • improved procurement, with greater use of generics, competitive bidding, and international procurement agencies • improved storage and distribution, with better storage conditions, inventory controls, security systems, and use of prepacked kits. At the same time, several factors constrain better management of pharmaceuticals. Considerable resources are needed to adequately monitor NDPs, and implementation can be difficult (Petrova 2002). Furthermore, much of the pharmaceutical use is outside the control of the public sector: two-thirds or more of health problems are self-medicated. Though the public sector may strive to inform consumers, patients’ nonadherence remains high (Le Grand, Hogerzeil, and Haaijer-Ruskamp 1999). As in management of other inputs, political considerations can thwart managerial responses. The Republic of Korea decided to divide its prescribing and dispensing functions precisely to address high levels of pharmaceutical overuse and misuse, but it subsequently faced strikes and stiff opposition from those same stakeholders (Kwon 2003). Management of pharmaceuticals thus presents a complicated challenge, requiring significant investment and flexible responses. Using Intellectual Resources The process of generating, disseminating, and using knowledge is frequently imperfect. Pang and others (2003) have argued that a well-functioning health care system must have in place mechanisms that allow it to access and use research and the products of research. They highlight the weaknesses of much of the existing health care research, including fragmentation, overspecialization, and damaging competition among researchers, who are frequently isolated from other researchers and from the policy-making community. Drawing on concepts of the functions of a health system, they identify a series of four roles for a health research system: • stewardship, which includes defining and articulating a vision for a national research system, identifying appropriate priorities, and setting and monitoring ethical standards • financing, which includes obtaining research funds and allocating them accountably • creating and sustaining resources, which includes the physical and human capacity to conduct, absorb, and use research • producing and using research, which includes generating valid research outputs; translating research into formats that inform health policy, practices, and public opinion; and promoting the use of research to support innovation. Such a system must be able to answer the many different questions requiring research, from basic laboratory science, such as new drug development, through health services research, such as comparisons of the cost-effectiveness of different drug regimens, to organizational research, such as the best way of delivering the most cost-effective drug regimen. Although the majority of health systems and services research continues to be undertaken in the industrial countries, a growing volume of research addresses the needs of low- and middleincome countries, such as that by the participants in the Effective Health Care Alliance Programme (EHCAP), an international research network that is undertaking systematic research within the framework of the Cochrane Collaboration (http://www.liv.ac.uk/lstm/ehcap/introduction.htm). Establishing Relationships The debate about the relative benefits of vertical (in which a single disorder is tackled by a program managed across levels from the Ministry of Health to the health care provider) and horizontal (in which health care for a wide range of disorders is delivered through a system that is integrated at each level) 1346 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 256 systems of health care delivery has been examined in detail in a major review of relevant literature by Oliveira-Cruz, Kurowski, and Mills (2003). They note how many activities lie on a continuum between the two extremes, with the Global Polio Eradication Initiative more vertical than the Expanded Programme on Immunization, which in turn is more vertical than the integrated management of childhood illness approach. They identify certain features that are often associated with vertical programs and that promote success: specific objectives, clear work schedules, well-defined techniques, and frequent supervision. They also identify characteristics that are often associated with horizontal programs and that can hamper effectiveness: shortage of essential drugs, lack of adequate staff training, intermittent supervision, and limited backup. However, they note that horizontal programs have considerable potential to deliver effective services if they are adequately funded, staffed, and managed, largely because of their economies of scale and scope. To some extent, the approach is determined by the nature of the program. Vertical programs are most effective when the technology involved is very sophisticated or when it includes procedures different from the usual tasks and thus requires specialist skills. Vertical programs may be more appropriate when there is a need to rapidly achieve major reductions in the burden of a disease, although this situation does not preclude embedding the management of the program within existing organizations. These programs are often a response to weak management capacity in the existing system, although it is argued that they can perpetuate this problem or even undermine what does exist, diverting the attention of staff members from their usual tasks. Such programs often have a short time horizon, either being absorbed into existing systems or brought to an end. In part, their duration is linked to the source of their funding, which is often from donors who themselves have a short time horizon. Integrating previously vertical programs into mainstream systems can be successful, as with schistosomiasis programs in Saudi Arabia (Ageel and Amin 1997) and Brazil (Coura Filho and others 1992). However, a systematic review of integration failed to identify consistent benefits, largely because of the very limited extent of the evidence available and the context-specific nature of this process (Briggs, Capdegelle, and Garner 2001). The authors of that review concluded that the question facing policy makers is not whether one approach is invariably better than the other; rather, it is how best to build on the synergies among them to maximize overall benefits. They note, for example, how the many successes of the Malaria Eradication Programme in the 1950s and 1960s were not sustained because active case surveillance was not integrated into routine health services (see also Bradley 1998). Successful vertical programs are likely to involve community participation, but not to the extent that there is overdependence and subsequent attrition of volunteers. The programs’ developers will have learned lessons from other similar programs, in relation to both organizational and technical issues. Where several vertical programs coexist, the programs’ developers should explore how they can share common elements. Contracting for Services The setting of contracts by public agencies to purchase health care services is increasingly common in a number of low- and middle-income countries. The theoretical case for contracting out identifies potential advantages from combining public finance with private provision. However, there may also be difficulties, such as ensuring that competition takes place among potential contractors, that competition leads to efficiency, and that contracts and the process of contracting are effectively managed; consequently, these advantages may not always be realized (McPake and Banda 1994). Unfortunately, the question of whether the advantages outweigh the disadvantages has been the subject of relatively little empirical study in low- and middle-income countries, and what exists is often highly context specific. For example, in Zimbabwe, a comparison of a hospital owned by a colliery, from which services were purchased by the government, and a nearby government hospital found that the colliery hospital offered services of at least comparable quality at prices lower than the unit costs of the government hospital after capital costs were included (McPake and Hongoro 1995). However, failure to establish policies on thresholds for use meant that growth in expenditure on the colliery hospital was not controlled. The authors argue that contracted facilities can achieve powerful bargaining positions if there are no viable competitors and the government does not retain the ability to offer an alternative service. They also identify a need for specific skills to manage contracts at all levels. Where a policy of contracting is a response to crises arising from civil service retrenchment and public expenditure cuts, these skills are unlikely to be developed. Another study examined the economic arguments for contracting for district hospital care in South Africa, by using private for-profit providers, and in Zimbabwe, by using nongovernmental (mission) providers (Mills, Hongoro, and Broomberg 1997). In the South African setting, there were no significant differences in quality among three contractor hospitals and three government-run hospitals, but the contractor hospitals provided care at significantly lower unit costs. However, the overall cost to the government was similar for the two options because of the additional cost of contracting, with the efficiency gains captured almost entirely by the contractor. In Zimbabwe, two district-designated mission hospitals delivered similar quality care at lower cost than did two government hospitals. However, the contract between the government and Strategic Management of Clinical Services | 1347 ©2006 The International Bank for Reconstruction and Development / The World Bank 257 the missions was implicit, rather than explicit, and was of long standing. As in the other Zimbabwean example, the authors identified the importance of developing the government’s capacity to design and negotiate contracts that allow the government to derive significant efficiency gains from contractual arrangements. Increasing Provider Autonomy A review of cross-country experiences with enhanced autonomy of hospitals found improvements in service delivery. The most successful cases—in Hong Kong (China) and Tunisia— applied private sector management techniques and training, with appropriate performance assessment systems for staff. In countries where reforms were considered less successful, managers had been granted greater autonomy without suitable performance-oriented incentives (New Zealand) or vice versa (Indonesia) (Hawkins and Ham 2003). In the Kenyatta National Hospital, greater autonomy led to the introduction of performance appraisal linked to incentives, enabling the dismissal of poor performers and increased benefits and greater responsibilities for good performers. This change was coupled with clarification of clinical management roles. Complementing increased salaries for staff nurses, these changes helped improve the hospital’s strategic management, donor accountability, and performance reporting (Collins and others 1999). Implementing such management strategies in a coherent fashion is not an easy task. Hospital governance in several Eastern European countries, which has been transferred to local governments to improve responsiveness, has included measures such as performance-based payment mechanisms. Performance did not improve as expected because of an “inconsistent incentive environment”; rewards and sanctions were not linked to performance. Important factors in that failure to improve were weak stewardship functions and an absence of effective governance at the regional level, which made it difficult to change the initial configuration of the hospital system. Instead, increased hospital autonomy was used to ensure the survival of the institution rather than to meet the needs of the population. Thus, a continuing excess of capacity, inefficiency, and poor responsiveness to patient expectations remains (Healy and McKee 2002a). A review of experience with programs that increased autonomy in Sub-Saharan Africa also identified only modest success in achieving the stated goals (McPake 1996). Public or Private Provision? Although there has been considerable enthusiasm for privatizing state facilities because of the supposed efficiency gains achieved in the private sector, in reality the evidence is somewhat mixed. Thus, a study of dispensaries run by the government and by nongovernmental organizations in Tanzania found considerable variation in both sectors (Gilson 1995). This finding was consistent with another study in Tanzania of primary care providers in Dar es Salaam. In the latter, although the quality of care offered by private providers was, on average, better, much low-quality care was found in both types of facilities (Kanji and others 1995). Considerable variation in providers of both types, although with overall better quality in the private sector, was also reported in a study in Senegal (Bitran 1995). In summary, there is very little evidence to support the contention that private provision is better than public provision, and what evidence exists indicates considerable variations in both. Strategic Purchasing The quest to deliver effective health care is a dynamic process, adapting continually to changing health needs and the opportunities that arise that make it possible to respond in new and better ways. However, health systems that have failed in the past to respond to these changing circumstances face even greater problems. The pace of change is constantly increasing, with factors such as greater population mobility contributing to the reemergence of infectious diseases and with demographic changes and lifestyle changes giving rise to a new burden of chronic diseases. Health care providers have faced difficulties in responding to this challenge on their own. Although they may possess a great deal of information about the patient sitting in front of them and, on the basis of their training and accumulated experience, about what might be done to help that patient, health care providers confront several important information gaps: • First, they may know little about those who, despite being in need of health care, do not seek help. These people will often be the most disadvantaged in a society, with few means of making their voices heard. • Second, they may have inadequate knowledge about newly emerging treatments or more effective ways of providing those treatments, especially if the treatments involve creating multidisciplinary teams with new sets of skills, working in ways outside their experience. • Third, even if providers introduce changes, they may have inadequate knowledge of whether such changes have been effective. These knowledge gaps provide the justification for action to improve the delivery of health care at several levels above that of the individual encounter between the patient and the health professional. Strategic purchasing brings together a series of interlinked activities: assessing health needs, using appropriate 1348 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 258 Health strategy SUMMARY Assess needs Monitor outcome Specify care model Purchase care Source: McKee and Brand 2005. Figure 73.1 A Framework for Strategic Purchasing evidence to develop models of care that meet priority needs, creating the appropriate combination of regulations and incentives to implement those models, and then evaluating the response and reassessing whether the need remains (Figure 73.1). All of these activities should take place within an overall health strategy that takes into account the goals of a health care system, such as those defined by WHO (2000), of increasing health attainment, providing services responsive to the population’s needs and expectations, and financing those services equitably. The development of a strategic purchasing function is complicated, requiring high levels of information resources, both on health needs and on effectiveness. Strategic purchasing involves using technical and political skills, determining the needs of the population, identifying evidence of the effectiveness of different care packages, and setting priorities within limited resources. The last of these components is arguably the most difficult, given the high level of need and the scarcity of resources in many places. This list of components highlights why, in addition to having skills in financial and personnel management, the effective health service manager needs at least a working knowledge of clinical epidemiology and economic evaluation. Even in industrial countries, the strategic purchasing function is often poorly developed. Given its many interlinked components and the problem of isolating any benefits from wider changes in the health system, this function is very difficult to evaluate. Nonetheless, it is included here as a model from which concepts may be adopted in low- and middle-income settings. Health systems worldwide face unprecedented challenges in responding to the increasing complexity of health care. Systems that were capable of providing basic care to populations for whom diseases were either simple or complex but self-limiting confront a fatal struggle to keep up with the increasing opportunities that modern science has provided. The challenge is especially great for health systems in low- and middle-income countries because the global community is no longer willing to sit back while millions of people die from treatable diseases such as malaria and tuberculosis and fail to receive lifeprolonging therapies for AIDS. As a consequence, some of the resources, primarily pharmaceuticals, are being made available to those who need them. However, the challenge that health systems face is not simply a lack of money to purchase pharmaceuticals; effective management systems are requisite as well to create the infrastructure to identify those in need, establish appropriate treatments, and ensure provision of these treatments as long as necessary. Emerging challenges must be identified, and the necessary resources to deal with them must be brought together and applied effectively. Many countries have a clear need to invest in the development of human resources. Although in many cases this investment will require new and wide-ranging human resource strategies involving training, career progression, and retention, there seems to be scope for rapid gains from some shorter forms of training, particularly, in communications skills. Although the evidence for the effectiveness of current models of management strengthening is somewhat mixed, gains may be realized by identifying and filling key gaps, such as those in financial expertise. Changing the skill mix can do much to match available skills to tasks. Much also can be done to manage the capital stock better or, in most cases, to manage it at all. For example, mechanisms such as capital charging can focus greater management attention on this issue, although this will work only if sufficient capacity can be focused. Important gains can be made from better management of pharmaceuticals, an issue of increasing importance because of the new funds made available for their purchase. Modern health care is based on the growth of knowledge, and it is as important to manage intellectual resources as it is to manage people and equipment. Doing so means investing in a health research strategy that includes the generation, synthesis, and adoption of knowledge. Finally, it is necessary to bring these resources together optimally, which raises issues of relationships between different levels of the system, between the public and private sectors, and between vertical and horizontal programs. Unfortunately, despite the large amount of rhetoric on these often highly ideological issues, there is surprisingly little research to inform Strategic Management of Clinical Services | 1349 ©2006 The International Bank for Reconstruction and Development / The World Bank 259 policy. More than ever, the issue of context specificity reemerges, leading once more to the answer “it depends.” The delivery of optimal health care requires well-developed managerial skills to apply methods that are appropriate for the setting in which they are being applied. However, it also requires governments to provide oversight of their health systems and to anticipate changes and give managers the tools with which to respond to those changes. REFERENCES Ageel, A. R., and M. A. Amin. 1997. “Integration of SchistosomiasisControl Activities into the Primary-Health-Care System in the Gizan Region, Saudi Arabia.” Annals of Tropical Medicine and Parasitology 91: 907–15. Aiken, L., J. Sochalski, and E. Lake. 1997. “Studying Outcomes of Organizational Change in Health Services.” Medical Care 35 (11 Suppl.): NS6–18. Andreev, E. M., E. Nolte, V. M. Shkolnikov, E. Varavikova, and M. McKee. 2003. “The Evolving Pattern of Avoidable Mortality in Russia.” International Journal of Epidemiology 32: 437–46. Anell, A., and H. Barnum. 1998. “The Allocation of Capital and Health Sector Reform.” In Critical Challenges for Health Care Reform in Europe, ed. R. B. Saltman, J. Figueras, and C. Sakellarides, 179–96. Buckingham, U.K.: Open University Press. Arrow, K. 1963.“Uncertainty and the Welfare Economics of Medical Care.” American Economic Review 53: 941–73. Barnett, E., and S. Ndeki. 1992. “Action-Based Learning to Improve District Management: A Case Study from Tanzania.” International Journal of Health Plan Management 7: 299–308. Baumol, W., J. Panzar, and R. Willig. 1982. Contestable Markets and the Theory of Industrial Structure. New York: Harcourt Brace Jovanovich. Bergman, H., F. Beland, and A. Perrault. 2002. “The Global Challenge of Understanding and Meeting the Needs of the Frail Older Population.” Aging Clinical and Experimental Research 14: 223–25. Berman, P. 2000. “Organization of Ambulatory Care Provision: A Critical Determinant of Health System Performance in Developing Countries.” Bulletin of the World Health Organization 78: 791–802. Bitran, R. 1995. “Efficiency and Quality in the Public and Private Sectors in Senegal.” Health Policy and Planning 10: 271–83. Bradley, D. J. 1998.“The Particular and the General: Issues of Specificity and Verticality in the History of Malaria Control.” Parassitologia 40: 5–10. Briggs, C. J., P. Capdegelle, and P. Garner. 2001. “Strategies for Integrating Primary Health Services in Middle- and Low-Income Countries: Effects on Performance, Costs, and Patient Outcomes.” Cochrane Database of Systematic Reviews (4): CD003318. Cassels, A., and K. Janovsky. 1995. Strengthening Health Management in Districts and Provinces. Geneva: World Health Organization. Collins, D., G. Njeru, J. Meme, and W. Newbrander. 1999. “Hospital Autonomy: The Experience of Kenyatta National Hospital.” International Journal of Health Planning and Management 14: 129–53. Conn, C. P., P. Jenkins, and S. O. Touray. 1996. “Strengthening Health Management: Experience of District Teams in The Gambia.” Health Policy and Planning 11: 64–71. Connor, C. A., C. C. Wright, and C. D. Fegan. 2002. “The Safety and Effectiveness of a Nurse-Led Anticoagulant Service.” Journal of Advanced Nursing 38: 407–15. Coura Filho, P., R. S. Rocha, E. De Lima, M. F. Costa, and N. Katz. 1992. “A Municipal Level Approach to the Management of Schistosomiasis Control in Peri-Peri, MG, Brazil.” Revisto do Instituto de Medicina Tropical de São Paulo 34: 543–48. Davis, D., M. Thomson, A. Oxman, and R. Haynes. 1995. “Changing Physician Performance. A Systematic Review of the Effect of Continuing Medical Education Strategies.” Journal of the American Medical Association 274: 700–05. de Soto, H. 2000. The Mystery of Capital: Why Capitalism Succeeds in the West and Fails Everywhere Else. Boulder CO: Basic Books. England, R. 2000. Contracting and Performance Management in the Health Sector: A Guide for Low and Middle Income Countries. London: Department for International Development Health Systems Resource Centre. Ensor, T., and B. Amannyazova. 2000. “Use of Business Planning Methods to Monitor Global Health Budgets in Turkmenistan.” Bulletin of the World Health Organization 78: 1045–53. Enthoven, A. C. 1985. Reflections on the Management of the National Health Service. London: Nuffield Provincial Hospitals Trust. Evans, R. G. 1997. “Health Care Reform: Who’s Selling the Market, and Why?” Journal of Public Health Medicine 19: 45–49. Farmer, P., L. Reichman, and M. Iseman. 1999. The Global Impact of Drug Resistant Tuberculosis. Boston, MA: Harvard Medical School and Open Society Institute. Faxelid, E., B. M. Ahlberg, S. Freudenthal, J. Ndulo, and I. Krantz. 1997. “Quality of STD Care in Zambia: Impact of Training in STD Management.” International Journal for Quality in Health Care 9: 361–66. Fayol, H. 1949. General and Industrial Management. Translated from the original, Administration Industriele et Générale, 1916. London: Pitman. Ferlie, E., and S. Shortell. 2001. “Improving the Quality of Health Care in the United Kingdom and the United States: A Framework for Change.” Milbank Quarterly 79: 281–315. Few, R., and T. Harpham. 2003. “Urban Primary Health Care in Africa: A Comparative Analysis of City-Wide Public Sector Projects in Lusaka and Dar es Salaam.” Health and Place 9: 45–53. Foster, S. 1991. “Supply and Use of Essential Drugs in Sub-Saharan Africa: Some Issues and Possible Solutions.” Social Science and Medicine 32: 1201–18. Garraoui, A., P. Le Feuvre, and M. Ledoux. 1999. “Introducing Management Principles into the Supply and Distribution of Medicines in Tunisia.” Bulletin of the World Health Organization 77: 525–29. Gilson, L. 1995. “Management and Health Care Reform in Sub-Saharan Africa.” Social Science and Medicine 40: 695–710. Griffiths, R. 1984. National Health Service Management Inquiry Report. London: Department of Health and Social Security. Hanson, K., L. Atuyambe, J. Kamwanga, B. McPake, O. Mungule, and F. Ssengooba. 2002. “Towards Improving Hospital Performance in Uganda and Zambia: Reflections and Opportunities for Autonomy.” Health Policy 61: 73–94. Hanson, K., M. K. Ranson, V. Oliveira-Cruz, and A. Mills. 2003. “Expanding Access to Priority Health Interventions: A Framework for Understanding the Constraints to Scaling Up.” Journal of International Development 15: 1–14. Harrison, S., and C. Pollitt. 1994. Controlling Health Professionals. Buckingham, U.K.: Open University Press. Hawkins, L., and C. Ham. 2003. “Reviewing the Case Studies: Tentative Lessons and Hypotheses for Future Testing.” In Innovations in Health Service Delivery: The Corporatization of Public Hospitals, ed. A. S. Preker and A. L. Harding. Washington, DC: World Bank. Heald, D., and D. Scott. 1996. “Assessing Capital Charging in the National Health Service.” Financial Accountability and Management 12: 225–44. Healy, J., and M. McKee. 2002a. “Implementing Hospital Reform in Central and Eastern Europe.” Health Policy 61: 1–19. 1350 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 260 ———. “The Role and Function of Hospitals.” In Hospitals in a Changing Europe, ed. M. McKee and J. Healy, 59–80. Buckingham, U.K.: Open University Press. Helander, E. 2000. “25 Years of Community-Based Rehabilitation.” Asia Pacific Disability Rehabilitation Journal 11: 4–9. Hood, C. 1991. “A Public Management for All Seasons.” Public Administration 69: 3–19. Hussein, R., and S. Al-Junid. 2003. “Corporatization of a Single Facility: Reforming the Malaysian National Heart Institute.” In Innovations in Health Service Delivery: The Corporatization of Public Hospitals, ed. A. S. Preker and A. L. Harding. Washington, DC: World Bank. Institute of Medicine. 2001. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: Institute of Medicine. Jakab, M., and A. S. Preker. 2003. “The Missing Link? Hospital Reform in Transition Economies.” In Innovations in Health Service Delivery: The Corporatization of Public Hospitals, ed. A. S. Preker and A. Harding. Washington, DC: World Bank. Johri, M., F. Beland, and H. Bergman. 2003. “International Experiments in Integrated Care for the Elderly: A Synthesis of the Evidence.” International Journal of Geriatric Psychiatry 18: 222–35. Kanji, N., P. Kilima, N. Lorenz, and P. Garner. 1995. “Quality of Primary Outpatient Services in Dar-es-Salaam: A Comparison of Government and Voluntary Providers.” Health Policy and Planning 10: 186–90. Kanlisi, N. 1991. “Strengthening District Health Teams in Ghana: The Experience of Ejisu District.” Tropical Doctor 21: 98–100. McPake, B., and E. E. Banda. 1994. “Contracting Out of Health Services in Developing Countries.” Health Policy and Planning 9: 25–30. McPake, B., and C. Hongoro. 1995. “Contracting Out of Clinical Services in Zimbabwe.” Social Science and Medicine 41: 13–24. Mehotra, S. 2000.“Integrating Economic and Social Policy: Good Practices from High Achieving Countries.” Innocenti Working Paper 80, UNICEF Innocenti Research Centre, Florence, Italy. Mills, A., C. Hongoro, and J. Broomberg. 1997. “Improving the Efficiency of District Hospitals: Is Contracting an Option?” Tropical Medicine and International Health 2: 116–26. Mossialos, E., M. Mrazek, and T. Walley. 2004. Regulating Pharmaceuticals in Europe: Striving for Efficiency, Equity, and Quality. Buckingham, U.K.: Open University Press. Nabiswa, A. K., J. D. Makokha, and R. C. Godfrey. 1993. “Malaria: Impact of a Standardized Protocol on Inpatient Management.” Tropical Doctor 23: 25–26. OECD (Organisation for Economic Co-Operation and Development). 2003. OECD Health Data. Paris: OECD. Oliveira-Cruz, V., K. Hanson, and A. Mills. 2003. “Approaches to Overcoming Constraints to Effective Health Service Delivery: A Review of the Evidence.” Journal of International Development 15: 41–65. Oliveira-Cruz, V., C. Kurowski, and A. Mills. 2003. “Delivery of Priority Health Services: Searching for Synergies with the Vertical versus Horizontal Debate.” Journal of International Development 15: 67–86. Kinzett, S., and J. Bates. 2000. Bangladesh: Contraceptive Logistics System, Review of Accomplishments and Lessons Learned. Arlington, VA: Family Planning Logistics Management and John Snow Inc. Pang, T., R. Sadana, S. Hanney, Z. A. Bhutta, A. A. Hyder, and J. Simon. 2003. “Knowledge for Better Health: A Conceptual Framework and Foundation for Health Research Systems.” Bulletin of the World Health Organization 81: 815–20. Krause, G., J. Benzler, R. Heinmuller, M. Borchert, E. Koob, K. Ouattara, and H. J. Diesfeld. 1998. “Performance of Village Pharmacies and Patient Compliance after Implementation of Essential Drug Programme in Rural Burkina Faso.” Health Policy and Planning 13: 159–66. Petrova, G. I. 2002. “Prescription Patterns Analysis—Variations among Bulgaria, Romania, Macedonia, and Bosnia Herzegovina.” Central European Journal of Public Health 10 (3): 100–03. Kwon, S. 2003. “Pharmaceutical Reform and Physician Strikes in Korea: Separation of Drug Prescribing and Dispensing.” Social Science and Medicine 57: 529–38. Preker, A. S., A. Harding, and P. Travis. 2000. “‘Make or Buy’ Decisions in the Production of Health Care Goods and Services: New Insights from Institutional Economics and Organizational Theory.” Bulletin of the World Health Organization 78: 779–90. Le Grand, A., H. V. Hogerzeil, and F. M. Haaijer-Ruskamp. 1999. “Intervention Research in Rational Use of Drugs: A Review.” Health Policy and Planning 14 (2): 89–102. Quick, J. D. 1997. Managing Drug Supply: The Selection, Procurement, Distribution, and Use of Pharmaceuticals. West Hartford, CT: Kumarian Press. Le Grand, J., and W. Bartlett. 1993. Quasi-Markets and Social Policy. London: Palgrave Macmillan. Shortell, S., C. Bennett, and G. Byck. 1998. “Assessing the Impact of Continuous Quality Improvement on Clinical Practice: What It Will Take to Accelerate Progress.” Milbank Quarterly 76: 593–624. Levy-Bruhl, D., A. Soucat, R. Osseni, J. M. Ndiaye, B. Dieng, X. De Bethune, and others. 1997. “The Bamako Initiative in Benin and Guinea: Improving the Effectiveness of Primary Health Care.” International Journal of Health Planning and Management 12 (Suppl. 1): S49–79. Litwin, H., and E. Lightman. 1996. “The Development of Community Care Policy for the Elderly: A Comparative Perspective.” International Journal of Health Services 26: 691–708. Shortell, S. M., and A. D. Kaluzny. 1983. Health Care Management: A Text in Organisation Theory and Behaviour. New York: Wiley. Sibbald, B., J. Shen, and A. McBride. 2004. “Changing the Skill-Mix of the Health Care Workforce.” Journal of Health Services Research and Policy 2004 9: 28–38. Maynard, A. 1998. “Competition and Quality: Rhetoric and Reality.” International Journal for Quality in Health Care 10: 379–84. Stenson, B., G. Tomson, and L. Syhakhang. 1997. “Pharmaceutical Regulation in Context: The Case of Lao People’s Democratic Republic.” Health Policy and Planning 12: 329–40. McKee, M., and H. Brand. 2005. “Purchasing to Promote Population Health.” In Effective Purchasing for Health Gain, ed. J. Figueras, R. Robinson, and E. Jakubowski, 140–63. Buckingham, U.K.: Open University Press. Stewart, J. 1998. “Advance or Retreat: From the Traditions of Public Administration to the New Public Management and Beyond.” Public Policy and Administration 13: 27. McKee, M., and J. Healy. 2000. “The Role of the Hospital in a Changing Environment.” Bulletin of the World Health Organization 78: 803–10. ———. 2001. “The Changing Role of the Hospital in Europe: Causes and Consequences.” Clinical Medicine, 1: 299–304. McKee, M., and E. Nolte. 2004. “Responding to the Challenge of Chronic Disease: Ideas from Europe.” Clinical Medicine 4: 336–42. McPake, B. 1996. “Public Autonomous Hospitals in Sub-Saharan Africa: Trends and Issues.” Health Policy 35: 155–77. Stromberg, A., J. Martensson, B. Fridlund, L. A. Levin, J. E. Karlsson, and U. Dahlstrom. 2003. “Nurse-Led Heart Failure Clinics Improve Survival and Self-Care Behaviour in Patients with Heart Failure: Results from a Prospective, Randomised Trial.” European Heart Journal 24: 1014–23. Vrijhoef, H. J., J. P. Diederiks, and C. Spreeuwenberg. 2000. “Effects on Quality of Care for Patients with NIDDM or COPD When the Specialised Nurse Has a Central Role: A Literature Review.” Patient Education and Counseling 41: 243–50. Strategic Management of Clinical Services | 1351 ©2006 The International Bank for Reconstruction and Development / The World Bank 261 Vrijhoef, H. J., J. P. Diederiks, C. Spreeuwenberg, and B. H. Wolffenbuttel. 2001. “Substitution Model with Central Role for Nurse Specialist Is Justified in the Care for Stable Type 2 Diabetic Outpatients.” Journal of Advanced Nursing 36 (4): 546–55. ———. 2000. World Health Report 2000: Health Systems: Improving Performance. Geneva: WHO. Vrijhoef, H. J., J. P. Diederiks, G. J. Wesseling, C. P. Van Schayck, and C. Spreeuwenberg. 2003. “Undiagnosed Patients and Patients at Risk for COPD in Primary Health Care: Early Detection with the Support of Non-Physicians.” Journal of Clinical Nursing 12: 366–73. ———. 2002. Community Home-Based Care in Resource-Limited Settings: A Framework for Action. Geneva: WHO. Waghorn, A., and M. McKee. 2000. “Why Is It So Difficult to Organise an Outpatient Clinic?” Journal of Health Services Research and Policy 5: 140–47. WHO (World Health Organization). 1985. “The Rational Use of Drugs: Review of Major Issues.” Paper prepared for the Conference of Experts on the Rational Use of Drugs, Nairobi, Kenya, November 25–29. ———. 2001. Innovative Care for Chronic Conditions. Geneva: WHO. Yip, W., and W. Hsiao. 2003. “Autonomizing a Hospital System: Corporate Control by Central Authorities in Hong Kong.” In Innovations in Health Service Delivery: The Corporatization of Public Hospitals, ed. A. S. Preker and A. L. Harding. Washington, DC: World Bank. 1352 | Disease Control Priorities in Developing Countries | Alexander S. Preker, Martin McKee, Andrew Mitchell, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 262 Chapter 1 Measuring the Global Burden of Disease and Risk Factors, 1990–2001 Alan D. Lopez, Colin D. Mathers, Majid Ezzati, Dean T. Jamison, and Christopher J. L. Murray In an era when most societies must cope with increasing demand for health resources, they will inevitably have to make choices about the provision of health services, even if those choices are, by default, to continue current practices. Strategic health planning can accelerate health development and the attainment of health goals or reduce the cost of reaching such goals. Such planning must take into account the needs that the health system must address; that is, policy makers must be aware of the comparative burden of diseases and injuries and the risk factors that cause them, and how this burden is likely to change with the adoption of various policies and interventions. Needs are, of course, not the only factors determining service provision, but should be a critical component of the decision-making and planning processes. The issue then becomes how to assess the comparative importance of risks to health and their outcomes in different demographic groups of the population.What is needed is a framework for integrating, validating, analyzing, and disseminating the fragmentary, and at times contradictory, information that is available on a population’s health, along with some understanding of how that population’s health is changing, so that the information is more relevant for health policy and planning purposes. The Global Burden of Disease (GBD) framework is the principal, if not the only, attempt to do so. Features of the GBD framework include the incorporation of data on nonfatal health outcomes into summary measures of population health, the development of methods for assessing the reliability of data and imputing missing data, and the use of a common metric to summarize the disease burden from diagnostic categories of the International Classification of Diseases and the major risk factors that cause those health outcomes. Figure 1.1 presents a simplified version of this framework and indicates the causal chain of events that matter for health outcomes, identifying the key components and determinants of health status that require quantification. Many countries and health development agencies have adopted the GBD approach as the standard for health accounting and for guiding the determination of health research priorities, for example, Australia (Mathers, Vos, and Stevenson 1999); the state of Andra Pradesh, India (Mahapatra 2002); Mauritius (Vos and others 1995); Mexico (Lozano and others 1995); South Africa (Bradshaw and 1 ©2006 The International Bank for Reconstruction and Development / The World Bank 263 Distal socioeconomic and environmental causes Proximal determinants (risk factors) Diseases and injuries Impairments Functional limitations (disability) Nonhealth well-being Death Source: Mathers and others 2002. Note: This presentation is intended as a broad schema: for example, some exposures, such as environmental factors, can be proximate causes of disease, and injuries can lead directly to death. Figure 1.1 Overview of Burden of Disease Framework others 2003); Thailand (Bundhamcharoen and others 2002); Turkey (Baskent University 2005); the United States (McKenna and others 2005); and the World Health Organization (WHO 1996). This chapter begins with a brief history of the work on burden of disease, including a discussion of the nature and origins of the disability-adjusted life year (DALY) as a measure of disease burden. Next it discusses applications of burden of disease analysis to the formulation of health policy. The chapter then summarizes the methods and findings of the 2001 GBD study, reported in more detail in chapters 3 and 4 of this volume. A concluding section takes stock of the work on disease burden since the early 1990s and suggests some key areas for further work. Following this introductory and summarizing chapter, chapter 2 describes the demographic underpinnings for the epidemiological assessments that follow and provides context by briefly reviewing recent changes (from 1990 to 2001) in key demographic parameters. The chapter also assesses changes in the cause distribution of mortality among children under five between 1990 and 2001 and the difficulties of reliably assessing trends in mortality. Chapters 3 and 4 provide the definitive methods and results of the 2001 GBD study. Chapter 3 reports on deaths and the disease and injury burden by age, sex, and 136 disease and injury categories. Chapter 4 reports on the disease and injury burden resulting from 19 risk factors, specifically for a number of important conditions. Both chapters present results using the World Bank’s classification of low- and middle-income countries into six regional groups. Chapter 5 then explores the robustness of the major findings to uncertainties in the data and to alternative assumptions concerning construction of the DALY. Chapter 6 examines the implications of including stillbirths in a global burden of disease assessment. Their inclusion is potentially significant, both because the numbers are large (3.3 million in 2001), and because including stillbirths raises major questions about how to assess the DALY loss associated with deaths near the time of birth. HISTORY OF BURDEN OF DISEASE STUDIES In 1992, the World Bank commissioned the initial GBD study to provide a comprehensive assessment of the disease burden in 1990. The study was undertaken for the world as a whole and for 8 regions (Lopez and Murray 1998; Murray and Lopez 1996a,d; Murray, Lopez, and Jamison 1994; World Bank 1993). In order to recommend intervention packages for countries at different stages of development, the estimates were combined with analyses of the costeffectiveness of interventions in different populations (World Bank 1993; Jamison and Jardel 1994). Whereas earlier attempts to quantify global cause of death patterns (Hakulinen and others 1986; Lopez 1993) were valuable initial contributions to building the evidence base for policy, they were largely restricted to broad cause of death groups, for example, all infections and parasitic diseases combined, and did not address nonfatal health outcomes. The methods and findings of the 1990 GBD study have been widely published and, as noted earlier, have spawned multiple disease burden exercises (Murray and Lopez 1996c,d; 1997a,b,c).One of the basic principles guiding a burden of disease assessment is that almost all sources of health data are likely to contain useful information provided they are carefully screened for validity and completeness.With appropriate methods, investigator commitment, and expert judgment, obtaining internally consistent estimates of the global descriptive epidemiology of major conditions is possible. To prepare internally consistent estimates of incidence, prevalence, duration, and mortality for almost 500 sequelae of the diseases and injuries under consideration, a mathematical model, DisMod, was developed for the 1990 GBD study to convert partial, often nonspecific, data on disease and injury occurrence into a consistent description of the basic epidemiological parameters in each region by age group (Barendregt and others 2003; Murray and Lopez 1996b). Many diseases, for example, neuropsychiatric conditions and hearing loss, and injuries may cause considerable ill health but no or few direct deaths. Therefore separate 2 | Global Burden of Disease and Risk Factors | Alan D. Lopez, Colin D. Mathers, Majid Ezzati, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 264 Box 1.1 Disability-Adjusted Life Years The DALY is a health gap measure that extends the concept of potential years of life lost due to premature death to include equivalent years of healthy life lost by virtue of individuals being in states of poor health or disability (Murray 1996). One DALY can be thought of as one lost year of healthy life and the burden of disease as a measure of the gap between current health status and an ideal situation where everyone lives into old age free from disease and disability. This conceptualization of DALYs as a measure of health, and not of lost utility, is analogous to the principles of measuring gross domestic product as summarized by Eisner (1989, p. 7): “Our focus . . . is on measures of all economic activity related to welfare [for example, gross domestic product], but not of welfare itself.” Information on calculating DALYs, on time discounting, and on age weights is provided in chapter 3. DALYs for a disease or health condition are calculated as the sum of YLL in the population and YLD for incident cases of the health condition. YLL is calculated from the number of deaths at each age multiplied by a global stan- measures of survival and of health status among survivors, while useful inputs when formulating health policy, need to be combined in some fashion to provide a single, holistic measure of overall population health. To assess the burden of disease, the 1990 GBD study used a time-based metric that measures both premature mortality (years of life lost because of premature mortality or YLL) and disability (years of healthy life lost as a result of disability or YLD, weighted by the severity of the disability). The sum of the two components, namely, DALYs, provides a measure of the future stream of healthy life (years expected to be lived in full health) lost as a result of the incidence of specific diseases and injuries in 1990 (box 1.1). The effect of fatal cases (of disease or injury) is captured by years of life lost, while YLD captures the future health consequences in terms of sequelae of diseases or injuries of incident cases in 1990 that were not fatal. (For a more complete account of the DALY measure and the philosophy underlying parameter choices, see Murray 1996; Murray, Salomon, and others 2002). DALYs are not unique to the GBD study. The World Bank used a variant of DALYs in its seminal review of health sector priorities (Jamison and others 1993), and they are derived from earlier work to develop time-based measures that better reflect the public health impact of death or illness dard life expectancy for the age at which death occurs. To estimate YLD for a particular cause for a particular time period, the number of incident cases in that period is multiplied by the average duration of the disease and a weight factor that reflects the severity of the disease on a scale from 0 (perfect health) to 1 (dead). The weights used in the 2001 GBD study are listed in detail elsewhere (see annex tables 3A.6 to 3A.8 in chapter 3). In addition, in calculating DALYs, the GBD study used 3 percent time discounting and non-uniform age weights which give less weight to years lived at young and older ages. For the results reported in this volume and used in the Disease Control Priorities in Developing Countries, second edition (DCP2) 3 percent time discounting was applied but not non-uniform age weights. A death in infancy then corresponds to 30 DALYs, and deaths at age 20 to around 28 DALYs. Thus a disease burden of 3,000 DALYs in a population would be the equivalent of around 100 infant deaths or to approximately 5,000 persons aged 50 years living one year with blindness (disability weight 0.6). at young ages (Dempsey 1947; Ghana Health Assessment Project Team 1981). Much of the comment on, and criticism of, the GBD study focused on the construction of DALYs (Anand and Hanson 1998; Hyder, Rotllant, and Morrow 1998; Williams 1999), particularly the social choices pertaining to age weights and severity scores for disabilities. Relatively little criticism was directed at the vast uncertainty of the basic descriptive epidemiology for some populations, especially in Sub-Saharan Africa (see chapter 5 in this volume), which is likely to be far more consequential for setting health priorities (Cooper and others 1998). The results of the 1990 GBD study confirmed what many health workers had suspected for some time,namely,that noncommunicable diseases and injuries were a significant cause of health burden in all regions, and in some rapidly industrializing regions such as East Asia and Pacific, were already by far the leading cause of death and disability.Neuropsychiatric disorders and injuries in particular were major causes of lost years of healthy life as measured by DALYs, and were vastly underappreciated when measured by mortality alone. The original GBD study estimated that noncommunicable diseases, including neuropsychiatric disorders, caused 41 percent of the global burden of disease in 1990, only slightly less than Measuring the Global Burden of Disease and Risk Factors, 1990–2001 | 3 ©2006 The International Bank for Reconstruction and Development / The World Bank 265 communicable, maternal, perinatal, and nutritional conditions combined (44 percent), and that 15 percent of the burden was due to injuries. Earlier assessments of global health priorities based on mortality data attributed no deaths to mental health disorders and less than half (7 percent) of that suggested by DALYs to injuries (Lopez 1993). Estimates of the disease and injury burden caused by exposure to major risk factors are likely to be a much more useful guide to policies and priorities for prevention than a “league table” of the disease and injury burden. In recent decades, researchers have attempted to quantify the effects of specific exposures, for instance, tobacco smoking, on mortality from major diseases such as cancers (Doll and Peto 1981; Parkin and others 1994) or from multiple diseases (Peto and others 1992; United States Department of Health and Human Services 1992), either in individual countries or across groups of countries using comparable methods. Specific country studies have examined the impact of several leading risk factors (Holman and others 1988; McGinnis and Foege 1993), but prior to the 1990 GBD study, no global assessments of the fatal and nonfatal burden of disease and injury resulting from exposure to multiple major health risks had been attempted. The 1990 study quantified 10 risk factors based on information about causation, prevalence, exposure, and disease and injury outcomes available at the time. The study attributed almost 16 percent of the entire global burden of disease and injury to malnutrition; another 7 percent to poor water and sanitation; and 2 to 3 percent to such risks as unsafe sex, tobacco, alcohol, and occupational exposures (Lopez and Murray 1998; Murray and Lopez 1996a; Murray and Lopez 1997a; Murray, Lopez, and Jamison 1994; World Bank 1993). APPLICATIONS OF BURDEN OF DISEASE ANALYSIS Burden of disease analyses are useful for informing health policy in at least five major ways as outlined in this section. Estimates of deaths by cause or years of life lost serve these same purposes, but for some uses, less well. Assessing Performance The burden of disease provides an indicator that can be used to judge progress over time within a single country or region or relative performance across countries and regions. In this application, burden of disease may be considered analogous to national income and product accounts, developed by Simon Kuznets and others in the 1930s and culminating in 1939 with a complete national income and product account for the United Kingdom prepared at the request of the treasury. In subsequent decades, national income and product accounts have transformed the empirical underpinnings of economic policy analysis. As one leading scholar put it, “The national income and product accounts for the United States . . . , and kindred accounts in other nations, have been among the major contributions to economic knowledge over the past half century . . . Several generations of economists and practitioners have now been able to tie theoretical constructs of income, output, investment, consumption, and savings to the actual numbers of these remarkable accounts with all their fine detail and soundly meshed interrelations” (Eisner 1989, p. 1). Generating Forums for Informed Debate of Values and Priorities In practice, assessing the disease burden involves participation by a broad range of disease specialists, epidemiologists, and often, policy makers. Debating the appropriate values for, say, disability weights or for years of life lost at different ages helps clarify values and objectives for national health policy. Discussing the relationships between diseases and their risk factors in the light of local conditions sharpens consideration of priorities and of programs to address them. Identifying National Control Priorities Many countries now identify a relatively short list of interventions whose full implementation becomes an explicit priority for national political and administrative attention. Examples include interventions to control tuberculosis, poliomyelitis, HIV/AIDS, smoking, and specific micronutrient deficiencies. Because political attention and high-level administrative capacity are in relatively fixed and short supply, the benefits from using those resources will be maximized if they are directed toward interventions that are both cost-effective and aimed at problems associated with a high disease burden. National assessments of disease burden are one input into the process of establishing a shortlist of disease control priorities. Creating Knowledge Medical schools offer a fixed number of instructional hours, and training programs for other levels and types of health workers are similarly limited. A major instrument for 4 | Global Burden of Disease and Risk Factors | Alan D. Lopez, Colin D. Mathers, Majid Ezzati, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 266 implementing health policy priorities is to allocate this fixed time resource well. This implies allocating time to training for interventions where the disease burden is high and costeffective interventions exist. Information on the disease or risk factor burden is also a vital input for informing resource allocation for research and development. In particular, whenever a fixed effort will have a benefit proportional not only to the size of that effort, but also to the size of the problem being addressed, estimates of the disease burden become essential for formulating and implementing research and development priorities. For example, developing a vaccine for a broad range of viral pneumonias would have perhaps hundreds of times the impact of a vaccine against hantavirus infection. Allocating Resources across Health Interventions A key task for priority-setting analyses in health is to create the evidence base to stimulate the reallocation of resources to interventions that, at the margin, will generate the greatest reduction in health loss. When there are major fixed costs in mounting an intervention, as is the case with political and managerial attention for national control priorities, burden estimates are required to improve resource allocation. Similarly, major fixed costs may be associated with the universalization (or major expansion) of an intervention and, if so, the cost-effectiveness of the expansion will depend in part on the size of the burden. IMPROVING THE COMPARATIVE QUANTIFICATION OF DISEASES, INJURIES, AND RISK FACTORS: THE 2001 GBD STUDY The 1990 GBD study represented a major advance in the quantification of the impact of diseases, injuries, and risk factors on population health globally and by region.Government and nongovernmental agencies alike have used its results to argue for more strategic allocations of health resources to disease prevention and control programs that are likely to yield the greatest gains in terms of population health. The results have also greatly increased understanding of the basic descriptive epidemiology of diseases and injuries worldwide. Following publication of the initial results of the GBD study, several national applications of the methods it used have led to substantially more data on the descriptive epidemiology of diseases and injuries becoming available, as well as to improvements in analytical methods and mortality data in a number of countries. By emphasizing substan- tially more sophisticated approaches than in the past to the interpretation and presentation of population health data to policy makers, national burden of disease studies have stimulated efforts to improve and extend the collection of the health information data that are the basis for such analyses. A good example is the Islamic Republic of Iran where, over the last five years, the government has implemented a system of death registration with medical information on the cause of death that has been extended from four provinces initially to include 26, or almost all of the country’s provinces. Another example is the government of Thailand’s extensive verbal autopsy study aimed at addressing major coding deficiencies in Thailand’s national mortality data (Choprapawon and others 2005). Critiques of the original study’s approach, particularly of the methods used to assess the severity weightings for disabling health states, have led to fundamental changes in the way that investigators incorporate health state valuations, that is, the use of population-based rather than expert opinion as used in the 1990 study, and to substantially better methods for improving the cross-national comparability of survey data on health status (Murray, Tandon, and others 2002; Salomon and Murray 2004). Better methods for modeling the relationship between the level of mortality and the broad cause of death structure in populations that are based on proportions rather than rates have led to greater confidence in cause of death estimates for developing countries (Salomon and Murray 2002). In addition, improved population surveillance for some major diseases such as HIV/AIDS, and the wider availability of data from verbal autopsy methods, particularly in Sub-Saharan Africa, have lessened the dependence on models for cause of death estimates, although substantial uncertainty in the use of such data persists. For more details on these and other methodological advances, see chapter 3 in this volume. Perhaps the major methodological progress since the 1990 GBD study has been with respect to the quantification of the disease burden from risk factors. The initial study quantified the population health effects of 10 risk factors, but serious concerns exist about the comparability of the methods and estimates used. Different risk factors have different epidemiological traditions, particularly with regard to the definitions of hazardous exposure, the strength of the evidence on causality, and the availability of epidemiological research on exposure and hazard. As a result, comparability across estimates of the disease burden caused by different risk factors has been difficult to establish. In particular, much of classical risk factor research has treated exposures as dichotomous, with individuals either exposed Measuring the Global Burden of Disease and Risk Factors, 1990–2001 | 5 ©2006 The International Bank for Reconstruction and Development / The World Bank 267 or not exposed, with exposure defined according to an often arbitrary threshold value, for example, systolic blood pressure of 140 millimeters of mercury as the threshold for hypertension. Recent evidence for such continuous exposures as cholesterol, blood pressure, and body mass index suggests that such arbitrarily defined thresholds are inappropriate, because the hazards for these risks decline continuously across the entire range of measured exposure levels, with no obvious threshold (Eastern Stroke and Coronary Heart Disease Collaborative Research Group 1998; Ezzati and others 2004; Rose 1985; WHO 2002). For the 2001 GBD study, a new framework for risk factor assessment was defined that examines changes in the disease burden that would be expected under alternative population distributions of exposure to a risk factor or groups of risk factors (Murray and Lopez 1999). Attributable fractions of disease due to a risk factor were then calculated based on a comparison of the disease burden expected under the current estimated distribution of exposure by age, sex, and region with that expected under a counterfactual distribution of exposure. One such counterfactual distribution was defined for each risk factor as the population distribution of exposure that would lead to the lowest levels of disease burden. Thus, for example, in the case of tobacco, this theoreticalminimum-risk counterfactual exposure would be 100 percent of the population being never-smokers, for overweight and obesity it would be a narrow distribution of body mass index centered around an optimal level of 21 kg/m2 and so on. The distributions of the theoretical-minimum-risk exposure for the risk factors quantified in the World Health Organization’s study of comparative risk assessment (the methodological and empirical basis for the 2001 GBD study) were developed by expert groups for each risk factor based on available scientific knowledge of risk factor hazard. The study also used systematic reviews and analyses of extant sources on risk factor exposure and hazard in an iterative process that increased comparability across risk factors (Ezzati and others 2002, 2004). These methods and results are described in more detail in chapter 4 in this volume. Risk factors may affect disease and injury outcomes through other intermediate factors. For instance, some of the effects of diet and physical activity on cardiovascular diseases are mediated through changes in such intermediate factors as weight, blood pressure, and cholesterol. Risk factors may also affect disease and injury outcomes in combination with one another. For example, people who smoke and have elevated blood pressure and cholesterol have substantially higher probabilities of cardiovascular events. Finally, some risks have common social and behavioral determinants. For instance, members of poor households in rural areas are the most likely to be undernourished, use unsafe water sources, and be exposed to indoor smoke from solid fuels. Because of these epidemiological and social characteristics of risk factor exposure and hazard, policy-relevant analysis should include an assessment of the health benefits of simultaneous reductions in multiple risks. Multicausality also means that a range of interventions can be used for disease prevention, with the specific choices determined by such factors as costs, technology availability, infrastructure, and preferences. A novel aspect of the analysis of risk factors in the 2001 GBD study is the development and application of methods for estimating the disease burden attributable to the combined hazards of multiple risk factors (Ezzati and others 2003). The basic units of analysis in the 1990 GBD study were the eight World Bank regions defined for the World Bank’s (1993) World Development Report 1993. Designed to be geographically contiguous, these regions were nonetheless extremely heterogeneous with respect to health development, for example, the region referred to as Other Asia and Islands included countries with such diverse epidemiological profiles as Myanmar and Singapore. This seriously limited the applicability of these regions to comparative epidemiological assessments. Thus the 2001 GBD study followed a more refined approach. Estimates of overall mortality were first developed for World Health Organization member states using different methods for countries at different stages of health development. The choice of methods was largely determined by the availability of data (Lopez and others 2002). Age- and sex-specific death rates for countries were essentially determined using one of three standard approaches: the use of routine life table methods for countries with complete vital registration; the application of standard demographic methods to correct for underregistration of deaths; or the application of model life tables where no vital registration or survey data on adult mortality were available (Lopez and others 2002; Murray and others 2003). The detailed methodological approaches adopted for estimating cause-specific mortality for countries and the descriptive epidemiology of nonfatal conditions for countries or subregions are described elsewhere (Mathers and others 2002; chapter 3 in this volume). This focus on individual countries as the unit of analysis, as well as the systematic application of standardized approaches for all countries in any given category of data availability, has vastly improved the cross-population comparability of disease and injury quantification. A final major advance of the 2001 GBD study has been the systematic attempts to quantify some of the uncertainty 6 | Global Burden of Disease and Risk Factors | Alan D. Lopez, Colin D. Mathers, Majid Ezzati, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 268 in both national and global assessments of the disease burden (see chapter 5 in this volume). This uncertainty must be taken into account when making cross-national comparisons and needs to be carefully communicated to and interpreted by epidemiologists and policy makers alike. for almost one-fifth of all deaths. In other words, the epidemiological transition from infectious to chronic noncommunicable diseases in this group of countries is already well established and is of major relevance to health planning. Leading Causes of Disability MAJOR FINDINGS OF THE 2001 GBD STUDY This section, and tables 1.1 and 1.2, summarize the principle findings of the 2001 GBD study. More detailed findings are reported in chapters 3 and 4. Global and Regional Mortality Slightly more than 56 million people died in 2001, 10.5 million (or nearly 20 percent) of whom were children younger than five years of age. Almost 4 million children died before 1 month of age, with an additional 3.3 million stillbirths (see chapter 6). Of these child deaths, 99 percent occurred in low- and middle-income countries. Low- and middleincome countries also account for a comparatively large number of deaths at young and middle adult ages: 30 percent of all deaths occur at ages 15 to 59, compared with 15 percent in high-income countries. The causes of death at these ages, as well as in childhood, are thus important for assessing public health priorities. Worldwide, one death in every three is from what the GBD study terms Group I causes (communicable diseases, maternal and perinatal conditions, and nutritional deficiencies) (see table 1.1). This proportion remains almost unchanged from 1990, with one major difference. Whereas HIV/AIDS accounted for only 2 percent of Group I deaths in 1990, it accounted for 14 percent in 2001. Excluding HIV/AIDS, Group I deaths fell from one-third of total deaths in 1990 to less than one-fifth in 2001. Virtually all Group I deaths are in low- and middle-income countries. In low- and middle-countries, Group II causes (noncommunicable diseases) are now responsible for more than 50 percent of deaths in adults ages 15 to 59 in all regions except South Asia and Sub-Saharan Africa, where Group I causes, including HIV/AIDS, remain responsible for onethird and two-thirds of deaths, respectively. Outside these two regions, developing countries are now facing a triple burden of disease from communicable diseases, noncommunicable diseases, and injuries (Group III causes). Among low- and middle-income countries as a group, the three leading causes of death in 2001 included ischemic heart disease and cerebrovascular disease, which together accounted The 1990 GBD study brought the previously largely ignored burden of nonfatal illnesses, particularly neuropsychiatric disorders, to the attention of health policy makers. The findings of the 2001 GBD study, based on updated data and analyses, confirm that disability and states of less than full health caused by diseases and injuries play a central role in determining the overall health status of populations in all regions of the world. Neuropsychiatric conditions, vision disorders, hearing loss, and alcohol use disorders dominate the overall burden of nonfatal disabling conditions. In all regions, neuropsychiatric conditions are the most important causes of disability, accounting for more than 37 percent of YLD among adults aged 15 years and older worldwide. The disabling burden of neuropsychiatric conditions is almost the same for males and females, but the major contributing causes are different. While depression is the leading cause of disability for both males and females, the burden of depression is 50 percent higher for females than males, and females also have higher burdens from anxiety disorders, migraine, and senile dementia. In contrast, the male burden for alcohol and drug use disorders is nearly six times higher than that for females and accounts for a quarter of the male neuropsychiatric burden. More than 85 percent of disease burden from nonfatal health outcomes occurs in low- and middle-income countries, and South Asia and Sub-Saharan Africa account for 40 percent of all YLD. Even though the prevalence of disabling conditions such as dementia and musculoskeletal disease is higher in countries with long life expectancies, this is offset by lower contributions to disability from conditions such as cardiovascular disease, chronic respiratory diseases, and long-term sequelae of communicable diseases and nutritional deficiencies. In other words, people living in developing countries not only face shorter life expectancies than those in developed countries, but also live a higher proportion of their lives in poor health. Burden of Disease and Injuries The results of the 2001 GBD study reinforce some of the conclusions of the 1990 GBD study about the importance of including nonfatal outcomes in a comprehensive assessment Measuring the Global Burden of Disease and Risk Factors, 1990–2001 | 7 ©2006 The International Bank for Reconstruction and Development / The World Bank 269 Table 1.1 Deaths and Burden of Disease by Cause—Low- and Middle-Income Countries, High-Income Countries, and World, 2001 Low- and middle-income Deaths All causes Total number (thousands) Rate per 1,000 population Age-standardized rate per 1,000b Selected cause groups: I. COMMUNICABLE DISEASES, MATERNAL AND PERINATAL CONDITIONS AND NUTRITIONAL DEFICIENCIES Tuberculosis HIV/AIDS Diarrheal diseases Measles Malaria Lower respiratory infections Perinatal conditions Protein-energy malnutrition II. NONCOMMUNICABLE CONDITIONS Stomach cancers Colon and rectum cancers Liver cancer Trachea, bronchus, and lung cancers Diabetes mellitus Unipolar depressive disorders Alcohol use disorders Cataracts Vision disorders, age-related Hearing loss, adult onset Hypertensive heart disease Ischemic heart disease Cerebrovascular disease Chronic obstructive pulmonary disease Cirrhosis of the liver Nephritis and nephrosis Osteoarthritis Congenital anomalies Alzheimer and other dementias III. INJURIES Road traffic accidents Falls Self-inflicted injuries Violence High-income DALYs(3,0)a Deaths 48,351 9.3 11.4 1,386,709 265.7 281.7 17,613 (36.4) 552,376 (39.8) 552 (7.0) 1,590 (3.3) 2,552 (5.3) 1,777 (3.7) 762 (1.6) 1,207 (2.5) 3,408 (7.0) 2,489 (5.1) 241 (0.5) 35,874 (2.6) 70,796 (5.1) 58,697 (4.2) 23,091 (1.7) 39,961 (2.9) 83,606 (6.0) 89,068 (6.4) 15,449 (1.1) 26,023 (53.8) 696 (1.4) 357 (0.7) 505 (1.0) 771 (1.6) 757 (1.6) 10 (<.1) 62 (0.1) 0 (0.0) 0 (0.0) 0 (0.0) 760 (1.6) 5,699 (11.8) 4,608 (9.5) 2,378 (4.9) 654 (1.4) 552 (1.1) 2 (<.1) 477 (1.0) 173 (0.4) 4,715 (9.8) 1,069 (2.2) 316 (0.7) 749 (1.5) 532 (1.1) DALYs(3,0)a 7,891 149,161 8.5 160.6 5.0 128.2 Number in thousands (percent) World Deaths DALYs(3,0)a 56,242 9.1 10.0 1,535 871 249.8 256.5 8,561 (5.7) 18,166 (32.3) 560,937 (36.5) 16 (0.2) 22 (0.3) 6 (<.1) 1 (<.1) 0 (0.0) 345 (4.4) 32 (0.4) 9 (0.1) 219 (0.1) 665 (0.4) 444 (0.3) 23 (<.1) 9 (<.1) 2,314 (1.6) 1,408 (0.9) 130 (<.1) 1,606 (2.9) 2,574 (4.6) 1,783 (3.2) 763 (1.4) 1,208 (2.1) 3,753 (6.7) 2,522 (4.5) 250 (0.4) 36,093 (2.3) 71,461 (4.7) 59,141 (3.9) 23,113 (1.5) 39,970 (2.6) 85,920 (5.6) 90,477 (5.9) 15,578 (1.0) 678,483 (48.9) 9,616 (0.7) 5,060 (0.4) 7,945 (0.6) 10,701 (0.8) 15,804 (1.1) 43,427 (3.1) 11,007 (0.8) 28,150 (2.0) 15,364 (1.1) 24,607 (1.8) 9,969 (0.7) 71,882 (5.2) 62,669 (4.5) 33,453 (2.4) 13,633 (1.0) 9,076 (0.7) 13,666 (1.0) 23,533 (1.7) 9,640 (0.7) 6,868 (87.0) 146 (1.9) 257 (3.3) 102 (1.3) 456 (5.8) 202 (2.6) 3 (<.1) 23 (0.3) 0 (0.0) 0 (0.0) 0 (0.0) 129 (1.6) 1,364 (17.3) 781 (9.9) 297 (3.8) 118 (1.5) 111 (1.4) 3 (<.1) 30 (0.4) 207 (2.6) 129,356 (86.7) 1,628 (1.1) 3,175 (2.1) 1,223 (0.8) 5,397 (3.6) 4,192 (2.8) 8,408 (5.6) 4,171 (2.8) 493 (0.3) 1,525 (1.0) 5,387 (3.6) 1,209 (0.8) 12,390 (8.3) 9,354 (6.3) 5,282 (3.5) 2,146 (1.4) 929 (0.6) 3,786 (2.5) 1,420 (1.0) 7,468 (5.0) 32,891 (58.5) 842 (1.5) 614 (1.1) 607 (1.1) 1,227 (2.2) 960 (1.7) 13 (<.1) 84 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 889 (1.6) 7,063 (12.6) 5,390 (9.6) 2,676 (4.8) 771 (1.4) 663 (1.2) 5 (<.1) 507 (0.9) 380 (0.7) 807,839 (52.6) 11,244 (0.7) 8,236 (0.5) 9,169 (0.6) 16,099 (1.0) 19,997 (1.3) 51,835 (3.4) 15,178 (1.0) 28,643 (1.9) 16,889 (1.1) 29,994 (2.0) 11,178 (0.7) 84,273 (5.5) 72,024 (4.7) 38,736 (2.5) 15,778 (1.0) 10,005 (0.7) 17,452 (1.1) 24,952 (1.6) 17,108 (1.1) 155,850 (11.2) 32,017 (2.3) 13,582 (1.0) 17,674 (1.3) 18,132 (1.3) 471 (6.0) 121 (1.5) 71 (0.9) 126 (1.6) 24 (0.3) 11,244 (7.5) 3,045 (2.0) 1,459 (1.0) 2,581 (1.7) 765 (0.5) 5,186 (9.2) 1,189 (2.1) 387 (0.7) 875 (1.6) 556 (1.0) 167,094 (10.9) 35,063 (2.3) 15,041 (1.0) 20,255 (1.3) 18,897 (1.2) Source: Chapter 3. Notes: Numbers in parentheses indicate percentage of column total. Broad group totals in bold are additive but should not be summed with all other conditions listed in table. a. DALYs (3,0) refer to the version of the DALY based on a 3% annual discount rate and uniform age weights. b. Age-standardized using the WHO World Standard Population. c. Includes only causes responsible for more than 1% of global deaths or DALYs in 2001. of global population health. They also confirm the growing importance of noncommunicable diseases in low- and middle-income countries and highlight important changes in population health in some regions since 1990. HIV/AIDS is now the fourth leading cause of the burden of disease globally and the leading cause in Sub-Saharan Africa, where it is followed by malaria in second place. Seven other Group I causes also appear in the top 10 causes for this 8 | Global Burden of Disease and Risk Factors | Alan D. Lopez, Colin D. Mathers, Majid Ezzati, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 270 region. The epidemiological transition in low- and middleincome countries has resulted in a 20 percent reduction in the per capita disease burden due to Group I causes since 1990. Without the HIV/AIDS epidemic and the associated lack of decline in the burden of tuberculosis, this reduction would have been closer to 30 percent. The per capita disease burden in Europe and Central Asia has increased by nearly 40 percent since 1990, and population health in this region is now worse than all other regions except South Asia and Sub-Saharan Africa. This reflects the sharp increase in adult male mortality and disability in the 1990s, leading to the highest male-female differential in the disease burden in the world. A significant factor in this increase is probably the high level of harmful alcohol consumption among men, which has led to high rates of accidents, violence, and cardiovascular disease. From 1991 to 1994, the risk of premature adult (15 to 59 years) death increased by 50 percent for Russian males. It improved somewhat between 1994 and 1998, but subsequently increased. The burden of noncommunicable diseases is increasing, accounting for nearly half the total global burden of disease, a 10 percent increase from estimated levels in 1990. Almost 50 percent of the adult disease burden in low- and middle income countries is now attributable to noncommunicable diseases. The implementation of effective interventions for Group I diseases, coupled with population aging and the spread of risks for noncommunicable disease in many lowand middle-income countries, are the likely causes of this shift. Ischemic heart disease and stroke dominate the burden of disease in Europe and Central Asia and together account for more than a quarter of the total disease burden. In contrast, in Latin America and the Caribbean these diseases account for 8 percent of the disease burden, but this region also has high levels of diabetes and endocrine disorders compared with other regions. Violence is the fourth leading cause of the disease and injury burden in Latin America and the Caribbean. Violence does not appear among the top 10 causes of burden in any other region, but is nonetheless significant. Injuries primarily affect young adults and often result in severe, disabling sequelae. All forms of injury accounted for 16 percent of the adult burden in 2001. In parts of Europe and Central Asia, Latin America and the Caribbean, and the Middle East and North Africa, more than 30 percent of the entire disease and injury burden among male adults aged 15 to 44 is attributable to injuries. Road traffic accidents, violence, and self-inflicted injuries are all among the top 10 leading causes of burden in these regions. The former Soviet Union and other high-mortality (among adults) countries of Eastern Europe have rates of injury death and disability among males that are similar to those in Sub-Saharan Africa. Burden of Disease Attributable to Risk Factors As described earlier, a major advance of the 2001 GBD study has been in creating a unified framework for quantifying the burden of disease and injury attributable to major risk factors and in applying this framework to exposure and hazard data for selected major risk factors based on comprehensive and systematic reviews of published literature and other sources. Notwithstanding the inherent uncertainties in assessing the population-level health effects of risk factors, the quantification of the burden of disease attributable to the individual and joint hazards of selected risks suggests that the leading causes of mortality and disease burden include risk factors for Group I conditions (for example, undernutrition; indoor smoke from household use of solid fuels; poor water, sanitation, and hygiene; and unsafe sex), whose burden is primarily concentrated in South Asia and Sub-Saharan Africa, and risk factors for Group II conditions (especially, smoking, alcohol, high blood pressure and cholesterol, and overweight and obesity), which are widespread globally (see table 1.2). In low- and middle-income countries, the leading causes of disease burden included risk factors prevalent among the poor and associated with Group I conditions (for example, childhood underweight [8.7 percent of the disease burden in these regions]; unsafe water, sanitation, and hygiene [3.7 percent]; and indoor smoke from household use of solid fuels [3.0 percent]), unsafe sex (5.8 percent), and risk factors for noncommunicable diseases (for example, high blood pressure [5.6 percent], smoking [3.9 percent], and alcohol use [3.6 percent]). Across high-income countries, risk factors associated with Group II and Group III conditions were the leading causes of loss of healthy life (smoking [12.7 percent], high blood pressure [9.3 percent], overweight and obesity [7.2 percent], high cholesterol [6.3 percent], and alcohol use [4.4 percent]). An estimated 45 percent of global mortality and 36 percent of the global burden of disease were attributable to the joint hazards of the 19 selected global risk factors. The joint hazards were even larger in regions where a relatively small number of diseases and their risk factors were responsible for large losses of life (HIV/AIDS and risk factors for child mortality in Sub-Saharan Africa; cardiovascular risks, including smoking and alcohol use in Europe and Central Asia). Globally, large fractions of major diseases such as diarrhea, lower respiratory infections, HIV/AIDS, lung cancer, Measuring the Global Burden of Disease and Risk Factors, 1990–2001 | 9 ©2006 The International Bank for Reconstruction and Development / The World Bank 271 Table 1.2 Deaths and Burden of Disease Attributable to Risk Factors—Low- and Middle-Income Countries, High-Income Countries, and World, 2001 Low- and middle-income Deaths Total number (thousands) Rate per 1,000 population Age-standardized rate per 1,000b 48,351 9.3 11.4 High-income DALYs(3,0)a Deaths 1,386 709 265.7 281.7 World DALYs(3,0)a 7,891 8.5 5.0 Deaths 149,161 160.6 128.2 DALYs(3,0)a 56,242 9.1 10.0 1,535 871 249.8 256.5 Number in thousands (percent) Risk factor Childhood and maternal undernutrition Childhood underweight Iron-deficiency anemia Vitamin A deficiency Zinc deficiency 3,630 (7.5) 613 (1.3) 800 (1.7) 849 (1.8) 120,579 (8.7) 23,933 (1.7) 24,686 (1.8) 27,631 (2.0) 0 (0.0) 8 (0.1) 0 (0.0) 0 (0.0) 67 (<0.1) 789 (0.5) 0 (0.0) 5 (<0.1) 3,630 (6.5) 621 (1.1) 800 (1.4) 849 (1.5) 120,647 (7.9) 24,722 (1.6) 24,686 (1.6) 27,636 (1.8) 6,223 (12.9) 3,038 (6.3) 1,747 (3.6) 2,308 (4.8) 1,559 (3.2) 78,063 (5.6) 42,815 (3.1) 31,515 (2.3) 32,836 (2.4) 22,679 (1.6) 1,392 (17.6) 842 (10.7) 614 (7.8) 333 (4.2) 376 (4.8) 13,887 (9.3) 9,431 (6.3) 10,733 (7.2) 3,982 (2.7) 4,732 (3.2) 7,615 (13.5) 3,880 (6.9) 2,361 (4.2) 2,641 (4.7) 1,935 (3.4) 91,950 (6.0) 52,246 (3.4) 42,248 (2.8) 36,819 (2.4) 27,411 (1.8) Addictive substances Smoking Alcohol use Illicit drug use 3,340 (6.9) 1,869 (3.9) 189 (0.4) 54,019 (3.9) 49,449 (3.6) 7,890 (0.6) 1,462 (18.5) 24 (0.3) 37 (0.5) 18,900 (12.7) 6,580 (4.4) 2,024 (1.4) 4,802 (8.5) 1,893 (3.4) 226 (0.4) 72,919 (4.7) 56,029 (3.6) 9,914 (0.6) Sexual and reproductive health Unsafe sex Non-use and use of ineffective methods of contraception 2,819 (5.8) 162 (0.3) 80,270 (5.8) 7,411 (0.5) 32 (0.4) 0 (0.0) 909 (0.6) 23 (<0.1) 2,851 (5.1) 162 (0.3) 81,179 (5.3) 7,434 (0.5) Environmental risks Unsafe water, sanitation, and hygiene Urban air pollution Indoor smoke from household use of solid fuels 1,563 (3.2) 735 (1.5) 1,791 (3.7) 51,622 (3.7) 8,707 (0.6) 41,731 (3.0) 4 (<0.1) 76 (1.0) 0 (0.0) 289 (0.2) 664 (0.4) 2 (<0.1) 1,567 (2.8) 811 (1.4) 1,791 (3.2) 51,911 (3.4) 9,371 (0.6) 41,734 (2.7) 407 (0.8) 65 (0.1) 8,974 (0.6) 5,381 (0.4) 4 (<0.1) 6 (<0.1) 76 (<0.1) 699 (0.5) 412 (0.7) 71 (0.1) 9,050 (0.6) 6,079 (0.4) 22,014 (45.6) 500,066 (36.1) 3,473 (44.0) 51,092 (34.3) 25,488 (45.3) 551,158 (35.9) Other nutrition-related risk factors and physical activity High blood pressure High cholesterol Overweight and obesity Low fruit and vegetable intake Physical inactivity Other selected risks Contaminated injections in health care setting Child sexual abuse All selected risk factors together Source: Chapter 4. Note that mortality and disease burden attributable to individual risk factors cannot be added due to multi-causality. See Chapter 4 for details. a. (some footnote as Table 1.1) b. Age-standardized using the WHO World Standard Population chronic obstructive pulmonary disease, ischemic heart disease, and stroke were attributable to the joint effects of the risk factors considered in this volume. The joint hazards of these 19 risks for a number of other important diseases and injuries, such as perinatal and maternal conditions, selected other cancers, and intentional and unintentional injuries, which have more diverse risk factors, were smaller, but nonnegligible. The relatively small number of risk factors that account for a large fraction of the disease burden underscores the need for policies, programs, and scientific research to take advantage of interventions for multiple major risks to health (Ezzati and others 2003). CONCLUSIONS The substantial scientific and policy interest in the methods and findings of the 1990 GBD study, the widespread application of the methods by countries at all levels of health development, and the adoption of the framework as the preferred method for health accounting by international health agencies such as the World Health Organization attest to the critical need for objective and systematic assessments of the disease burden for priority setting in health. The vast and comprehensive effort to quantify the disease burden worldwide dramatically changed views about the 10 | Global Burden of Disease and Risk Factors | Alan D. Lopez, Colin D. Mathers, Majid Ezzati, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 272 importance of some conditions, particularly psychiatric disorders, and drew global public health attention to the unrecognized burden of injuries. The methodological developments over the past decade, a more systematic approach to collecting key data and research findings on the health of populations, and the results of numerous national and subnational burden of disease studies have dramatically improved the methodological armamentarium and the empirical base for disease burden assessment, in particular, the comparability of the estimated contributions of diseases, injuries, and risk factors to this burden. As reported in this volume, the 2001 GBD study provides a comprehensive update of the comparative importance of diseases, injuries, and risk factors for global health. The study incorporates a range of new data sources to develop internally consistent estimates of incidence, prevalence, severity and duration, and mortality for 136 major causes by sex and by eight age groups. Estimates of deaths by cause, age, and sex were carried out separately for 226 countries and territories, drawing on a total of 770 country-years of death registration data, 535 additional sources of information on levels of child and adult mortality, and more than 2,600 data sets providing information on specific causes of death in regions not well covered by death registration systems. Together with the more than 8,500 data sources (epidemiological studies, disease registers, notifications systems, and so on) used to estimate incidence, prevalence, and YLD by cause, the 2001 GBD study has incorporated information from more than 10,000 datasets relating to population health and mortality (see chapter 3). This represents one of the largest syntheses of global information on population health carried out to date. Much of the research on the burden of disease undertaken over the past decade or so has relied on the methodological and empirical efforts that defined the 1990 GBD study as a major advance in global public health statistics. Progress in updating the epidemiological basis for assessing the disease burden from the various diseases and injuries of interest has been uneven, although improvements in the data and methods available for assessing global and regional mortality by cause have been substantial, and some advances have been made in the data for, and epidemiological understanding of some major causes of ill health such as HIV/AIDS and diabetes mellitus. Nevertheless, making more reliable estimates of global, regional, and national disease burdens still faces many methodological and empirical challenges. The substantive agenda, mapped out over a decade ago (Murray, Lopez and Jamison, 1994) remains equally valid today and needs to be addressed more systematically if the burden of disease framework is to gain greater acceptance as the international tool for health accounting. Assessing and documenting in detail the state of the world’s health at the beginning of the millennium is a useful undertaking. This volume will provide scholars today and in the future with a definitive historical record of the leading causes of the burden of disease for major regions of the world at the start of the 21st century. An account of global health at the beginning of the 20th century, or earlier, would no doubt have been of more than just historical interest, but given the methods of scientific interchange and the state of scientific and methodological knowledge at the time, this was impossible. In presenting the comprehensive findings of the 2001 GBD study, this volume is, in many respects, a culmination of the effort launched in 1990 and represents the end of the beginning of global disease burden assessments. The widespread use of disease burden concepts by national and international bodies since the first results were published and the heightened interest in improving the basic descriptive epidemiology of diseases, injuries, and risk factors by both countries and agencies has laid the foundations for future population health assessments. As programs and policies to improve health worldwide become more widespread, so too will the need for more comprehensible, credible, and comparable assessments to periodically monitor world health and the success, or otherwise, of measures to promote health and reduce the burden of disease. New initiatives, and perhaps new global institutions, are required to measure the burden of disease worldwide and how it is changing, more reliably than hitherto. This book provides the baseline against which such progress with global health development will be measured. REFERENCES Anand, S., and K. Hanson. 1998. “DALYs: Efficiency Versus Equity.” World Development 26 (2): 307–10. Barendregt, J. J., G. J. van Oortmarssen, T. Vos, and C. J. L. Murray. 2003. “A Generic Model for the Assessment of Disease Epidemiology: The Computational Basis of DisMod II.” Population Health Metrics 1 (1): e4. Baskent University. 2005. Burden of Disease Final Report. Ankara, Turkey: Baskent University and the School of Public Health, Refik Saydam Hygiene Center, Ministry of Health. Bradshaw, D., P. Groenewald, R. Laubscher, N. Nannan, B. Nojilana, R. Norman, D. Pieterse, M. Schneider, D. E. Bourne, I. M. Timaeus, R. Dorrington, and L. Johnson. 2003. Initial Burden of Disease Estimates for South Africa, 2000. Cape Town: South African Medical Research Council. http://www.mrc.ac.za/bod/bod.htm. Measuring the Global Burden of Disease and Risk Factors, 1990–2001 | 11 ©2006 The International Bank for Reconstruction and Development / The World Bank 273 Bundhamcharoen, K., Y. Teerawatananon, T. Vos, and S. Begg. 2002. Burden of Disease and Injuries in Thailand: Priority Setting for Policy. Bangkok: Ministry of Public Health. Lozano, R., C. J. L. Murray, J. Frenk, and J. Bobadilla. 1995. “Burden of Disease Assessment and Health System Reform: Results of a Study in Mexico.” Journal for International Development 7 (3): 555–64. Choprapawon, C., Y. Porapakkham, O. Sablon, R. Panjajaru and B. Jhantharatat. 2005. “Thailand’s National Death Registration Reform: Verifying the Causes of Death between July 1997 and December 1999.” Asia-Pacific Journal of Public Health 17 (2): 110–116. Mahapatra, P. 2002. Estimating National Burden of Disease: The Burden of Disease in Andhra Pradesh, 1990s. Hyderabad, India: Institute of Health Systems. Cooper, R. S., B. Osotimehin, J. S. Kaufman, and T. Forrester. 1998. “Disease Burden in Sub-Saharan Africa: What Should We Conclude in the Absence of Data?” Lancet 351 (9087): 208–10. Mathers, C. D., T. Vos, and C. Stevenson. 1999. The Burden of Disease and Injury in Australia. Canberra: Australian Institute of Health and Welfare. http://www.aihw.gov.au/publications/index.cfm/title/5180. Doll, R., and R. Peto. 1981. The Causes of Cancer. Oxford, U.K.: Oxford University Press. Mathers, C. D., C. Stein, M. Ma Fat, C. Rao, M. Inoue, K. Shibuya, N. Tomijima, C. Bernard, and H. Xu. 2002. The Global Burden of Disease 2000 Study (version 2): Methods and Results. Discussion Paper 50. Geneva: Global Program on Evidence for Health Policy, World Health Organization. http://www.who.int/evidence. Eastern Stroke and Coronary Heart Disease Collaborative Research Group. 1998. “Blood Pressure, Cholesterol, and Stroke in Eastern Asia.” Lancet 352 (9143): 1801–7. McGinnis, J. M., and W. H. Foege. 1993. “Actual Causes of Death in the United States.” Journal of the American Medical Association 270 (18): 2207–12. Eisner, R. 1989. The Total Incomes System of Account. Chicago and London: University of Chicago Press. McKenna, M. T., C. M. Michaud, C. J. L. Murray, and J. S. Marks. 2005. “Assessing the Burden of Disease in the United States Using DisabilityAdjusted Life Years.” American Journal of Preventive Medicine 28 (5): 415–23. Dempsey, M. 1947. “Decline in Tuberculosis: The Death Rate Fails to Tell the Entire Story.” American Review of Tuberculosis 56 (2): 157–64. Ezzati, M., A. D. Lopez, A. Rodgers, and C. J. L. Murray. 2004. Comparative Quantification of Health Risks: The Global and Regional Burden of Disease Attributable to Selected Major Risk Factors. Geneva: World Health Organization. Ezzati, M., A. D. Lopez, A. Rodgers, S. Vander Hoorn, C. J. L. Murray, and the Comparative Risk Assessment Collaborating Group. 2002. “Selected Major Risk Factors and Global and Regional Burden of Disease.” Lancet 360 (9343): 1347–60. Murray, C. J. L. 1996. “Rethinking DALYs.” In The Global Burden of Disease, ed. C. J. L. Murray and A. D. Lopez, 1–89. Cambridge, MA: Harvard University Press . Murray, C. J. L., and A. D. Lopez. 1996a. “Evidence-Based Health Policy: Lessons from the Global Burden of Disease Study.” Science 274 (5288): 740–43. Ezzati, M., S. Vander Hoorn, A. Rodgers, A. D. Lopez, C. D. Mathers, C. J. L. Murray, and the Comparative Risk Assessment Collaborating Group. 2003. “Estimates of Global and Regional Potential Health Gains from Reducing Multiple Major Risk Factors.” Lancet 362 (9380): 271–80. . 1996b. “Global and Regional Descriptive Epidemiology of Disability: Incidence, Prevalence, Health Expectancies, and Years Lived with Disability.” In The Global Burden of Disease, ed. C. J. L. Murray and A. D. Lopez, 201–46. Cambridge, MA: Harvard University Press. Ghana Health Assessment Project Team. 1981. “Quantitative Method of Assessing the Health Impact of Different Diseases in Less Developed Countries.” International Journal of Epidemiology 10 (1): 73–80. . 1996c. Global Health Statistics: A Compendium of Incidence, Prevalence, and Mortality Estimates for over 200 Conditions. Cambridge, MA: Harvard University Press. Hakulinen, T., H. Hansluwka, A. D. Lopez, and T. Nakada. 1986. “Global and Regional Mortality Patterns by Cause of Death in 1980.” International Journal of Epidemiology 15 (2): 226–33. , eds. 1996d. The Global Burden of Disease, vol. 1. Cambridge, MA: Harvard University Press. Holman, C. D. J., B. K. Armstrong, L. N. Arias, C. A. Martin, W. M. Hatton, L. D. Hayward, M. A. Salmon, R. E. Shean, V. P. Waddell. 1990. The Quantification of Drug Caused Morbidity and Mortality in Australia 1988. Canberra: Commonwealth Department of Community Services and Health. Hyder, A. A., G. Rotllant, and R. Morrow. 1998. “Measuring the Burden of Disease: Healthy Life Years.” American Journal of Public Health 88 (2): 196–202. . 1997a. “Global Mortality, Disability, and the Contribution of Risk Factors: Global Burden of Disease Study.” Lancet 349 (9063): 1436–42. . 1997b. “Mortality by Cause for Eight Regions of the World: Global Burden of Disease Study.” Lancet 349 (9061): 1269–76. . 1997c. “Regional Patterns of Disability-Free Life Expectancy and Disability-Adjusted Life Expectancy: Global Burden of Disease Study.” Lancet 349 (9062): 1347–52. . 1999. “On the Comparable Quantification of Health Risks: Lessons from the Global Burden of Disease Study.” Epidemiology 10 (5): 594–605. Jamison, D. T., and J.-P. Jardel. 1994. “Comparative Health Data and Analyses.” In Global Comparative Assessments in the Health Sector: Disease Burden, Expenditures, and Intervention Packages, ed. C. J. L. Murray and A. D. Lopez, v–vii. Geneva: World Health Organization. Murray, C. J. L., A. D. Lopez, and D. T. Jamison. 1994. “The Global Burden of Disease in 1990: Summary Results, Sensitivity Analyses, and Future Directions.” Bulletin of the World Health Organization 72 (3): 495–508. Jamison, D. T., W. H. Mosely, A. R. Measham, and J. L. Bobadilla, eds. 1993. Disease Control Priorities in Developing Countries. New York: Oxford University. Murray, C. J. L., J. A. Salomon, C. D. Mathers, and A. D. Lopez. 2002. Summary Measures of Population Health: Concepts, Ethics, Measurement, and Applications. Geneva: World Health Organization. Lopez, A. D. 1993. “Causes of Death in the Industrialized and Developing Countries: Estimates for 1985–1990.” In Disease Control Priorities in Developing Countries, ed. Dean Jamison, W. Henry Mosely, A. R. Measham, and J. L. Bobadilla, 15–30. New York: Oxford University Press. Murray, C. J. L., A. Tandon, J. A. Salomon, C. D. Mathers, and R. Sadana. 2002. “New Approaches to Enhance Cross-Population Comparability of Survey Results”. In Summary Measures of Population Health: Concepts, Ethics, Measurement, and Applications, ed. C. J. L. Murray, J. A. Salomon, C. D. Mathers, and A. D. Lopez, 421–32. Geneva: World Health Organization. Lopez, A. D., and C. J. L. Murray. 1998. “The Global Burden of Disease, 1990–2020.” Nature Medicine 4 (11): 1241–43. Lopez, A. D., O. B. Ahmad, M. Guillot, B. D. Ferguson, J. A. Salomon, C. J. L. Murray, and K. Hill. 2002. World Mortality in 2000: Life Tables for 191 Countries. Geneva: World Health Organization. Murray, C. J. L., B. D. Ferguson, A. D. Lopez, M. Guillot, J. A. Salomon, and O.B. Ahmad. 2003. “Modified Logit Life Table System: Principles, Empirical Validation, and Application.” Population Studies 57 (2): 165–82. 12 | Global Burden of Disease and Risk Factors | Alan D. Lopez, Colin D. Mathers, Majid Ezzati, and others ©2006 The International Bank for Reconstruction and Development / The World Bank 274 Parkin, D. M., P. Pisani, A. D. Lopez, and E. Masuyer. 1994. “At Least One in Seven Cases of Cancer Is Caused by Smoking: Global Estimates for 1985.” International Journal of Cancer 59 (4): 494–504. Peto, R., A. D. Lopez, J. Boreham, M. Thun, and C. Heath. 1992. “Mortality from Tobacco in Developed Countries: Indirect Estimates from National Vital Statistics.” Lancet 339 (8804): 1268–78. Rose, G. 1985. “Sick Individuals and Sick Populations.” International Journal of Epidemiology 14 (1): 32–38. Salomon, J. A., and C. J. L. Murray. 2002. “The Epidemiologic Transition Revisited: Compositional Models for Causes of Death by Age and Sex.” Population and Development Review 28 (2): 205–28. . 2004. “A Multimethod Approach to Measuring Health State Valuations.” Health Economics 13: 281–90. United States Department of Health and Human Services. 1992. Smoking and Health in the Americas. Report of the Surgeon General, in Collaboration with the Pan-American Health Organization, Department of Health and Human Services publication (CDC) 92-8419. Washington, DC: Office on Smoking and Health. Vos, T., M. Tobias, H. Gareeboo, F. Roussety, S. Huttley, and C. J. L. Murray. 1995. Mauritius Health Sector Reform, National Burden of Disease Study, Final Report of Consultancy. Port Louis, Mauritius: Ministry of Health and Ministry of Economic Planning and Development. Williams, A. 1999. “Calculating the Global Burden of Disease: Time for a Strategic Appraisal?” Health Economics 8 (1): 1–8. World Bank. 1993. Investing in Health: World Development Report 1993. New York: Oxford University Press. WHO (World Health Organization). 1996. Investing in Health Research and Development. Report of the Ad Hoc Committee on Health Research Relating to Future Intervention Options. Geneva: WHO. . 2002. Reducing Risks: Promoting Healthy Life. World Health Report 2002. Geneva: WHO. Measuring the Global Burden of Disease and Risk Factors, 1990–2001 | 13 ©2006 The International Bank for Reconstruction and Development / The World Bank 275 ©2006 The International Bank for Reconstruction and Development / The World Bank 276