lEGTlll'ill'E'NllTlES

EPIDEMIMUGY
Second Edition

fllllllll RflSlIill Khan

IUL Naravan
Published by
4 Sepoy Lines
10450 Penang, Malaysia
Tel: +604 226 3459
Fax: +604 228 4285
Website: http://www.pmc.edu.my

Designed and Layout by

ACE DESIGN & PRINTING
No.5, Jalan Putra 1, Taman Sri Putra,
Kempas, 81200 Johor Bahru, Johor, Malaysia.
Tel: +6016-608 4808 / +607-562 6264
Fax: +607-554 3144
E-mail: ace1printers@gmail.com

All rights reserved.

Copyright © Abdul Rashid Khan
Design © ACE Design & Printing.

No part of this book may be reproduced in any form or by any means without
prior permission from Abdul Rashid Khan.

First Print 2010

Perpustakaan Negara Malaysia Cataloguing-in-Publication Data
Abdul Rashid Khan
Lecture Notes on Epidemiology / Abdul Rashid Khan, K. A. Narayan.
ISBN 978-967-950-302-9
1. Epidemiology. I. Narayan,K.A II. Title.
614.4

i
Foreword

Epidemiology is the bread and butter of public health. It contributes the knowl
edge of the what, where, when, why and how of diseases and its occurrence. Stu
dents pursuing medicine must be equipped with the epidemiology knowledge and
skills to be good medical and health practitioners. It will help strengthen the prac
tice of evidence-based medicine and public health. Many books on epidemiology
have been written for this purpose. However, to be more relevant for the local set
ting, epidemiology books must include data and examples from the local medical
and health scene. This is then the value of a local book on epidemiology, where lo
cal data and examples brings relevance and meaningfulness for the local medical
and health students.
In Malaysia, local books in epidemiology are lacking. This book, ‘Lecture
Notes on Epidemiology’ Second Edition is written by Professor Dr Abdul Rashid
Khan and Professor K.A. Narayan, both of whom are experienced medical teach
ers in community medicine. It is the improved and refined version of the first edi
tion. The second edition maintains the qualities of an excellent text book for under
graduate medical students in Malaysia. This book will help build a strong epidemi
ology base for their future medical and health careers. This edition continues to
provide local relevance and meaningfulness by incorporating local epidemiology
data and scenarios. It covers many important areas of epidemiology and is written
in a lecture note format making it simple for students to understand and follow.
While the book is written mainly for undergraduate medical students, the book
is also suited for teachers and practitioners of public health and for other under
graduate students in the field of medical and health sciences. I am sure many stu
dents in Malaysia will benefit from this book and will go on to become good medi
cal and heath practitioners for the country. I must congratulate both the authors for
making the second edition of this book possible and hope that their effort will en
courage others to write such books in other fields of public health.

ii
Professor Dr Zulkifli bin Ahmad,
Department’of Community Medicine, School of Medical Sciences
Deputy Dean (Research & Postgraduate studies), School of Dental Sciences
Universiti Sains Malaysia,
Kubang Kerian, Kelantan.

iii
Preface

“Lecture Notes on Epidemiology” was originally written with the view of helping
undergraduates have a basic knowledge of the subject. Since the book was pub
lished in 2007, many students have given us valuable feedback. However, what
was pleasantly surprising was when many colleagues, both academicians and
health service personnel, informed us that they found the book helpful in their
work.
We had the choice of either reprinting the first edition or writing a second edi
tion. We chose the latter, as based on the feedback received, we felt changes could
be made to make the book better and more up to date. More examples, explana
tions and newer data have been added. References to websites which give addi
tional reading have been included. The main focus of the book, however, remains
i.e keeping its relevance to Malaysia.
We do not purport this to be a text book on epidemiology but rather an easy
guide to improve the understanding of epidemiology. Much of the material of this
book has been sourced from text books, journals and internet. We encourage the
readers to refer to the source materials of this book which has been mentioned in
the text, figures/tables, reference section and in the additional reading section.
We fervently hope that this book will benefit the students as well as practicing
doctors.
Abdul Rashid Khan
K.A. Narayan

iv
Acknowledgement

We are indebted to Professor Amir S. Khir, the Dean and President of Penang
Medical College, who agreed to publish the reprint as an e-book and to allow read
ers to access it for free.
We dedicate this book to our families; Dr. Azizah, Sarah Diyanah and Pushpa
Narayan, for their patience and support. Finally, a note of thanks to all our col
leagues and students for their invaluable suggestions and comments.
Abdul Rashid Khan
K.A. Narayan

v
 CHAPTER 1

Introduction to
Epidemiology
Man has always lived with disease and it has been his endeavour to conquer it.
Medical knowledge has been derived to a great degree from intuitive and observa
tional propositions and cumulative experiences gleaned from others.
Health and disease can be studied in three basic ways, (i) observation of effects
on individuals, (ii) laboratory experiments and (iii) measuring their distribution in
population (epidemiology).
The origin of the word epidemiology is from the Greek word “epi” meaning
“upon”, “demos” meaning “people” and “logos” meaning “doctrine”, the literal
translation would be “the doctrine of what is upon the people”.
The International Epidemiological Association defines epidemiology as “the
study of the distribution and determinants of health related states and events in the
populations and the application of this study to the control of health problems”.
The primary unit of concern is groups of persons not individuals.

HISTORICAL PERSPECTIVE
Medicine has drawn richly from, firstly traditional cultures of which it is a
part, later from biological and social sciences, and more recently from social and
behavioural sciences. Medicine has now evolved into a social system heavily bu
reaucratized and politicized. It is now more complex and the costs of health care
are ever increasing.

6
 In ancient times, health and illness were interpreted in cosmological and anthro
pological perspective. Medicine was dominated by magical and religious beliefs
which were an integral part of ancient cultures and civilizations. Later the belief
of disease causation shifted from spiritual to environmental factors.
Hippocrates attempted to explain disease occurrence from a rational in
stead of a supernatural viewpoint. He suggested that environmental and host fac
tors such as behaviours might influence the development of disease.
John Graunt (April 24, 1620 - April 18, 1674) is often regarded as the founder
of vital statistics. He quantified patterns of disease, birth and death, noted male
female disparities, high infant mortality, urban-rural differences and seasonal
variations.
Dr. John Snow, (15 March 1813 – 16 June 1858) a British physician
and a leader in the adoption of anaesthesia and medical hygiene, is considered
to be one of the fathers of epidemiology, because of his work in tracing the
source of a cholera outbreak in Soho, England, in 1854, long before the organ
ism was found by Robert Koch. Snow studied the epidemics of cholera in 1849
and hypothesized that the disease was caused by polluted drinking water, con
trary to the ‘Miasma’ theory prevalent at the time. In 1854, Snow used a spot
map to illustrate how cases of cholera were centred around the pump in
Broadstreet. He also made a solid use of statistics to illustrate the connection
between the quality of the source of water and cholera cases. He showed that com
panies taking water from sewage-polluted sections of the Thames delivered wa
ter to homes with an increased incidence of cholera. Snow’s study was a major
event in the history of public health and can be regarded as the founding event of
the science of epidemiology.
In the 1800’s William Farr, (November 30, 1807 - April 14, 1883) considered
as the father of modern vital statistics and surveillance, began to systematically
collect and analyze Britain’s mortality statistics. His most important contribution
was to set up a system for routinely recording the causes of death. Although he did
not agree with Snow’s waterborne theory, he gave him a great deal of help in col
lecting data to support it, in particular by providing the addresses of people who
had died.
7
 The more recent development of epidemiology can be illustrated by the work
of Sir William Richard Shaboe Doll (28 October 1912 – 24 July 2005) a British
physiologist who became the foremost epidemiologist of the 20th century, turning
the subject into a rigorous science. He along with Sir Austin Bradford Hill (July 8,
1897 - April 18, 1991), English epidemiologist and statistician, studied the rela
tionship between cigarette smoking and lung cancer. Studies such as these brought
in the important concept of risk. Hill pioneered the randomized clinical trial.
The Framingham Heart Study which started in 1948 saw the influence of
“many exposures” such as smoking, obesity and elevated blood pressure on the
outcome measure i.e. Coronary Heart Disease.

THE EPIDEMIOLOGICAL PERSPECTIVE

Epidemiology is about information: the information needed for health plan
ning, supervision and evaluation of the health-promotion and disease-control ac
tivities. The key components of the data needed can be approached through a se
ries of questions.

Mr. How returns from his epidemiological expedition with a new wife

8
‣Who? – Who is affected? Referring to age, sex, social class, ethnic group, oc
cupation, heredity and personal habits. (These are person factors).
‣Where? – Where did it happen? In relation to place of residence, geographi
cal distribution and place of exposure (Place factors).
‣When? – When did it happen? In terms of months, seasons or years (Time fac
tors).
‣What? – What is the disease or condition, its clinical manifestation and diag
nosis?
‣How? – How did the disease occur? In relation to: the interplay of the spe
cific agent, vector and source of infection, susceptible groups and other con
tributing factors.
‣Why? – Why did it occur? In terms of the reasons for the disease outbreak.
‣What now? – The most important question - What action is now to be taken
as a result of the information gained?

SCOPE OF EPIDEMIOLOGY
Epidemiology covers all major health problems in the community including:
‣Communicable diseases
‣Chronic degenerative, metabolic, neoplastic diseases
‣Nutritional deficiencies
‣Occupational health and injuries
‣Mental and behavioural disorders
‣Population issues and demographic trends

9
USES OF EPIDEMIOLOGY
‣The most important use of epidemiology is to increase the understanding of
disease by looking at communities or populations. It is used to determine the
health of a population by the design, conduct and interpretation of studies
(See Community Diagnosis in Chapter 7).

‣Epidemiology is concerned with describing the natural history of disease, in

cluding not only the clinical stages seen in hospitals and medical practice but
inapparent, sub-clinical and carrier states and precursor states of chronic dis
eases.
‣Epidemiology is used to monitor the health of populations (surveillance) to
chart changes over time, place and person as well as determine which dis
eases are of most public health importance. By analysing trends it is able to
predict and devise methods of control. For example, in the 1900’s infectious
diseases were common but these have been replaced by chronic diseases of
long duration (See chapter “Patterns of disease”).
‣Epidemiology attempts to identify causative agents, the factors in the
web of causation, the populations at highest risk, environmental and other
determinants. For example, one can determine what are the contributory fac
tors for lung cancer, eg. smoking, occupations etc, determine what is the risk
of developing a disease in the presence of that factor (risk due to smoking)
and what can be achieved by removing that risk (benefits of cessation of
smoking).
‣It supplies information necessary for health planning and development, man
agement of programmes of disease prevention and control and supplies tools
for evaluating health programmes (Operations research).
‣It provides a foundation for public policy and for making regulatory deci
sions especially those relating to environmental problems.

10
THE EPIDEMIOLOGICAL RESPONSIBILITIES OF THE MEDICAL
OFFICER
The responsibility of the medical officer extends to the entire community and
includes the healthy, the sick, those seeking help as well as those who do not and
the unborn or dead.
The clinician examines the patient and has to recognise and identify the pathog
nomonic significance of the clinical signs and symptoms to reach a specific diag
nosis and to prescribe the appropriate treatment. The epidemiologist looks at the
population and has to select the diagnostic indicators most suitable for case defini
tion of the diseases in that population; he must pre-select the methods and tests
which can be applied for mass diagnosis. Neither approach is self-sufficient; they
complement each other in the overall approach to solving health problems of the
community.

SOME OF THE RESPONSIBILITIES OF AN EPIDEMIOLOGIST ARE:

‣Carry out public health administration.
‣Conduct disease surveillance for the area under his charge.
‣Investigate and bring about rapid control of disease outbreaks.
‣Plan and supervise specific health programmes for disease control.
‣Train health personnel in epidemiology and disease surveillance and control.
‣Conduct health surveys, operations research and programme evaluation.
‣Produce reports useful for decision-making in disease control and prevention.
‣Describe the natural history of a disease and identify its causes.

11
ADDITIONAL READING
http://www.ph.ucla.edu/epi/snow.html
http://www.medscape.com/viewarticle/567457
Doll R, Hill AB (1950). “Smoking and carcinoma of the lung; preliminary re
port”. BMJ ii (4682): 739–48. doi:10.1136/bmj.2.4682.739. PMID 14772469
Doll R, Hill AB (1954). “The mortality of doctors in relation to their smoking
habits; a preliminary report”. BMJ i (4877): 1451–5.
doi:10.1136/bmj.328.7455.1529. PMID 13160495

12
 CHAPTER 2

Dynamics of Health
The widely acceptable definition of health is that given by the WHO in 1948 in
the preamble to its constitution, which is as follows, “Health is a state of complete
physical, mental and social wellbeing and not merely an absence of disease or in
firmity”.

However health is defined, it derives principally from forces other than

medical care. Appropriate nutrition, adequate shelter, a non-threatening environ
ment and prudent lifestyle contribute far more to health and well-being than does
the medical care system.
Health in any society should be defined in terms of prevailing ecological condi
tions. Instead of setting universal health standards, each country should decide on
its own norms for a given set of prevailing conditions and then look into ways to
achieve that level.

PHILOSOPHY OF HEALTH
‣Health is a fundamental human right.
‣Health is the essence of productive life, and not the result of ever in
creasing expenditure on medical care.

‣Health is inter-sectoral.
‣Health is an integral part of development.
‣Health is central to the concept of quality of life.
‣Health involves individual, state and international responsibility.
‣Health and its maintenance is a major social investment.
‣Health is a world-wide goal.

13
SPECTRUM OF HEALTH
There is no clear-cut demarcation between health and disease. Health can
range from optimum well-being to various levels of dysfunction. Health is dy
namic, it is not static. There are levels of health as well as levels of sickness.

THE NATURAL HISTORY OF DISEASE

It is the course of disease over time unaffected by treatment see (Figure 2.1)

Figure 2.1 Natural History of Disease and Levels of Prevention

1. STAGE OF SUSCEPTIBILITY

At this stage the disease has not developed but the groundwork has been laid
by the presence of factors that favour its occurrence. Factors whose presence are

14
associated with the increased probability of the disease developing later are called
risk factors. Risk factors are immutable or susceptible to change. Neither will all
individuals with risk factor necessarily develop the disease nor will the absence of
risk factor ensure the absence of disease. Our inability to identify all the risk fac
tors contributing to risk of disease limits our ability to predict its occurrence.

2. STAGE OF PRE-SYMPTOMATIC DISEASE

At this stage there is no manifestation of disease but, usually through the inter
action of factors, pathogenic changes have started to occur.

3. STAGE OF CLINICAL DISEASE

By this stage sufficient end-organ changes have occurred so that there are rec
ognisable signs or symptoms of disease. Depending on a specific disease, these
are classified on morphological, functional or therapeutic considerations.

4. STAGE OF DISABILITY

There are a number of conditions which give rise to a residual defect of short
or long duration, leaving the person disabled to a greater or a lesser extent.

PREVENTION
Prevention is any activity which protects the individual or population from ex
posure to the causes of disease, disability or injury, or which enhances the ability
to withstand the onslaught of specific causative agents.

15
 Table 2.1: Levels of preventions and stages of diseases
Adapted version of Leavell H.R and EG Clark, Preventive Medicine for the Doctor in His Community; 3rd
Ed. New York, Mc Graw- Hill Book Company , 1965.

There are four levels of prevention and these are linked to the stage of the dis
ease as shown in Table 2.1.

Prevention and treatment is often viewed as mutually exclusive activities. How

ever prevention is as much a part of clinical medicine as it is of public health.
When we treat illness we are preventing death, complications and a multitude of
effects on the patient’s family and the community. The spectrum of prevention
should be viewed as integral to both public health and clinical medicine. Epidemi
ology provides the tools for assessing a disease and the rational basis on which ef
fective prevention programmes can be planned and implemented.

LEVELS OF PREVENTION
1. PRIMORDIAL PREVENTION

It is the prevention of the emergence or development of risk factors in coun

tries or population groups in which they may have not yet appeared, i.e. to avoid
the emergence and establishment of the social, economic and cultural patterns of
living that are known to contribute to an elevated risk of disease, e.g. increase in
salt and cholesterol intake, resulting in rise in prevalence of hypertension and con
sequent development of a stroke. The main intervention is through individual and
mass education.

16
 2. PRIMARY PREVENTION

It is any action taken prior to the onset of disease which ensures that the dis
ease will never occur. The actions are appropriate in stages of susceptibility. It is
directed at altering susceptibility or reducing exposure of susceptible individuals.

(a) Health Promotion

It is the provision of conditions at home, school and work environment that fa

vour healthy living, e.g. good nutrition, adequate clothing, shelter, rest and recrea
tion and physical exercise. One activity is health education which includes topics
ranging from instruction on hygiene to sex education, as applied to both infectious
and non-infectious diseases. Figure 2.2 shows the health campaign themes in Ma
laysia.

Figure 2.2 Health campaign themes of the Ministry of Health Malaysia.

Source: Official website ofMOH

17
 (b) Specific Protective Measures

This form of primary prevention is targeted at specific diseases or type of in

jury. Examples include immunizations, environmental sanitation and protection
against accident and occupational hazards. Fluoridation of water supply and phar
macological treatment of diseases to prevent subsequent end-organ damage (e.g.
renal failure) or complications (e.g. stroke), are also example’s of specific protec
tion. Table 2.2 shows the immunisation schedule for Malaysia – a specific protec
tive measure.

Malaysian Ministry of Health’s New Immunisation Chart

Age in Months School Year
Immunization
0 1 2 3 5 6 12 18 Standard 1 Standard 6 Form 3
BCG Dose1 No Scar*
Hepatitis B Dose 1 Dose 2 Dose 3
DPT+Hib Dose 1 Dose 2 Dose 3
DPT/DT B DT

OPV Dose1 Dose2 Dose3 B B

Measles Sabah
#
MMR Dose1 Additional
Dose
Tetanus B

Rubella Dose 1

Table 2.2 Immunisation Schedule in Malaysia

3. SECONDARY PREVENTION

Secondary prevention are strategies that are applied in early disease, i.e. pre
clinical and clinical stages. It is the early detection and treatment of disease. Sec
ondary prevention interrupts the disease process before it becomes symptomatic,
e.g. recognition and treatment of hypertension and transient ischemic attack for
preventing a stroke.

EARLY DIAGNOSIS AND PROMPT TREATMENT

This is the main strategy of secondary prevention. Early diagnosis including

screening and individual case finding and prompt treatment of the condition will

18
result in full restoration of function when the disease has only produced reversible
body malfunction. For many infectious and non-infectious diseases the develop
ment of screening tests have made it possible to detect latent and sub clinical dis
ease in individuals considered at risk.

4. TERTIARY PREVENTION
(a) DISABILITY LIMITATION

It is the prevention of complications of a disease before irreversible changes

set in. These actions would limit disability, e.g. early mobilization or splinting for
stroke patients to prevent contractures.

Figure 2.3 Splinting to prevent contractures in a patient suffering from stroke.

Source: Effects ofSplinting on Wrist Contracture After StrokeA Randomised Controlled Trial Stroke.
2007;38:111-116

(b) REHABILITATION

It is appropriate in a stage of advanced disease or disability. It is the alleviation

of disability resulting from disease and attempts to restore effective functioning
(disease has already occurred and left residual damage; emphasis is on remaining
abilities than losses). Modern rehabilitation includes psychosocial, vocational and
medical components, e.g. a person who has suffered a stroke needs to be rehabili
tated physically, mentally and socially to take part in daily social life and be a pro

19
ductive member of the society. Figure 2.4 shows the services involved in manag
ing a patient with stroke.

PATIENT

Figure 2.4 Services involved in rehabilitating a stroke patient

HEALTH DEVELOPMENT
Health development is defined as “the process of continuous progressive im
provement of health status of a population”. Its product is, raising the level of hu
man well-being, marked not only by reduction of illness but increase of positive
physical and mental health related to satisfactory economic functioning and social
integration.

20
 MEASURES OF HEALTH
Indicators are required not only to measure the health status of a community,
but also to compare the health status of one community or country with that of an
other for assessment of health care needs, for allocation of scarce resources, and
for monitoring and evaluation of health services, activities and programmes.
Indicators help to measure the extent to which the objectives and targets of a
programme are being attained. Indicators are often only an indication of a given
situation or a reflection of that situation at a given time. If measured over time,
they can indicate direction and speed of change and serve to compare different ar
eas or groups of people at the same moment in time.
As people continue to live longer, the goal of survival is to live as long as possi
ble in good health and free of disease. However, positive health cannot be defined
in measurable terms. Thus measurements of health have been framed in terms of
illnesses (or lack of health), the consequences of ill health (morbidity, disability)
and economic, occupational and domestic factors that promote ill health and all
the antitheses.

A. MORTALITY INDICATORS
They are the most often used indicators of health. As infectious diseases are be
ing brought under control, mortality rates are losing their sensitivity as health indi
cators in developing countries, but they continue to be used as the starting point in
health status evaluation.

1. Crude Death Rate

It is defined as the number of deaths per 1000 population per year in a given
community. It indicates the rate at which people are dying. Its usefulness is lim
ited because of the influence of age-sex composition of the population. Crude
death rate can be affected by disease, natural and man-made calamites. However,
in the long term a decrease in death rate provides a good tool for assessing the
overall health improvement in a population.

21
 2. Expectation of Life
It is the average number of years remaining at a given age. Life expectancy at
birth is the average number of years that will be lived by those born alive into a
population if the current age-specific mortality rates persist. Life expectancy at
birth is highly influenced by a high infant mortality rate.
Life expectancy at the age of one excludes the influence of infant mortality.
Life expectancy at age of five excludes the influence of child mortality. It is esti-
mated for both sexes separately.

Figure 2.5 Life expectancy in Malaysia and the World - 1975 - 2025
Source: http://earthtrends.wri.org: Population. Health. and Human Well-Being -Malaysia

An increase in the expectation of life is regarded as an improvement in health

status. It is a good indicator of socio-economic development in general. As an indi
cator of long- term survival, it can be considered as a positive health indicator. It
has been adopted as a global health indicator. A minimum life expectancy at birth
of 60 years was the goal for health for all by 2000 AD. Figure 2.5 shows the trend
in life expectancy for Malaysia in comparison to the world.

3. Healthy Life Expectancy (HALE)

Healthy life expectancy or health-adjusted life expectancy is based on life ex
pectancy at birth but includes an adjustment for time spent in poor health. It is

22
most easily understood as the equivalent number of years in full health that a new
born can expect to live based on current rates of ill-health and mortality. Healthy
ageing is an important policy issue in the face of demographic challenges to the so
cietal well-being and economic prosperity. If the population can remain healthy as
they get older, they can also remain active, contributing to society and reducing
strains on health and social systems. In Malaysia, a study by the Ministry of
Health among the elderly found 81.4% suffered from at least one chronic medical
illness and 12.7% had three or more chronic illnesses.

4. Infant Mortality Rate

It is the ratio of deaths of children under one year of age in a given year to the
total number of live births in the same year, usually expressed as the rate per 1000
live births. It is one of the most accepted indicators of health status not only for in
fants, but also for the whole population, and the socioeconomic conditions under
which they live.
Infant mortality is an important component of under-5 child mortality. Not
only does this indicator reflect health conditions, but also, and critically, it is a ro
bust and sensitive measure of the social, economic and environmental conditions
in which children (and others) live. One reason for this is that the post-neonatal
contribution to infant mortality, i.e. deaths after the first 28 days of life, is almost
entirely due to exogenous socioeconomic and environmental factors.

5. Child Mortality Rate

It is defined as the number of deaths at 1 - 4 years in a given year per 1000 chil
dren in that age group at the mid-point of the year concerned. It correlates with in
adequate Maternal and Child Health services. It also relates to malnutrition, low
coverage by immunisation and adverse environmental exposure and other exoge
nous agents.

23
 6. Under-5 Mortality
The under-5 mortality rate is the probability (expressed as a rate per 1000 live
births) of a child dying before reaching its fifth birthday. As an indicator, it pro
vides similar insights into a broad range of development factors, and has the
added advantage in that it captures almost all mortality of children below age 5.
The MDG (millennium development goals) targets are to “Reduce by two
thirds the mortality rate among children under five”. Figure 6.2 shows the under-5
mortality rates for Malaysia from 1970 to 2002.

7. Under-5 Proportionate Mortality Rate

It is the proportion of deaths occurring in the under-5 age group as a proportion
of total deaths. This rate can be used to reflect both infant and child mortality
rates. High rates reflect high birth rates, high child mortality rates and a shorter
life expectancy.

8. Maternal (Puerperal) Mortality Ratio

The maternal mortality ratio (MMR) is the number of women who die from
any cause related to or aggravated by pregnancy or its management (excluding ac
cidental or incidental causes) during pregnancy and childbirth or within 42 days of
termination of pregnancy, irrespective of the duration of pregnancy, per 100,000
births. Such deaths are affected by various factors, especially general health
status, nutrition, education, and all obstetrics services and care during pregnancy
and childbirth. It accounts for the greatest proportion of deaths among women of
reproductive age in most of the developing world. The MDG is to reduce by three
quarters the maternal mortality ratio.

24
 9. Disease Specific Mortality
Mortality can be computed for specific diseases. It is defined as the number of
deaths from a stated cause in a year per average (mid-year) population. This rate
gives an idea of the burden of the particular disease on the community.

10. Proportional Mortality Rate

The simplest measure of estimating the burden of a disease in the community
is proportional mortality rate. This measure tells us the relative importance of a
specific cause of death in relation to all deaths in the population group. It is de
fined as the number of deaths from a given cause in a specified time period di
vided by the total deaths in the same period. However this rate has to be inter
preted with caution as it fluctuates with place and time, and as the proportion for
one disease becomes higher another will come down simultaneously.

B. MORBIDITY INDICATORS
They reveal the burden of ill health in the community. All of the following are
the morbidity rates used for assessing ill health in a community:

1. Incidence and Prevalence

These are the most important measures of disease frequency at a particular
place and time. Incidence and prevalence give the burden of a given disease in a
community and are used for comparison of diseases between places and at differ
ent times.

2. Notification Rates
For each country or state, certain diseases, usually those that have an impact on
public health, need to be informed or notified to health authorities for action. The
notification rates give the burden of the disease and help identify its trends.

25
 3. Hospital & Clinic Statistics
This is a good source of information provided the data is stored and analysed.
The disadvantage is that they do not relate to a defined population and vary with
the services offered, e.g. a cardiovascular centre would have a high rate of cardio
vascular diseases.
The following are examples of hospital statistics:
a) Attendance rates at out-patient departments, health centre’s, etc.

b) Admission, readmission and discharge rates.

c) Duration of stay in hospital.

d) Spells of sickness or absence from work and school.

C. DISABILITY RATES

These are used to supplement morbidity and mortality rates.

1. Event Type Indicators

Number of days of restricted activity
Bed disability days
Work loss (school loss days) within a specific period.

2. Person Type Indicators

Limitation of mobility
Limitation of activity.

26
 Figure 2.6 Global distribution of Disease in Disability Adjusted Life Years (DALYs) 2000 (000s)
Source: WHO

Figure 2.7 Major burden of disease – leading 10 selected risk factors and leading 10 diseases and injuries,
for low mortality developing countries.

Source: http://www.who.int/whr/2002/whr2002_annex14_16.pdf

27
 3. DALY (Disability - Adjusted Life Year)
The national burden of disease can be measured by DALY. The burden of dis
ease is a measurement of the gap between current health status and an ideal situa
tion where every one lives into old age, free of disease and disability. DALY is an
indicator of the time lived with a disability and the time lost due to premature mor
tality. DALY is the only quantitative indicator of burden of disease that reflects
the total amount of healthy life lost to all causes whether from premature mortal
ity or from some degree of disability during a period of time. (Figure 2.6, 2.7).

4. Years of Potential Life Lost

It is a measure of the impact of premature mortality on a population; it is calcu
lated as the sum of the differences between some predetermined end point and the
ages of death for those who died before the end point. Two most commonly used
end points are age 65 years and average life expectancy, e.g. 76.2 for females and
70.4 for males in Malaysia (2004).

D. NUTRITIONAL INDICATORS

These are often used to measure the health status of population groups (figure
2.8). As communities develop they often move from under nutrition to over nutri
tion and their associated health problems. The common ones are:
‣Anthropometric measurements of pre-school children
‣Heights of children at school entry
‣Prevalence of low birth weight
‣Percentage of obese population
‣Per capita Calorie and Protein consumption

28
 Figure 2.8 Prevalence of Underweight children in 3 countries by year.
Source: http://www.dcp2.org/pubs/PIH/6/Box/6.1 PIHFM.pdf

Table 2.3 Selected Health care delivery Indicators.

Source: Ministry of Health health_facts_2006.pdf

29
 E. HEALTH CARE DELIVERY INDICATORS

These indicators show the availability and utilization of health services (table
2.3). They are useful for planning and allocation of resources. Often alternative
systems of health care are also included in these statistics. The common ones are:
‣Doctor- population ratio
‣Doctor- nurse ratio
‣Population -bed ratio.
‣Population per health / sub-centre
‣Population per traditional birth attendant

Figure 2.9 Population per Doctor.

Source: Department of statistics, Malaysia.

30
 F. UTILISATION RATES

It is expressed as the proportion of people in need of service who actually re

ceive it in a given period, usually in a year. Health care utilisation is affected by
factors of availability and accessibility of health services and the attitude of an in
dividual towards his or her health and the health care system, e.g. proportion of in
fants who are fully immunised, proportion of pregnant women who receive antena
tal care, percentage of population using various methods of family planning, bed
occupancy rate and bed turn-over rate.

G. INDICATORS OF SOCIAL AND MENTAL HEALTH

These cannot be measured directly; hence it is necessary to use indirect meas

ures, i.e. indicators of social and mental pathology, which include suicide, homi
cide, other acts of violence and crime, road traffic accidents, juvenile delinquency,
alcohol and drug abuse, smoking, obesity, family violence, battered-baby and
battered-wife syndrome, and neglected and abandoned youth in the neighbour
hood. All these act as a guide for social action towards improving the health of the
people.

H. ENVIRONMENTAL INDICATORS
They reflect the quality of the physical and biological environment in which
diseases occur and in which people live. These are indicators relating to pollution
of air and water, radiation, solid wastes, noise and toxic substances in food. The
most useful indicators are those measuring the proportion of population having ac
cess to safe drinking water and sanitation facilities.

I. SOCIO-ECONOMIC INDICATORS

The following developmental goals do not measure health directly but are very
valuable for policy makers and planners as overall development influences health.
‣Rate of population increase

31
 ‣Per capita GNP
‣Level of unemployment
‣Dependency ratio
‣Literacy rates (especially female literacy rates)

‣Family size
‣Housing - number of people per room
‣Per capita calorie availability

J. HEALTH POLICY INDICATORS

‣The proportion of GNP spent on health-related activities

‣Proportion of GNP spent on health services
‣Proportion of total health resources devoted to primary health care

K. QUALITY OF LIFE INDICATORS

Quality of life is difficult to define and even more difficult to measure. The
physical quality of life index (PQLI) consolidates three indicators: infant mortal
ity, life expectancy at age one and literacy.

32
L. HUMAN DEVELOPMENT INDEX

(100.0)

Table 2.4 Malaysia’s human development index 2006 and underlying indicators in comparison with
selected countries.

33
 Figure 2.10 Human development Index Malaysia 2000 – 2005
Source : Ministry of Finance and Ministry of Education, Malaysia

The Human Development Index (HDI) provides a composite measure of three

dimensions of human development: living a long and healthy life (measured by
life expectancy), being educated (measured by adult literacy and enrolment at the
primary, secondary and tertiary level) and having a decent standard of living
(measured by purchasing power parity, income). It provides a broadened prism for
viewing human progress and the complex relationship between income and well
being.
The HDI for Malaysia was 0.823 in 2006, which gave the country a rank of
63rd out of 179 countries with data (Table 2.4 and Figure 2.10)

M. OTHER INDICATORS

These can be categorised as:

Social Indicators : population, family formation, families and households,
learning and educational services, earning activities, distribution of income, con
sumption and accumulation, social security and welfare services, health services
and nutrition, housing and environment, public order and safety, time use; leisure
and culture, social stratification and mobility.

34
 Basic needs indicators: calorie consumption, access to water, life expectancy,
deaths due to disease, literacy, doctors and nurses per population, rooms per per
son, GNP per capita.
Health for all indicators: heath policy indicators, social and economic indica
tors related to health, indicators for the provision of health care, health status indi
cators

SUMMARY BOX CHAPTER 2

• Health and disease are a continuum and two ends of the same spectrum.

• Health interventions can be made at four levels, primordial, primary,

secondary and tertiary.

• Disease evolves over time, and as this occurs pathologic changes may become
irreversible.

• The aim is to push back the level of detection and intervention to the precursors
and risk factors of disease thus emphasizing the role of prevention rather than
curative medicine. Health and disease are multi- factorial.

• The challenge for health professionals is to find out which factors are
associative and which are causative.

• The major burden of disease is hidden. The challenge is to detect it early and
treat it promptly. Screening methods are the major tools employed.

35
 CHAPTER 3

Population Dynamics
DEMOGRAPHY
Demography is the scientific study of human population. It is concerned with
growth, development and movement of human population and focuses its atten
tion on three readily observable human phenomena, i.e.
‣Changes in population size (growth or decline)
‣Composition of the population
‣The distribution of population in space
Knowledge of the interaction of demographic characteristics of a population
and its health status is important for health service providers.
Demography deals with five demographic processes namely:
‣Fertility
‣Mortality
‣Marriage
‣Migration

‣Social mobility
These five processes are continually at work within a population determining
size, composition and distribution.

FACTORS IN POPULATION DYNAMICS

Three factors determine the population of any defined area:
‣Birth ( fertility)

36
 ‣Deaths (mortality)
‣Migration

The balance of these three factors determines whether a population decreases,

remains stationary or increases in number.
The relation between births and deaths is referred to as natural increase. When
the net effect of migration is added to natural increase this is referred to as total in
crease and is also called growth rate. The total population of Malaysia, according
to the 2000 Census, was 23.27 million compared to 18.38 million in 1991, thus
giving an average annual population growth rate of 2.6% over the 1991 - 2000 pe
riod.

SOURCES OF VITAL AND HEALTH STATISTICS

‣Census: is conducted every 10 years. In Malaysia the fourth census was con
ducted in June 2000 by the Department of Statistics.
‣Registration: of vital events, i.e. births, deaths and marriages are done by
the National Registration Department in Malaysia.
‣Notification: of diseases (refer to chapter on communicable diseases).
‣Hospital and health facility records
‣Surveys: e.g. National Health Survey, 2nd national health and morbidity sur
vey 1996, 3rd national health and morbidity survey 2006
The quality of information obtained from these various sources vary according
to country or source, which often makes comparison difficult. Common problems
include:
✴Absence of a uniform and standard method of collection.
✴Lack of completeness with respect to data record and extent of coverage.
✴Inaccuracy in recording cause of death and age.

37
 ✴Reporting agencies, especially in rural areas, being persons who are not
aware of the relevance and importance of registration.
✴Concealment, under notification, and inaccurate diagnosis of notifiable dis
eases.

DEMOGRAPHIC CYCLE
This is the term used to describe the demographic trends of a given country,
and it passes through five stages depending on the birth and death rates. These are
greatly influenced by socio economic factors.
The stages are:
1. HIGH STAGE (HIGH STATIONARY)

This stage is characterised by a high birth rate and a high death rate which can
cel each other and the population remains stationary.

2. SECOND STAGE (EARLY EXPANDING)

The death rate begins to decline, while the birth rate remains unchanged. Many
countries in South-East Asia and Africa are in this phase. Death rates decrease rap
idly as a result of improved health conditions and sanitation.

3. THIRD STAGE (LATE EXPANDING)

The death rate declines still further and the birth rate tends to fall. The popula
tion continues to grow because births exceed deaths.

4. FOURTH STAGE ( LOW STATIONARY)

This stage is characterised by low birth and low death rate with the result that
the population becomes stationary. Zero population growth has already been re
corded in many industrialised countries.

38
 5. FIFTH STAGE (DECLINING)

The population begins to decline because birth rate is lower than the death rate.
In the 1970s, Malaysia had relatively high birth and death rates. By 1990, Ma
laysia experienced a more rapid decline in birth and death rates. The proportion of
the young population continued to decline while the older age group increased.
This trend continued into the current decade and now Malaysia is considered to be
in the third stage of demographic transition.

WORLD POPULATION TRENDS

Nearly 2000 years ago, world population was estimated to be around 250 mil
lion. In the year 1800 the world population reached one billion, the second billion
came around 1930, the third billion came around 1960, the fourth billion in 1974,
the fifth billion in 1987, and sixth billion in 1999.
The global population trends show a paradox, on the one hand in many devel
oping countries fertility is declining rapidly and there are low fertility rates in de
veloped countries, but on the other hand there is a massive increase in world popu
lation. In the “World Population Assessment and Projection The 1996 edition”, the
United Nations Population Division projects a global population of 8.04 billion
for the year 2025 and 9.37 billion for 2050. According to this medium variant, an
increase of some 2.35 billion people can be expected worldwide between 1995
and 2025; and an additional 1.3 billion between 2025 and 2050.
In 1998 about three fourths of the world’s population was living in the develop
ing countries. The most populous region in the world was the Western Pacific re
gion with about 28% of the world’s population; about 25% inhabited South-East
Asia region. India, Indonesia and Bangladesh are among the most populous coun
tries in the world and account for 88% of the region’s population.

39
 POPULATION TREND IN MALAYSIA - SIZE AND GROWTH
Malaysia’s population more than doubled between 1970 and 2000, rising from
10.4 million in 1970 to reach 23.2 million in 2000. When non-citizens, mainly con
sisting of migrant labour, is included, the figure was about 23.5 million. On an av
erage the growth rate has declined from decade to decade, but became apparent
only in the 1990s. Almost 80% of the population is located in Peninsular Malaysia
and just under 10% each in Sabah and Sarawak. In Peninsular Malaysia the states
show varying growth rates. These are greatly influenced by immigration. Sabah
has shown the fastest growth due to high fertility levels and very high levels of im
migration.

POPULATION STRUCTURE
There are two broad uses of the knowledge about the population composition
and structure. ( Figure 3.1)
It provides evidence of past events in the history of populations. Ecological
processes that human populations have gone through may leave marks on the
structure of populations.

Figure 3.1 Population Pyramid Malaysia

Source: Department of Statistics, Malaysia http:// www.statistics.gov.my

40
 It allows the assessment of the limits of organisational development in a par
ticular population.

Figure 3.2 Population, Health, and human well-being-malaysia 1975-2025

Source: http://www.earthtrends.wri.org

In the 1970s, Malaysia had a typical age pyramid with a high percentage of
young children. By 1990, due to a rapid decline in birth and death rates, the pro
portion of the young population began to decline while the older age group in
creased. This trend has continued up to the present. Figure 3.2 shows the pro
jected change in the proportion of the population from 1975 to 2025.

DEMOGRAPHIC MEASURES
DEPENDENCY RATIO

This measures the proportion of those economically productive to those who

are dependent:

41
 A high dependency ratio is a reflection of great strain on the productive mem
bers of the population to provide for non-productive members.
Population aged <15+65 years and above
x 100
Population aged between 15-64 years

RATE OF NATURAL INCREASE

This rate gives a measure of the overall gain and loss in a population through
the addition of births and the subtraction of deaths.
Number of births which occur in a given population and geographic
area in a year minus the corresponding number of deaths
x 100
Mid-year population of the given geographic area during the same year

FERTILITY RATES
CRUDE BIRTH RATE (CBR)
Total number of live births in a year
x 1000
Mid-year population of all ages

This is a very rough indicator of fertility, since everyone in the population - fe

male, male, young, old - contributes equally to the denominator of this rate even
though only females of childbearing age are actually at risk of giving birth.

GENERAL FERTILITY RATE (GFR)

Total number of live births in a year
x 1000
Females in reproductive age (15-44) in the mid-year population

The GFR although an improvement on the CBR, it is still affected by differ

ences in age composition of the female population of the reproductive age groups.

42
 AGE SPECIFIC FERTILITY RATE
Number of live births to women of a particular age group in a year
x 1000
Total number of women of that particular age group in the mid year population

DEVELOPING AND DEVELOPED COUNTRIES

A developing country has a relatively low standard of living, an undeveloped

industrial base, and a moderate to low Human Development Index (HDI) score. In
developing countries, there is low per capita income, widespread poverty and low
capital formation.

Development entails a modern infrastructure (both physical and institutional),

and a move away from low value-added sectors such as agriculture and natural re
source extraction.
The characteristics of a developing country:
‣Low income
‣Inequitable distribution of wealth
‣Low literacy rates
‣Mainly rural populations
‣Mainly young populations

Some 80% of the world’s population lives in developing countries. The popula
tion characteristics of a developing country:
‣High birth rate
‣High infant mortality rate
‣High child mortality rate
‣High mortality rate
‣Short life expectancy

43
 Developed countries, in comparison, usually have economic systems based on
continuous, self-sustaining economic growth in the tertiary and quaternary sec
tors and high standards of living.

Summary Box Chapter 3

• The three important factors that are involved in population dynamics are birth,
death and migration.
• There are five stages in the demographic cycle. Malaysia is in the third stage of
the cycle.
• World population is showing a paradox with decreasing fertility rates but a
massive increase in the world population.

44
 CHAPTER 4

Epidemiological
Concepts
The epidemiological concept of disease holds that health and disease in an individ
ual or community are outcomes of the dynamic relationship between the agent,
the host and the environment. A state of equilibrium between these factors indi
cates no disease; any disturbances of this equilibrium brought about by changes in
the inherent characteristics of the agent, the host and the environment results in
disease (See Figure 2.1 and 4.1).
The concept which is sometimes called the ecological concept of disease or the
concept of multiple causations is based on three premises:
1. Disease results from an imbalance between the disease agent and the host.
2. The nature and the extent of the imbalance depends upon the nature and char
acteristics of the agent and the host.
3. The characteristics of the agent and the host and their interactions are di
rectly related to and largely dependent on the nature of the physical, biological
and social environment.

1. AGENTS
The agent has been defined as an element, a substance or a force either animate
or inanimate, the presence or the absence of which may, following effective con
tact with the susceptible human host and under proper environmental conditions,
serve as a stimulus to initiate or perpetuate a disease process. The classifications
of agents are:

45
 ‣Biological agents - due to living agents e.g. viruses, bacteria, fungi, protozoa.
‣Nutritional factors - both excess and deficiencies such as calories, proteins,
vitamins.
‣Chemical agents - e.g. lead, solvents.
‣Physical agents - humidity, vibration, heat, light, cold, radiation etc.
‣Mechanical agents - explosives, bullets, knives, etc.
‣Social and Psychological stressors - poverty, smoking, drug abuse, work
stress.

2. THE HUMAN HOST

The factors which influence the exposure of response include:
‣Age

‣Sex
‣Family size

‣Marital status
‣Religion
‣Occupation
‣Inter current disease
‣Ethnic or racial factors
‣Habits and customs
‣Inherent immunity or non-specific immunity
‣Immunity - passive immunity, active immunity

46
 3. THE ENVIRONMENT
The environment is the sum total of all external conditions and influences that
affect the life and development of an organism. It thus influences both the agent
and the host. Figure 4.2 and 4.3 shows how different factors influence the occur
rence of HIV and how they change over time.
The environment can be categorised as:
Biological environment - infectious agents of disease, reservoirs of infection,
vectors that transmit disease, plants and animals.
Social environment - the overall economic and political organisation of a soci
ety and of the institutions by which individuals are integrated into the society at
various stages in their lives.
Physical environment - heat, light, air, water, radiation, gravity, chemical
agents, etc.

RISK FACTORS: THE “BEINGS” MODEL

Epidemiological research has focused on life-threatening diseases such as can
cer. Majority of cancers are potentially preventable and are due to “extrinsic fac
tors”. However extrinsic or environmental factors have often been misinterpreted
to mean “man-made chemicals”. Hence the BEINGS is a helpful acronym to re
member the major categories of risk factors.
‣Biological factors and behavioural factors: gender, age, weight, smoking
behaviour, etc.
‣Environmental factors: rainfall, season, housing, air-conditioning, etc.

47
 Figure 4.1 Epidemiological Triad and their interaction

Figure 4.2 Reported Cumulative HIV cases by risk Categories, Malaysia

Source of Data: Malaysia, Ministry of Health 2003

48
 Figure 4.3 Reported HIV infections by risk category in Malaysia, 1986-2002
Source: Consensus Report On HIV And Aids Epidemiology In 2004: Malaysia (consensus_report.pdf)

‣Immunological factors: immunity and immunodeficiency

‣Nutritional factors: cholesterol in heart diseases
‣Genetic factors: thalassaemia, haemophilia, etc.
‣Services: social factors and spiritual factors.

SPECTRUM OF DISEASE
This term is a graphic representation of variations in the manifestations of dis
ease. It is akin to a spectrum of light where the colours vary from one end to the
other but difficult to determine where one colour ends and the other begins.
At one end of the disease spectrum are sub-clinical infections which are not or
dinarily identified and at the other end are fatal illnesses. In the middle of the spec

49
trum lie illnesses ranging in severity from mild to severe. These different manifes
tations are simply reflections of an individual’s different states of immunity and
receptivity (Figure 4.4).

For every disease there exists a spectrum of severity with a few exceptions
such as rabies. In the infectious diseases the spectrum of disease is also referred to
as the gradient of infection.
The sequence of events in the spectrum of disease can be interrupted by early
diagnosis and treatment, or by preventive measures which, if introduced at a par
ticular point, will prevent or retard the further development of the disease.

Figure 4.4 Status and Spectrum of Disease Severity

Source:Epidemiology Concepts for Disease in Animal Groups. John Gay

ICEBERG OF DISEASE
According to this concept, disease in a community may be compared with an
iceberg (Figure 4.5). The floating tip of the iceberg represents what the physician
sees in the community. i.e. clinical cases. The vast submerged portion of the ice
berg represents the hidden mass of disease i.e. latent, in-apparent pre
symptomatic, undiagnosed cases and carriers in the community.
The water line represents the demarcation between apparent and inapparent dis
ease. In some diseases (e.g. hypertension, diabetes, anaemia, malnutrition, mental
illness) the unknown morbidity far exceeds the known morbidity.

50
 The hidden part of the iceberg thus constitutes an important undiagnosed reser
voir of infection or disease in the community and its detection and control is a
challenge to modern techniques in preventive medicine.

Figure 4.5 Iceberg of Disease

One of the major deterrents in the study of chronic diseases of unknown aetiol
ogy is the absence of methods to detect the sub-clinical state-the bottom of the ice
berg.

51
 Summary Box Chapter 4

• Disease results from an imbalance between the disease agent and the host.
• The nature and the extent of the imbalance depend upon the nature and
characteristics of the agent and the host.
• The characteristics of the agent and the host and their interactions are
directly related to and largely dependent on the nature of the physical,
biological and social environment.
• For every disease there exists a spectrum of severity.
• A disease in a community may be compared with an iceberg. The tip of
the iceberg is what the physician sees

52
 CHAPTER 5

Measuring Health -
Morbidity
MORBIDITY
Morbidity is any departure, subjective or objective, from a state of physical,
mental and social well-being, whether due to disease, injury or impairment. Dis
ability is the restriction or the lack of ability to perform an activity in the manner
or within the range considered normal for a human being.
There is a need for accurate information on illness and death because of the
high economic loss, social disturbances, and the cost of medical care associated
with illnesses and to enable comparison of illness between societies at a given
point in time or over different time periods.
In morbidity statistics, the starting point of illness is defined as when either the
subject himself begins to be conscious of symptoms or some disability, or some
one else decides that the disease is present and cannot be ignored without risk to
the patient.

SOURCES OF MORBIDITY DATA

Data on morbidity can be obtained from the following sources:
‣Routine health service records
‣Routine data collection and notification systems
‣Occupational health services
‣Disease registries

53
 ‣Surveillance records of selected diseases
‣Reports to health ministries
‣Reports from voluntary bodies
‣National morbidity surveys
‣Absenteeism and medical-leave records.
‣Data from blood banks
‣Records from hospital admissions and out-patient attendance.

USES OF MORBIDITY STATISTICS

Morbidity statistics are useful in order to:
‣Evaluate control programmes of infectious diseases
‣Plan for development of preventive services
‣Plan for provision of adequate treatment services
‣Ascertain linkage to social factors
‣Conduct research into disease aetiology and pathogenesis and efficacy of pre
ventive and therapeutic measures
‣Measure economic importance of sickness
‣Study the distribution of disease and impairments.

ERRORS AND LIMITATIONS OF MORBIDITY STATISTICS

‣Problem of definition - the dividing line between illness and health is often
indefinite. In some cases there is no evidence of illness while in others infec
tion may be accompanied by various signs and symptoms.

54
 ‣Fixing the scale of illness - Normally there appears to be considerable fluctua
tions from one individual to another and in the same individual at different
times.
‣Problems of diagnosis
✴ The accuracy of diagnosis depends upon the doctor’s skills and the
diagnostic facilities available
✴Iceberg of disease phenomenon
‣Criteria of ill health
✴The opinion of the individual affected
✴Clinical examination by the physician
✴Diagnostic tests
‣Delay in reporting

‣Under reporting - stigma, ignorance, incomplete examination.

NUMERATOR AND DENOMINATOR

Often, morbidity data are expressed as counts, i.e. the number of cases occur
ring in an area or a given period. As a population science, however, epidemiology
is concerned with either the presence of health problems in a population or the oc
currence of new health events in a population. In both cases an epidemiological
measure (or expression) has at least two components: a numerator and a denomina
tor.
The numerator can be either existing (prevalent) cases as with measures of
prevalence or new (incidence) cases as with incidence.
The denominator is the population at risk or the population in which cases ex
ist or have occurred.

55
 RATIO, PROPORTION, RATE

There are three basic classes of mathematical quantity used to measure health
status and the occurrence of health events on populations (Table 5.1).
‣Ratio - is the general term that includes a number of more spe
cific measures, such as proportion, percentage and rate. A ratio is obtained by
dividing one quantity by another without implying any specific relation
ship between the numerator and the denominator. The value of a ratio can
range from minus to plus infinity.

‣Proportion - is a type of ratio in which those who are included in the numera
tor must also be included in the denominator, i.e. the numerator is a subset of
the denominator. The magnitude of proportions is usually expressed as a per
centage.
‣Rate - is a ratio in which there is a distinct relationship between the numera
tor and the denominator. A specified time period is an essential component of
the denominator.
Rates are used to compare an observed rate with a target rate of two different
populations at the same time (the two populations should be similar and be meas
ured in exactly the same way), and the same population at two different time peri
ods (used for studying time trends).

(a) Categories of rates

(i) Crude rates: are rates that apply to entire populations, without a refer
ence to any characteristics of the individuals in it. They are valid rates
but often misleading
(ii) Specific rates: are rates that are used when a population is divided
into more homogenous subgroups based on a particular characteristic of
interest.
(iii) Standard rates: are rates that are standardised to compare between two
or more different populations.

56
 Table 5.1 Common ratios, proportions and rates in epidemiology.

Table 5.2 and Figure 5.1 shows the differences between crude and standardized
rates for mouth cancers in Malaysia.

Table 5.2 Mouth Cancer Incidence per 100,000 population (CR) and Age-standardised incidence (ASR), by
sex, Peninsular Malaysia 2003

Source: Second report of the National Cancer Registry. Cancer incidence in Malaysia, 2003. National Cancer
Registry. Kuala Lumpur 2004.

57
 Figure 5.1 Mouth Cancer – Age-specific Incidence Rate per 100,000 population – Malaysia 2003
Source: Second report of the National Cancer Registry, cancer incidence in Malaysia, 2003. National Cancer
Registry. Kuala Lumpur 2004.

(b) Advantages and disadvantages of crude, specific and adjusted rates

(Table 5.3)

Table 5.3 Advantages and disadvantages of crude, specific and adjusted rates
Source: Epidemiology -An Introductory Text. 2nd Edition. Philadelphia. Mausner Judith S. Kramer Shira

58
MEASURES OF DISEASE FREQUENCY
A) PREVALENCE

Prevalence is an estimate of the proportion of individuals in the population

with a given disease, disability or health state at a particular point in time. It is
also the measure of the existence of a particular condition, i.e. prevalence meas
ures the probability of people having a disease at a given point in time.
Prevalence is not strictly a rate although it is sometimes referred to as one. It is
a proportion and should usually be reported as one. The major difference between
incidence and prevalence is that knowledge of time of onset is not required in a
prevalence study. For example for a disease like hypertension, some may have de
veloped it a day before, some a month ago and some a year ago. For prevalence
we do not take the duration of disease into consideration; consequently the nu
merator has a mix of people with different durations of disease. Denominators in
prevalence rates always include the entire population since the numerator contains
old as well as new cases.
Prevalence depends on two factors: the number of people who have been ill in
the past (previous incidence) and the duration of their illness, P~ I x D; if inci
dence and duration have been stable over a long period of time then this formula
becomes P= I x D.
Prevalence may change over time in response to:
‣Changes in incidence
‣Changes in disease duration and chronicity
‣Intervention programmes
‣Selective attrition
‣Changing classification of what constitutes an active case and whether an ar
rested/ cured case is counted or not. (Table 5.4)

59
 Table 5.4 Influences on prevalence of disease

Figure 5.2 Numbers and prevalence of hypertension in two villages of Kota Kula Muda, Kedah, Malaysia
(2006)

Figure 5.2 shows the difference in the age trend when prevalence is calculated
as compared to absolute numbers. When absolute numbers are given it appears
that the peak age is in the 41 - 50 age group, whereas when the prevalence is calcu
lated the age group of 61 - 70 shows the highest prevalence.
There are two types of prevalence rates: point prevalence and period preva
lence.

1) POINT PREVALENCE

Number of existing cases of a disease at a point of time

Point prevalence =
Total population at that point of time

60
 This attempts to measure disease at one point in time - the numerator in point
prevalence is the number of persons with a particular disease or attribute on a par
ticular date.
Point prevalence is preferred over period prevalence since it is more precise.

2) PERIOD PREVALENCE

Number of existing cases of a disease during a period or interval

Period prevalence =
Average population during a period or interval (usually at mid-point)

This describes the prevalence of disease over a period of time - it is a com

pound measure and is constructed from prevalence at a point in time plus new
cases (incidence) and recurrences during a succeeding time period. Period preva
lence is preferred to point prevalence or incidence for analysing data on mental ill
ness. It is also used for the planning of hospital beds.

USES OF PREVALENCE

‣Important in determining workload for chronic diseases, as it is a use

ful tool in the planning of facilities and man-power needs, e.g. insulin require
ments for diabetics, facilities for training of mentally challenged.
‣May also be used to express the burden of some attribute or condition in a
population.
‣Useful for monitoring control programmes for chronic conditions.
‣May be used to estimate the importance of a disease in a population,
but with the realization that prevalence may not be a good estimator of inci
dence, in the absence of necessary data for calculating incidence.
‣Useful in tracking changes in disease patterns over time through periodic esti
mation of point prevalence.

61
 B) INCIDENCE

Incidence measures the number of new cases or new events of disease which
develop in a given population during a specified time period. Incidence rates meas
ure the probability that healthy people will develop a disease during a specified pe
riod of time.
To determine incidence, it is necessary to follow prospectively a defined
group of people and determine the rate at which new cases of disease appear. Cer
tain basic requirements must be met if incidence rates are to be calculated:
‣There must be adequate grounds on which to assess the health of individuals
in a population and to classify people as diseased or not diseased.
‣Determination of date of onset is necessary for studies of incidence.
‣Specification of numerator - number of persons affected vs. number of epi
sodes of the condition, e.g. number of children who had diarrhoea in a year
vs. number of episodes of diarrhoea in a group of children, as one child may
suffer from several episodes.
‣Specification of denominator - since incidence covers a period of
time, the number of persons “at risk” is likely to change. The simplest
way to handle this problem is to let the population at mid-point of the
time period represent the average population at risk. Population at
risk should form the denominator - the denominator should not include
those who already have the disease or those who are not susceptible be
cause they already had it or have previously been immunized. In contrast,
denominators in prevalence rates always include the whole population, i.e.
diseased and not diseased.
‣Incidence rates must always be stated in terms of a definite pe
riod of time. The time period for which a rate is calculated should be long
enough to ensure stability of the numerator.

Incidence may change with the following factors:

‣Introduction of a new risk factor

62
 ‣Changing habits
‣Changing virulence of causative organism
‣Changing potency of treatment of intervention programmes
‣Selective migration of susceptible persons to an endemic area, which in
creases the incidence of the disease.

1) CUMULATIVE INCIDENCE (CI)

Cumulative incidence is the proportion of people in a total population at risk,

free of disease at the start of a particular time period, who become diseased or de
velop the incident condition during the specified time period. Cumulative inci
dence provides an estimate of the probability (or risk) that an individual will be
come diseased in the specified time period.
Cumulative incidence assumes that the entire population at risk has been fol
lowed for the whole time period; it also assumes that all participants are at risk of
the outcome of interest.
Cumulative incidence of a disease is dependent upon the incidence and the
length of the follow up period (T). CI = I x T. If either the incidence or follow up
period increases, then the CI will also increase.

Total
Number
population
of new cases(free
at risk of afrom
disease
disease
in a given
at beginning
period of period)
of time
CI = x 1000

A measure closely related to CI is survival. Survival is also a proportion. Usu

ally survival refers to survival from death and is therefore the reciprocal of cumu
lative mortality (CI where the incident condition is death).

2) INCIDENCE RATE (INCIDENCE DENSITY)

Often every individual in the denominator is not followed for the specified pe
riod of time. For a variety of reasons including absence of follow up, death or mi

63
gration, different individuals are observed for different lengths of time. Hence for
differing periods of observation, person time denominator must be used.
Incidence Rate is a true rate and is considered to be an instantaneous rate of de
velopment of disease in a population. The numerator is the number of new cases
or incident cases in the population.

Number
Person timeofofnew
observation
cases
Incidence rate = x 1000

Figure 5.3 Incidence of dengue per 100,000 population (1990 – 2006)

Source : http://www.wpro.who.int/NR/rdonlyres/3FB0A304-554E-4637-A3A0-3443036E56BC/0/MAA.pdf

The denominator is the sum of each individual’s time period of observation,

which is the total time during which individuals are at risk of developing the dis
ease of interest. The denominator is expressed as person time of observation. Fig
ure 5.3 gives an example of the incidence of Dengue in Malaysia and its utility in
comparing time periods.

64
 The use of person time denominator is valid only when:
1. The risk of disease or death is constant throughout the entire period of study.
2. The rate of disease or death amongst those lost to follow up is the same as
among individuals still under observation.
If the disease under study is so rapidly fatal that certain individuals are ob
served for less than a full unit of time, then the rate will be artificially high.

a) Attack Rate:
It is a variant of an incidence rate. When the study period spans the entire epi
demic, a special term is used to describe infectious disease outbreaks; the inci
dence rate is then referred to as an attack rate, e.g. outbreaks of food poisoning.
The attack rate is useful for pinpointing suspect causative agents

Number of people
Total
atnumber
risk in whom
of people
a certain
at riskillness develops
Attack rate = x 100

b) Secondary Attack Rate

This is a measure of the frequency of new cases of a disease among the con
tacts of known cases. To calculate the total number of household contacts, we sub
tract the number of primary cases from the total number of people residing in
those households. It indicates the infectivity of the organism.

Number of cases among

Total
contacts
numberofofprimary
contacts
cases during the period
Secondary AR = x 100

65
 BOX 5.1 Incidence and prevalence

Number of existing cases of a disease at a point of time x 1000

Point prevalance =
Total population at that point of time

period or interval
Period prevalance = Number of existing cases of a disease during a x 1000
Average population during a period or interval (usually at mid-point)

CI= Number of new cases of a disease in a given period of time

x 1000
Total population at risk (free from disease at beginning of period)

Incidence rate = Number of new cases x 1000

Person time of observation

Attack Rate = Number of people at risk in whom a certain illness develops x 100
Total number of people at risk

Secondary AR = Number of cases among contacts of primary cases during the period x 100
Total number of contacts

Summary Box Chapter 5

• Prevalence is an estimate of the proportion of individuals in the population

with a given disease, disability or health state at a particular point in time.
• Prevalence rate is the measure of disease burden in a community.
• Incidence is a direct measure of risk.
• Incidence rates are the fundamental tools for etiological studies of both
acute and chronic disease since they are direct indicators of disease.
• Cumulative incidence is estimation of the risk of a disease occurrence
within a group of people.
• Person time is used for calculating incidence rate.
• Attack rate is useful for pinpointing suspect causative agents.
• Secondary attack rate indicates the infectivity of the organism.
• All rates relate to a defined population.

66
 CHAPTER 6

Measuring Health &

Disease - Mortality
Although mortality is far from being an ideal measure of the health of a popula
tion, thanks to vital registration systems, it is often the most easily available and
accessible indicator that can be used by health agencies in the planning, implemen
tation and evaluation of health services. Morbidity is of course a better indicator
of health since it covers the whole spectrum of disease but there are numerous
problems and errors associated with it.
In most countries, vital registration systems exist for the documentation of vital
events such as births, deaths and marriages. However, in many countries many
of the deaths are not reported at all, resulting in an artificially low death rate. In
formation obtained from death certificates are of great importance in determining
the disease pattern in the country, and the diseases which are the most common
causes of death in the country. In many countries, however, the cause of death is
not medically certified. Various rates can be calculated and time trends can be
looked at to determine the effectiveness of various programmes in decreasing mor
tality for specific diseases.

IMPORTANCE OF MORTALITY STATISTICS

Mortality statistics:

‣are an index of the severity of a problem from both clinical and public health
standpoints
‣are indices of the risk of disease

67
 ‣are a good reflection of incidence rate under two conditions, i.e. when
the case fatality rate is high and when the duration of disease is short. For dis
eases that are highly lethal, mortality can approximate incidence and can
therefore provide strong evidence for emerging and increasing clinical and
public health problems.
‣are a less expensive method of surveillance
‣provide information on changing health care needs
‣provide clues for the changes in patterns of disease occurrence.

SOURCES OF MORTALITY STATISTICS

‣Vital statistics registration, births and deaths
‣National sample and special health surveys
‣Hospital records including health clinics/centres
‣Notification of diseases
‣Other government institutions such as old folks home
‣Voluntary health organisations, revenue agencies, police.
In Malaysia, mortality statistics are collected and produced by the National
Registration Department based on death certificates issued. These are classified as
medically certified and uncertified deaths.

ERRORS AND LIMITATIONS OF MORTALITY STATISTICS

Mortality statistics have the following limitations:

A degree of under reporting of deaths will occur.

There may be inaccuracies in the diagnosis of cause of death due to:

68
 (i) Inaccuracies in diagnosis made by doctors. When a doctor guesses the cause
of death based on sign and symptoms, lack of laboratory and other facilities
to confirm diagnosis and when patients suffer from more than one disease.

Table 6.1 Possible Explanation for Differences or Trends in Mortality

(ii)Differences in classification of diseases.

(iii)Errors in death certificates, e.g. missing information, wrong information.

MORTALITY RATES
Mortality rates are used for assessing the health status of communities and for
analysing trends. These can be crude, specific or standardised rates.

69
CRUDE RATES

One of the most commonly used indicators is the crude death rate since the
only information required is the total number of deaths observed in a population.
It is the actual observed mortality rate in a population. These rates refer to the to
tal population and may obscure the fact that sub-groups of the population may
show significant differences in risk.

SPECIFIC DEATH RATES

To overcome the limitations of crude rates where differences of the subgroup

of the populations are obscured, specific death rates can be used. Rates can be cal
culated for various groups in terms of age, sex and ethnic groups and also dis
eases. These specific death rates are very good for measuring the mortality forces
prevailing in a country and for analysing the trends. These rates are specific to a
particular group in a population and can be compared with a similar group in an
other country.

STANDARDISED RATES (ADJUSTED RATES)

These are crude rates that have been modified so as to allow for valid compari
son of rates by correcting for different age distributions in different populations.
This is done by using a standardised population.

Reasons for Standardisation

‣To remove the effect of composition of populations on crude rates.
‣To arrive at simple summary measures which are more readily compared
than a series of specific rates.
‣Unavailability of specific rates for one or more of the populations under
study.
‣Unreadibility of specific rates because of small numbers in numerator and de
nominator.

70
Direct Method Of Standardisation
This method requires the selection of some population called standard popula
tion. A standard population is defined as one for which the numbers in each age
and sex group are known.
A standard population can also be created by combining two populations. We
calculate the death rate which will occur in a standard population if the mortality
risks of a particular country are applied to it. The age-specific rates of the popula
tion whose crude death rate is to be adjusted or standardised is applied to the stan
dard population to get the standardised rates.

IndirectAdjustment of Rates
In circumstances where adjustment is required but the procedure for direct stan
dardisation cannot be applied because of small number of deaths in one group or
the age-specific rates may not be known, indirect standardisation is done by apply
ing age- specific rates of the population of interest to a population of known age
structure (standard population) to yield expected number of deaths.

MEASURES OF MORTALITY
A mortality rate is a measure of the frequency of occurrence of death in a de
fined population during a specified time period. Time must be specific in any mor
tality rate.

Number of deaths among residents in an area in a calendar year x 1000

Mortality rate= Size of the population among which the deaths occurred

When mortality rates are based on vital statistics (e.g. counts of death certifi
cates), the denominator most commonly used is the size of the population at the
middle of the time period.

71
 Crude Death Rate
It measures the proportion of the population dying every year or the number of
deaths in the community per 1000 population. Two factors contribute to the mag
nitude of a crude death rate: one is the probability of dying for individuals and the
other is the age distribution of the population.

Number of deaths among residents in an area in a calendar year x 1000

Crude death rate =
Average (mid-year) population in the area in that year

It is useful for short-term planning and evaluation, bearing in mind the limita
tions such as under reporting of deaths and population estimation. However for
long term analysis of trends and for comparison with other countries, these crude
death rates are unsatisfactory indicators since mortality is greatly influenced by
the age, sex, ethnic and other characteristics of the populations.
It measures the decrease in population due to deaths. It is widely used because
it is relatively easy to compute and gives a rough idea of the risk of dying for per
sons in the population. The disadvantage is that it does not take into account that
chance of dying varies according to age, sex, race, socioeconomic class and other
factors, and cannot compare different time periods or geographical areas.

Specific Rate

This is used when a restriction is placed on a rate, e.g. Age-specific mortality

rate, cause-specific mortality rate, sex-specific mortality rate, occupation
specific mortality rate, race -specific mortality rate.

Cause-specific death rate

This rate should be interpreted with caution especially when comparing with
other countries since the denominator used is the total population. The numerator
is based on medically certified deaths, which in developing countries only consti
tute a small proportion of the total deaths; the actual rate is likely to be higher than

72
that calculated. There may be changes in the definition and category of certain dis
eases. Increase may also be due to improvements in diagnosis or better reporting
systems rather than actual increase in mortality.

Number of deaths from a stated cause in a year x 1000

Cause specific death rate =
Average (mid-year) population

Age-specific death rate

Age specific death rates are-calculated separately for each sex. Mortality is gener
ally lower among the females for all ages. Mortality is high at two extremes of
lives, i.e. infants and the elderly.

Age-specific
death rate Number of deaths
Average (mid-year)
among persons of
population given
in athe age group
specified age group
in a year
= x 1000

Case Fatality Rates (CFR)

This rate is meaningful only when cases are actually followed until death or recov
ery.

Number of deaths from a disease

Case fatality rate = Number of clinical cases of that disease x 100

A disease is considered severe if the case fatality rate is high or if a substantial pro
portion of the surviving patients are left with a sequel. It can be used to measure
any benefits of new therapy for a disease. It is used mainly in acute infectious dis
eases and in studying the outcome of surgical and other procedures. A high CFR
may be due to a highly virulent infectious agent, an absence of effective curative
treatment or delayed detection, e.g. in avian flu the case fatality rate is 50%.

73
 Proportionate Mortality Ratio (PMR)

Number of deaths from a given cause in a specified time period x 100

PMR = Total deaths in the same period

This measure tells us the relative importance of a specific cause of death in rela
tion to all deaths in a population. Proportionate mortality rate is not a rate but a ra
tio, since the denominator is derived from deaths and not from the population at
risk. This measure answers the question, ‘What proportion of death is attributable
to disease X?’ In contrast, a cause-specific death rate answers the question, ‘What
is the risk of death from disease X for members of a population? - Figure· 6.1
shows the global distribution of causes of death for the year 2000.

Figure 6.1 Global Distribution of Causs of Death, 2000 (000s). Total Deaths 55,694,000
Source:WHO

74
 INFANT MORTALITY RATE (IMR)

Number of Number
deaths inof
a year of children
live births in theless
samethan 1 year of age
year x 1000
IMR =

It is one of the most used measures for comparing health services among na
tions. It is a sensitive index of the general health status of the population. It is
greatly influenced by improvements in socio-economic factors and health care
services. In developing countries it is usually around 100 deaths per 1000 live
births, and in the developed countries around 10 per 1000. Infant mortality rate
can be divided into two components, i.e. deaths occurring during the first 28 days
of life (neonatal period) and deaths occurring from 28 days to under one year
(post-neonatal period).

Figure 6.2 Infant and Under Five Mortality Rates – Malaysia 1970-2002
Source: Malaysia, Department of Statistics, Vital Statistics, various years.

75
a) Neo-natal Mortality Rate (NMR)
The neo-natal period is defined from birth up to but not including 28 days. The
main causes of death during this period include birth injuries, congenital defects,
prematurity and certain infections such as tetanus. It reflects the effectiveness of
health care services for mothers and neo-nates.

Death in a year of children <28 days of age

NMR = x1000
Number of live births in the same year

Figure 6.3 Post Neo Natal and Neo- Natal Mortality rate, Malaysia 1970 – 2000

b) Post-Neo-natal Mortality Rate (PNMR)

The main causes of death are respiratory and gastrointestinal infections espe
cially during the second six months of life because of decline of passively ac
quired natural immunity. With better socioeconomic development, this rate de
clines rapidly. (Figure 6.3)

Deaths in a year of children >= 28 days of age upto 1 year x 1000

PNMR =
Number of live births in the same year

76
Perinatal Mortality Rate

Births (28
Stillwks
Births
or more
+ Number
of gestation*)
of lives births
+ deaths
in the
in the
same year**
first week of life
Perinatal MR = x 1000

*Previously foetuses of less than 28 weeks gestation were not viable, now
with improved neonatal facilities the foetuses of less than 20 weeks are viable,
thus some countries may use 20 weeks as the cut-off point.
**Sometimes only live births are used as the denominator especially in the
countries where still births are grossly under reported:
Note: Because of uncertainty about the period of gestation, WHO recommends using the size of foetuses
or infants: all foetuses and infants weighing 500 g or more or having a crown-heel length of 35 cm, cor
responding to a gestational age of 28 weeks.

Congenital defects and birth injuries are the common causes of death during
this period. This rate reflects the quality of care given to mothers during preg
nancy and childbirth as well as care for the neonates.

Toddler Mortality Rate

Deaths of children aged 1 to 4 years in a given year
Toddler MR = Mid year population of children aged 1 to 4 years for the same year x 1000

It is an indicator of the nutritional status of a population since toddlers are the

group most vulnerable to under nutrition. In developing countries, large propor
tions of toddlers are malnourished and as a result are susceptible to infections.
Many die as a result of concurrent respiratory infections and gastrointestinal infec
tions.

77
Maternal Mortality Rate (MMR)
Since registration is more complete for live births than for fetal deaths, it has
been customary to express this rate in terms of live births only. Because MMR is
much less common than infant mortality, MMR is usually expressed per 100,000
live births. The international classification of diseases has recommended that ma
ternal deaths be grouped into two groups i.e. direct and indirect.

Number of deaths from pregnancy related causes in a year x 100,000

MMR =
Number of live births in the same year

The direct obstetric deaths are those resulting from obstetric complications of
pregnant state from interventions, omissions, incorrect treatment or chain of
events resulting from any of the above. Whereas indirect obstetric deaths are those
resulting from previous existing disease or disease that developed during preg
nancy and which was not due to direct obstetric causes but which was aggravated
by the physiologic effects of pregnancy.
It reflects the risk of dying from causes associated with childbirth. It is also an
index of the quality of care for women during the ante-natal period, childbirth and
also during the postpartum period. Hemorrhage, pre-ecclampsia and infections
and associated medical conditions are the commonest causes for death of mothers.
Figure 6.4 shows the maternal mortality ratios of Malaysia from 1970 to 2002.

78
 Figure 6.4 Maternal Mortality Ratios 1970-2002
Source: Malaysia, Department of Statistics, Vital Statistics, 2000g, 2001c, and 2003e.

Note: Since 1998 numbers of material deaths have been adjusted to take account of cause of death mis
classifications. Hence there appears a rise in the ratios.

Table 6.2 Vital rates in Malaysia 2003 and 2004. Source: Department of statistics, Malaysia.

79
 Figure 6.5 Common causes of death, life expectancy and maternal mortality in Malaysia
Source: WHO Mortality Country Fact Sheet 2006

Figure 6.6 Key demographic Indicators, Malaysia

80
 Box 6.1 Common Mortality Rates

Deaths in Number
a year ofofchildren
live births
>= 28 dayssame
in the of age
year
upto 1 year
PNMR = x 1000

Death
Number of live
in a year of births in the
children days year
<28same of age
NMR = x 1000
from
Number ofNumber
deaths of livepregnancy
births in the
related year in a year
samecauses
MMR = x 100,000
of
Number of Number
deaths inof
a year
live births
children
in theless
samethan 1 year of age
year
IMR = x 1000

Births (28 wks or more of gestation) + deaths in the first week of life
Perinatal MR = x 1000
Still Births + Number of lives births in the same year
of
Mid yearDeaths
populationchildren aged aged
of children 1to 4 1years in a given
to 4 years for the same year
year
Toddler MR = x 1000

Summary Box Chapter 6

• Mortality rates are not ideal health measures.

• However they are easily available as most countries have vital registration
systems.
• Deaths may be under reported thus causing distortion in the rates.
• Rates are standardised for comparison.

81
 CHAPTER 7

Epidemiological
Studies
Having measured the rates of disease occurrence and the associated factors, an un
biased comparison of those with or without a disease, risk factor or intervention
needs to be made. This is achieved by a good research design. Some research ques
tions can be answered by more than one type of research design. The choice of de
sign will depend on factors such as cost, speed and availability of data. Each de
sign has advantages and disadvantages.

OBJECTIVES OF EPIDEMIOLOGICAL STUDIES

1. Determine the extent of disease problems in the community.
2. Investigate the aetiology of disease and modes of transmission.
3. Study the natural history of disease.
4. Estimate the risk of developing a disease.
5. Develop the basis for prevention programmes.
6. Evaluate new preventive and therapeutic measures and new modes of health
care delivery.
7. Provide a foundation for developing public policy and regulatory decision
relating to environment problems.

82
EPIDEMIOLOGICAL REASONING

Epidemiological reasoning attempts to:

1. Determine whether there is a statistical association between a factor or char
acteristic and the development of disease by
a. Studying the characteristics of groups
b. Studying the characteristics of individuals.
2. Derive appropriate inferences regarding a possible causal relationship
from the patterns of statistical association which have been found.

DESIGN OF EPIDEMIOLOGICAL STUDIES

The design of an epidemiological study serves the function of a measuring in
strument. The researcher designs the study in such a way that valid estimates are
obtained. A number of design options exist, each with their own purposes,
strengths and weaknesses.
The selection of the type of study is the core of a research design and is proba
bly the single most important decision the investigator has to make. The strategy
must include definition of variables, their levels and their relationships to one an
other.

Figure 7.1 Types of Epidemiological Studies

83
 Epidemiological studies are broadly classified as shown in Figure 7.1
The type of study design chosen depends on:
‣The type of problem
‣The knowledge already available about the problem
‣The resources available for the study.

OBSERVATIONAL STUDIES (NON-EXPERIMENTAL)

In non-experimental studies, the assignment of subjects to exposure groups is
not determined by the researcher, but rather by the study subjects themselves or
by other factors. The possibility of bias in non-experimental studies is of constant
concern.
There are two types of non-experimental designs:
‣Descriptive studies

‣Analytical studies

DESCRIPTIVE STUDIES (WHO? WHAT? WHERE? WHEN? )

These studies involve the systematic collection, analysis and interpretation of

data to give a clear picture of a particular situation. The wealth of data obtained in
most descriptive studies allows the generation of hypothesis, which can then be
tested by analytical experimental design. Both qualitative and quantitative tech
niques may be used.
In descriptive epidemiology we organise and summarise data according to
time, place and person. These three characteristics are sometimes called the epide
miological variables.
Compiling and analysing data by time, place and person is desirable for several
reasons:

84
 ‣The investigator becomes intimately familiar with the data and the extent of
the public health problem being investigated.
‣This provides a detailed description of the health of a population that is easily
communicated.
‣Such analysis identifies the populations that are at greatest risk of acquir
ing a particular disease.
This information provides important clues to the causes of the disease, and these
clues can be turned into testable hypotheses.

Characteristics of Persons
‣Age - overall the most important epidemiological variable relating to ex
posure, susceptibility and pathogenesis. Age-specific rates to make compari
sons between populations must be determined. The Population pyramid of
the group studied should also be considered.
‣Sex – anatomical, physiological, psychological and behavioural characteris
tics account for many sex-specific disease association.
‣Ethnicity - genetic, physiologic, behavioural, environmental and socio
economic characteristics of importance are considered as determinants of dis
ease.
‣Place of origin - genetic pool, environmental, cultural, behavioural and die
tary factors.

‣Marital status - a selective process, and associated with differences in physi

cal state, behavioural and socio-economic determinants of health.
‣Occupation - reflects the physical, mental, psychological, environmental
and socioeconomic characteristics related to disease patterns.
‣Socio-economic - determines many of the above characteristics.

85
Characteristics of Place (geographic, landscape epidemiology)
‣Biologic environment - climatic and ecologic characteristics that determine
flora and fauna, including human factors.
‣Chemical and physical environment - quality of air, water and food.
‣Social environment - cultural, behavioural patterns that determine risks, per
ceptions and responses.

Characteristics of Time
‣Endemics - diseases which are regularly and continuously present.
‣Epidemic - a significant excess over that expected on basis of past experi
ence; an unusual clustering over time.

‣Pandemic - an epidemic occurring across several geographic regions

‣Short time variation - outbreaks, point epidemics
‣Periodic variations:
✴Seasonal changes - changes in disease occurrence with seasons,
✴Cyclical variations - changes in disease occurrence every two to three
years. This occurs due to accumulation of susceptible or changes in envi
ronmental conditions (Figure 7.2)
✴Secular variations - changes over decades. A secular trend can be influ
enced by intervention strategies e.g. introduction of immunisation (Figure
7.3).

86
 Figure 7.2 Cyclical Trend of Dengue/Dengue Hemorrhagic Fever World Wide (1968-1998)

Figure 7.3 Secular trend of Malaria Cases, Malaysia, 1960-2002

Source: UNDP: MALAYSIA: Achieving the Millennium Development Goals Successes and Challenges.

87
 Knowledge of seasonal and cyclical trend of disease is useful for surveil
lance, planning, preparedness and control of outbreaks and epidemics.

TYPES OF DESCRIPTIVE STUDIES

1. Case studies - are based on reports of a series of cases of a specific condi
tion or a series of treated cases, with no specifically allocated control group.
They represent the numerator of disease occurrence and should not be used
to estimate risks (see box 7.1)
2. Community diagnosis or needs assessment - entails collection of data on ex
isting health problems. Programmes, achievements, constraints, etc. Their
purpose is to identify existing needs and to provide base-line data for the de
sign of further studies or action.
3. Epidemiological description of disease occurrence - entails the collection of
data on the occurrence and distribution of disease in population according to
specific characteristics of individuals, place and time.
4. Ecological descriptive studies - are those where the unit of observation is an
aggregate (e.g. a family, clan or school) or an ecological unit (a village, town
or district).
5. Descriptive cross-sectional studies or community (population) surveys -
entail the collection of data on a cross-section of the population, which may
comprise the whole population or a sample. They provide a prevalence rate
at a point in time (point prevalence) or over a period of time (period preva
lence). The study population at risk is the denominator of these prevalence
rates. Included in these studies are surveys of the distribution of a disease,
disability, pathological condition, immunological condition, nutritional
status, etc. This design may also be used in health systems research to de
scribe prevalence by certain characteristics e.g. the pattern of health service
utilization and compliance or in opinion polls.

88
 Box 7.1 Example of a case series
Case Series Analysis of Oral Cancer and Their Risk Factors
Khan AR, Anwar N, Manan AB & Narayan KA
MDJ, 2008;29(1):46-50

Cancer causes approximately 12% of all deaths throughout the world and is
the third leading cause of death in developing countries. In Malaysia,
Indians have the highest incidence of mouth cancer compared to other
races, and females are more affected compared to males. Objective: The
main objective of this study was to analyze the cases of oral cancer treated
in the dental department of Penang hospital, Malaysia and to determine the
risk factors associated with oral cancer. Methodology: We reviewed the
medical reports of all the patients with oral cancer treated in the dental
department of Penang General Hospital from 1994 to 2004. Results: There
were 46 cases of oral cancer treated by the dental department of Penang
General Hospital during this time period. 22 were males and 24 females.
The mean age of the patients was 61.2 years. Indians comprised the
majority of the cases (n=23; 50%) followed by Malays (n=12; 26.1%) and
Chinese (n=11; 23.9%). Of these cases, 54.3% (n=25) had used quid,
39.1% (n=18) smoked cigarettes and 32.6% (n=15) consumed alcohol.
Indians made up 76% (n=19) of all quid users (p=<0.05). 56% (n=14) of
all quid users used the combination of betel leaves, areca nut and lime
(p=<0.05). Females made up 81% (n=17) of the quid users and smokers
were solely males (p=<0.05). Chinese were the highest among the races to
smoke (n=6; 54.5%) and consume alcohol (n=6; 54.5%). The most
common presentation of the tumours was swelling, pain and bleeding
(n=16; 34.8%). Oral mucosa was the commonest site of the tumours with
67.4% (n=31) followed by tongue (n=9; 19.6%) and jaw (n=6;13%).
Histopatological examination revealed 91.3% (n=42) of the cases were
squamous cell carcinoma. Conclusion: This study though with its
limitations, has shown the risk of cancer due to tobacco and betel quid use.
There is a need to develop focused promotion programmes such as
prevention of betel chewing among Indian women and reduction of
smoking among Chinese. Further analytical studies such as case-control and
qualitative studies are needed to determine other influencing factors.
6. Longitudinal studies - use ongoing surveillance or frequent cross-sectional
studies to measure trends of disease over a period of time in a given population.
By comparing these trends in disease rates with other changes in the society, the

89
 impact of these changes on disease occurrence can be assessed, e.g. effect
of introduction of vaccines, natural and man made disasters and economic
change.

*All the above studies can be the source of hypothesis generation. Both cross- sectional and eco
logical studies can be used in hypothesis testing (i.e. are analytical).

ANALYTICAL STUDIES (HOW? WHY?)

Analytical strategies are observational means used in epidemiological investiga
tions to test specific hypotheses. The term “analytical” implies that the study is de
signed to establish the cause of a disease by looking for association between expo
sure to a risk factor and disease occurrence.

Figure 7.4 Types of Analytical Studies

The basic approach in analytical studies is to develop a specific, testable hy

pothesis and to design the study to control for extraneous variables that could po
tentially confound the observed relationship between the studied factors and the
disease. The approach varies according to the specific strategy used.

90
 EXPERIMENTAL (INTERVENTIONAL)
An experiment or an interventional trial is designed to evaluate the effect of an
intervention in which the assignment of subjects to exposed and non-exposed
groups is designed by the researcher.
The researcher manipulates objects or situations and measures the outcome of
these manipulations. Usually (but not always) two groups are compared, one in
which the intervention takes place and another group that remains untouched.
There are two categories of interventional studies:
‣Experimental studies
‣Quasi-experimental studies.

EXPERIMENTAL STUDIES

An experimental design is the only type of study design that can actually prove
causation. The classical study design has three characteristics:
‣Manipulation - the researcher does something to one group of subjects in the
study.
‣Control - the researcher introduces one or more control groups to compare
with the experimental group
‣Randomisation - the researcher takes care to randomly assign the subjects to
the control and experimental groups. (Each subject is given an equal chance
of being assigned to either group.)
The strength of experimental studies is that by randomisation the researcher
eliminates the effects of confounding variables.

91
 QUASI-EXPERIMENTAL STUDIES

In this study at least one characteristic of a true experiment is missing, either

randomisation or the use of a separate control group. In this study, however it is
necessary to always include manipulation of an independent variable that serves
as the intervention.

Summary Box Chapter 7

Epidemiological studies are used to describe a disease by agent, host and

environmental characteristics and time, place and person. These studies
help in generating hypothesis for associated and causative factors.

92
 CHAPTER 8

Common Study
Designs in
Epidemiology
CROSS-SECTIONAL SURVEY
A cross-sectional survey is a survey of a population at a single point in time.
Many methods like interview or mass screening can be used in these surveys.
They are quick and relatively easy to perform and give a fair idea of the health
status of the community. They can also estimate the risk of developing diseases.
The survey could be descriptive (hypothesis generating) or analytical (hypothesis
testing).

DESCRIPTIVE CROSS-SECTIONAL STUDIES OR COMMUNITY (POPULATION)

SURVEYS
Many cross-sectional studies do not aim at testing a hypothesis about an asso
ciation and are thus descriptive. They provide prevalence rate at a point in time
(point prevalence) or over a period of time (period prevalence). The study popula
tion at risk is the denominator for these prevalence rates. Cross-sectional surveys
provide a “snapshot” of the population at a certain point of time. Both exposure
and disease outcomes are determined simultaneously. They are also called preva
lence studies as prevalent cases are identified.
A common procedure used in family planning service and now in other serv
ices is KAP survey (survey of Knowledge, Attitude and Practice). The design of a
non-analytic survey or case series involves one group in which either out-come is

93
present or exposure is present. This group is described without a comparison
group, in terms of either outcome or exposure.
Qualitative methods, which have long been associated with disciplines such as
sociology and cultural anthropology, are increasingly being used in surveys. Quali
tative research is characterised by an approach which seeks to describe and ana
lyse culture and behaviour. It emphasises on providing a comprehensive under
standing of social settings and relies on a research strategy which is flexible and
iterative.

ANALYTICAL CROSS-SECTIONAL STUDIES

In analytical cross-sectional study, the investigator measures exposure and dis

ease simultaneously in a representative sample of the population. By taking a rep
resentative sample it is possible to generalise the results obtained to the population
as a whole.
Both exposure and disease outcomes are determined simultaneously for each
subject. In this type of approach the cases identified are prevalent cases of the dis
ease in question because we know that they existed at the time of the study but
their duration is not known. For this reason this study is called prevalence study.
In many cases, this study develops into a case control study with follow up in
terviews to collect further data on exposure and possible confounding factors of a
sample of cases and controls. This study measures the association between the ex
posure variable and existing disease.
Rare diseases, conditions with short duration of illness and those with high fa
tality are often not detected by the one-time snapshot of the cross-sectional study.
Thus they are more appropriate for measuring the relationship between fairly per
manent characteristics in individuals and the incidence of chronic diseases or sta
ble conditions. Design of a cross sectional study involves the selection of a dy
namic population, observation of all members of the study population at one point
in time and then the classification of all individuals according to disease and expo
sure.

94
 Figure 8.1 Design of an analytical cross-sectional study

The data collected is examined by the prevalence of disease in different sub

groups and the presence of variables (or absence) in disease vs. non-disease.
The data is tabulated as in Table 8.1
The rates calculated are
(i) Prevalence of disease:

in exposed compared with non-exposed = a / (a + b) vs. c /(c + d)

Table 8.1 2 X 2 contingency table for analysing epidemiological studies.

(ii) Prevalence of exposure:

in disease and non-disease = a /(a + c) vs. b /( b + d)

95
ADVANTAGES OF CROSS SECTIONAL STUDIES

1. Can be done in a short time.

2. Are less costly.
3. Are a starting point in prospective cohort and case control studies.
4. Provide a wealth of data that can be used in health systems research.
5. Can be used for evaluating health safety services.
6. May be used in examining and identifying risk factors for acute diseases
where the time between exposure and out-come is very short.
7. Useful for monitoring control programmes for chronic conditions such as
mental illness.
8. Periodic surveys are useful in tracking changes in disease patterns over time.
(important: repeated cross-sectional surveys over time do not constitute a
longitudinal study).

DISADVANTAGES

1. Provide no direct estimate of risk (show association only).

2. Are prone to bias from selective survival.
3. Not possible to establish temporality.
4. Even if an association of exposure and disease is observed the association
may be due to survival.
5. Prone to selection bias, information bias, confounding bias.
6. Not suitable for rare diseases and remission.

96
 Box 8.1 Example of cross-sectional study

Body Mass Index and Nutritional Status of Adults in Two Rural Villages in
Northern Malaysia

Narayan, K.A., and Abdul Rashid Khan

MalaysianJournal of Nutrition, 2007;13 (1). 9 -17. ISSN 1394-035X

There has been a change in the lifestyles of populations, including reduced

physical activity and consumption of foods high in calories. Overweight and
obesity are now replacing the more traditional public health concerns such as
under-nutrition and infectious diseases as some of the most significant
contributors to ill health. Determination of the body mass index (BMI)
profile and nutritional status of adults of two rural coastal villages in
Northern Malaysia was part of a community diagnosis in a community
survey. Height and weight were measured and BMI calculated. Blood
pressure was measured using a manual sphygmomanometer according to
WHO guidelines. A standardized questionnaire was used to interview the
villagers concerning their health. Out of the total population, 504 were
above 20 years of age. Data was available for 441 persons for analysis. There
were 210 (47.6%) males and 231 (52.4%) females. The prevalence of
underweight was 9.8% (n=43), overweight 25.9% (n=114) and obesity 17%
(n=75). The problem of over-nutrition was significantly higher among
females, especially housewives. (p< 0.05). Those in ages 41- 70 years were
the majority with problems of over-nutrition (p=<0.05). More than half
(52.9%; n=39) of those who were obese had hypertension (p=<0.05). Results
show that a higher number of women especially housewives were obese and
more than half of those obese subjects had hypertension. A more thorough
nutritional profile using waist, hip and body fat measurement as well as an
assessment of the dietary intake and activity regime of these villagers is
needed. Interventions need to be carried out before more serious
complications of obesity become rooted in this community.

97
CASE CONTROL STUDIES (RETROSPECTIVE STUDY)
A case control study is useful as a first step when searching for a cause of an
adverse health outcome. The hallmark of this type of study is that it compares a
case group (with disease) with a control group (not diseased) with reference to
past exposure to possible risk factors. The cases and controls are selected from a
dynamic population and then compared.

SELECTION OF CASES

Cases can be selected from a variety of sources (hospitals, physician clinics,

community registries and more), and assessed by interviews, questionnaires and
direct measurement. The criteria for eligibility are carefully specified.
Ideally, incident (new) cases should be used but the problem with using the in
cident cases is that we must wait for new cases to develop and be diagnosed
whereas a large number of prevalent (existing) cases are often available for study.
Despite this practical advantage of using prevalent cases, it is generally preferable
to use incident cases as any risk factors identified by using prevalent cases may be
related to the survival with the disease rather than to the development of the dis
ease (incidence). Even if we use incident cases we will still be excluding patients
who may have died before the diagnosis is made.

SELECTION OF CONTROLS

The controls should ideally be from the same population which gave rise to the
cases, e.g. non hospitalized persons living in the community (hospital patients dif
fer from people in the community). However, most often, hospitalised patients ad
mitted for diseases other than that for which the cases were admitted are chosen as
controls. Multiple controls for each case are may be used. Such controls may be of
same or different type.
The controls can be either matched or unmatched, and ideally, selected from
the same population as the cases.

98
MATCHING

Matching is defined as the process of selecting the controls so that they are
similar to the cases in certain characteristics such as sex, age, race, socio
economic status and occupation. Matching removes the influence of that variable
on the causation of the disease.

(a) Group matching (or frequency matching)

Consists of selecting the controls in such a manner that the proportion of the
controls with a certain characteristic is identical to the proportion of cases with the
same characteristics, e.g. percent married. Cases should be selected first, the pa
rameter assessed, and then a control group in which the same characteristics occur
in the same proportions is selected.

(b) Individual Matching

In a matched case control study each case is individually matched to a control,

so that the effect of potentially confounding variables is removed. When there is
an unusual distribution of cases compared to controls with respect to a particular
variable individual matching helps to balance the data.
It may be difficult or impossible to find a control that is similar to the case in
all the characteristics. Once the controls are matched to cases according to given
characteristics the same characteristics cannot be studied as a risk factor. We only
match on variables that we are convinced are risk factors for the disease character
istics but are not interested in investigating in the study. Matching on variables
other than these is called overmatching.

99
 Figure 8.2 Design of a case control study

MEASURE OF ASSOCIATION/ RISK IN CASE CONTROL STUDIES

Case control study yields odds ratio. It is the odds of exposure in diseased sub
jects and the odds of exposure in non-diseased subjects.

The data from a case control study is tabulated as in Table 8.1

Odds ratio is obtained from ratio of the odds of exposure among the diseased
(a/b) to that among non-diseased (c / d). The ratio [(a / b) / (c / d)] is algebraically
equal to the cross product ratio: ad / bc.

ADVANTAGES OF CASE CONTROL STUDIES

a) Suitable for rare as well as common diseases

b) Usually less expensive

c) Performed relatively quickly

d) Many different exposures may be studied

e) Fewer subjects.

100
DISADVANTAGES
a) Incomplete information

b) Bias - selection bias, information bias and confounding

c) Problem in seeking control group and matching variables

d) Yields only odds ratio

e) Temporal relationship not clear.

Box 8.2 Example of a Case Control study

Role of family support in older adults defaulting treatment for depression: a

case-control study

AK Rashid, MA Rahmah

AsianJ Gerontol Geriatr 2011; 6: 29–34

ABSTRACT
Background - Only 10% of older adults who need mental health care receive
it, and most default the treatment. We therefore evaluated the role of family
support in compliance of depression treatment among older adults. Methods
- A case-control study was conducted. 148 depressed older adults (aged ≥60
years) who had defaulted treatment were the cases. Two control groups were
used: one consisted of 148 depressed older adults who were followed up
regularly and another consisted of 148 non-depressed older adults who were
followed up for other psychiatric illness. Results - Factors associated with
defaulting
2.64), low treatment
educationfor
level (OR, 2.64),
depression being
werelow income (OR, 1.61)
unemployed (oddsand [OR],
ratiolack of

family support (OR, 12.85). Multivariate logistic regression showed lack of

family support (OR, 12.72; 95% confidence interval [CI], 7.00-23.12), being
unemployed (OR, 3.83; 95% CI, 1.74-8.40), and being illiterate (OR, 2.49;
95% CI, 1.06-5.87) as significant predictors. Implications - Family members
should be aware that family support plays an important part in patient
adherence to treatment.

101
NESTED CASE CONTROL STUDY
This is a hybrid design in which a case control study is nested into a cohort
study. A population is identified and followed over time. At the time the popula
tion is identified, base line data are obtained from interviews, blood or urine test,
etc. Population is then followed up over a period of years and for most of the dis
eases that are studied, a small percent manifest the disease, whereas most do not.
A case control study is then carried out using persons in whom the disease de
velopes (cases) and a sample of those in whom the disease did not develop (con
trols). This kind of study is more economical to conduct and the problem of recall
bias is eliminated.

ADVANTAGES

Possibility of recall bias is eliminated since data on exposure are obtained be

fore disease develops.
Exposure data are likely to represent the pre-illness state since they are ob
tained much before clinical illness is diagnosed.
Costs are reduced compared to those of a prospective study since lab tests and
other investigations need to be done only on subjects who are later chosen as
cases or controls.

Figure 8.3 Nested case control study

102
COHORT STUDY (LONGITUDINAL STUDY)
A cohort is a group of persons who share a common experience within a de
fined time period. In a cohort study the investigator defines a cohort of a naturally
occurring non- diseased, exposed individuals and another cohort of non-diseased,
non-exposed individuals (the comparison population) and follows them over time
to determine disease incidence.
A definitive characteristic of a cohort (both retrospective and prospective)
study is that the subjects at the beginning of the study are free of the disease out
come.

TYPES OF COHORT STUDIES

There are two general types of cohort studies.

a) Prospective Cohort Studies (concurrent cohort/prospective study)
A cohort of individuals free from the disease are selected and grouped as per
their exposure or non-exposure to a suspect causative factor and then monitored
over a period of time for the development of disease.
This type of study is warranted when:
(i) There is good evidence of an association of the disease with certain exposure
(ii) Exposure is rare but incidence is high

(iii)Time between exposure and disease is short

(iv) Attrition can be minimized
(v) Adequate funds are available
(vi) Investigator has a long life expectancy.

103
 Figure 8.4 Design of a prospective cohort study

b) Retrospective Cohort Studies (historical cohort / non-concurrent prospective

study)
In this type of study the investigator goes back in time to define the exposure
and risk group and follows the members to the present time to see the outcomes. A
good health recording system is beneficial for this type of study as an estimate of
exposure to relevant variables in the past must be made based upon these records.
This kind of study can usually be conducted quickly and cheaply. It is efficient
for a cohort whose investigation for a disease with a long latency period requires
many years of follow up to accrue sufficient end points. Availability of relevant ex
posure data in adequate detail from pre-existing records is important. Data maybe
incomplete especially on confounding factors.

104
SELECTION OF EXPOSED POPULATION

Cohort study is often conducted among groups specifically chosen not only for
their exposure status, but also for their ability to facilitate the collection of rele
vant information. Choice of a particular group to serve as the study population for
any given study is related to both the hypothesis under investigation and specific
features of the design.

SELECTION OF COMPARISON GROUP

Major principle underlying this decision is that the groups being compared
should be as similar as possible with respect to all factors that may be related to
the disease except the determinant under investigation. It is important to ensure
that the information that can be obtained from the non-exposed group is adequate
for comparison with the exposed population. An internal comparison group can be
utilized, i.e. the experience of those cohort members classified as having a particu
lar exposure is compared with that of members of the same cohort who are either
non-exposed or exposed to a different degree.
The disadvantage of using the general population as a comparison group is that
its members may not be directly comparable to those of the study cohort. It may
be useful to have multiple comparison groups especially when no single group ap
pears sufficiently similar to those who are exposed, to provide assurance about the
validity of the comparison

105
 Box 8.3 Example of a cohort study

Kato I et al. Prospective study of gastric and duodenal ulcer and its relation
to smoking, alcohol and diet.

Source; American Journal of Epidemiology 1992; 135(5),521-530

To determine the effects of environmental exposure on the development of

gastric and duodenal ulcers a study was conducted from 1968 to 1990. The
subjects of this study were Japanese men born from 1900 to 1919 and
resident on the Hawaiian island of Oahu. They were identified by the
Honolulu heart programme in 1965 with the use of the comprehensive 1942
US selective service draft registration files. Interviews, dietary assessments
and clinical examinations were made. The incidence after examinations of
gastric and duodenal ulcers amongst men were determined by continuous
surveillance of all general hospitals on Oahu. Each diagnosed case was
confirmed by histological, radiological or endoscopic examinations. After
149,291 person years of observation, there were 280 incident cases of gastric
ulcer and 149 incident cases of duodenal ulcer.

SOURCES OF DATA

Information concerning the exposure may be obtained from a number of

sources, including records collected independently of the study, such as medical
employment records, information supplied by the study subjects themselves
through answering questionnaires, data obtained from medical examinations or
other testing of the participants, or direct measurement of the environment in
which cohort members have lived or worked. Outcome information can be ob
tained from records such as, death certificates and medical records or from ques
tionnaires and by physical examinations.
The goal is to obtain complete, comparable and unbiased information of the
subsequent health experience of every study subject. Combination of various
sources of the outcome data may be necessary to obtain complete follow- up infor
mation.

106
FOLLOW-UP

Collecting follow-up data on every person enrolled represents the major chal
lenge of a cohort study as well as the major cost in terms of time, fiscal resources
and ingenuity.

ANALYSIS

The basic analysis is the calculation of the rate of incidence of a specified out
come among the cohorts under investigation. Both relative and absolute measures
of association can be calculated.

RELATIVE RISK

Relative risk is the estimate of the association between exposure and disease
and indicates the likelihood of developing disease among the exposed individuals
relative to those not exposed, i.e. how much more likely one group is to develop a
disease than the other.

ATTRIBUTABLE RISK

Attributable risk is a method of attributing the occurrence of a disease to a spe

cific exposure which may be contributing to the development of the disease. Its
utility is that it represents the expected reduction in disease if the exposure could
be removed or never existed. It is the difference between the frequency measures
for the two populations.

ATTRIBUTABLE RISK PERCENT

This is the amount of the disease attributable to a given cause expressed as a

proportion. This is an estimate of the proportion of the disease in that group that
could be prevented by eliminating the exposure.

107
POPULATION ATTRIBUTABLE RISK

This estimates the excess rate of disease in the general population. It deter
mines which exposures have the most relevance to the health of a community.
Population Attributable Risk Percentage
This tells what proportion of disease in a population is attributable to a risk fac
tor.

ADVANTAGES OF COHORT STUDY

1. Suitable for rare as well as common exposure

2. Exposure data are often more accurate
3. Less information bias
4. Examines multiple effects of a single exposure
5. Provides absolute and relative effect measures
6. Can elucidate temporal relationship between exposure and disease
7. Allows direct measurement of incidence of disease in the exposed and non
-exposed groups.

DISADVANTAGES OF COHORT STUDY

1. Inefficient for the evaluation of rare diseases unless attributable risk present
is high
2. Expensive and time consuming (prospective)
3. Requires availability of adequate records (retrospective)
4. Validity can be seriously affected by losses to follow-up (attrition)
5. Large number of subjects required
6. Change over time - in criteria and methods
7. Non-response bias.

108
 Summary Box Chapter 8

• A cross sectional survey is a survey of a population at a single

point in time.
• Case control studies start with subjects who have a disease.
• Case control studies compare a group of cases with a matched
group of controls for rates of suspected exposure factors.
• Cohort studies are done in groups of people who are free of the
disease at the beginning.
•Cohort studies see the effect of a suspected exposure factor
on the population initially free of disease.

109
 CHAPTER 9

Risk

Mr I. Q. Low came for persistent cough to Dr. K Razy. The doctor found that
Mr Low was a smoker and advised him to stop, stating that research had shown
that smokers had increased risk of getting cancer. Low wanted the doctor to
clarify whether it meant he would definitely get cancer as it was difficult for
him to stop smoking given the nature of his job. Dr. K. Razy, though willing to
explain, was unable to explain the concept of risk vs getting the disease.

Often doctors and the lay public do not fully understand the implications of
studies reporting risk of exposure to disease agent. The confusion arises because
of the types of risk measures, how they are calculated and their implications.
Risk is the probability that an individual will incur a specified event within a
specified time period under specific conditions. It applies to individuals rather
than whole population whereas incidence and prevalence are group measures (i.e.
characteristics of populations not single individuals). Risks to individuals are esti
mated by measuring the corresponding cumulative incidence measure in the
whole population.
Risks vary according to the profile of causal risk factors to which an individ
ual is exposed. In practice, specific health risks faced by one individual cannot be
directly measured but may be estimated empirically by observing the health out
comes arising in a sufficiently large population of people who have characteristics
in common with the person for whom the risk assessment is to be made.
There are three types of risk: absolute, relative and attributable.
‣Absolute risk is the incidence of disease in a defined population
110
 ‣Relative risk is the ratio of the incidence rate in the exposed group to the inci
dence rate in the non-exposed group.
‣Attributable risk is the difference in the incidence rate between the exposed
and non- exposed groups.

RISK FACTOR
It is a characteristic which if present and active, clearly increases the probabil
ity of a particular disease in a group of persons compared with an otherwise simi
lar group of persons who do not have the characteristics. There are many types of
risk factors:
Physical e.g. toxins, infectious agents; social environment e.g. culture, loss of
spouse; behavioural e.g. smoking, drinking alcohol, chemical and genetic.

USE OF ESTIMATING RISK

(a) Prediction: The future of the disease in the population studied
(b) Diagnosis: Improvement in screening test
(c) Prevention: Removal of the risk factor can prevent the disease occurrence
(d) Counseling of patients: e.g. Smoking and lung cancer (OR, RR)
(e) Policy analysis: PAR% in preventive programmes

MEASURES OF ASSOCIATION (MEASURE OF RISK/EFFECT)

Analytical studies are designed to determine whether an association exists be
tween a factor (or exposure) and a disease, and if so, to determine the strength of
the association. The most efficient way to do this is to combine the measures of
disease frequency of the two populations being compared into a single expression.
The calculation of these measures of association or risk is facilitated by the use of
2x2 or a fourfold contingency table.

111
RELATIVE RISK (RISK RATIO)

This is the estimate of the association between exposure and disease and indi
cates the likelihood of the disease developing among the exposed individuals rela
tive to those not exposed, i.e. how much one group is more likely to develop a dis
ease than another. It is defined as the ratio of the incidence rate for persons ex
posed to a factor to the incidence rate for those not exposed.

Incidence rate among exposed

Relative Risk (RR) =
Incidence rate among non-exposed

Using 2 x 2 table (Table 8.1) it is the ratio of cumulative incidence of disease

among exposed to the cumulative incidence of disease among those unexposed.
Prospective studies permit direct calculation of the incidence rate of the dis
ease for the populations exposed and not exposed. This is because both groups rep
resent defined populations at risk that are followed for the development of dis
ease; therefore, excess risk caused by exposure to a given factor can be calculated
directly.

a/a+b
Relative Risk (RR) =
c/c+d

A risk ratio of 1 indicates identical risk in two groups. A risk ratio greater than
1 indicates an increased risk for the numerator group (exposed) while a risk ratio
less than 1 indicates a decreased risk for the numerator group (exposed). This
could also be called a protective effect and would be seen in those who are given
an effective vaccine (“exposed” to the vaccine).

Table 9.1 Hypothetical data of a cohort study amongst smokers and non smokers

112
 To respond to Mr I. Q. Low and understand the risk of developing lung cancer
due to smoking Dr. K. Razy found an article which described a cohort study com
paring the incidence of cancer among a group of men who smoked compared with
another group who did not smoke. If10 out of 1000 men who smoked over a 10 yr
period developed lung cancer compared with 20 of 4000 men who did not smoke
who developed lung cancer, the relative risk of developing cancer would be calcu
lated as follows.
Relative Risk = 1%/0.5% = 2
What is the meaning of this figure? How does Dr. K. Razy. Explain to Mr I.Q
Low about the risk of smoking and developing lung cancer.
“Given all other factors are equal, there is a two times greater chance (probabil
ity) of developing lung cancer if you smoke”. In other words the incidence of
lung cancer amongst smokers(1%) would be twice of that amongst non
smokers(0.5%).

ODDS RATIO (ESTIMATED RELATIVE RISK)

In case control studies, incidence rates cannot be derived directly because there
are no appropriate denominators (population at risk; i.e. groups (a + b) and (c +
d) do not represent the total populations exposed and not exposed to the factor).
Relative risk can be estimated from a retrospective study if the assumptions can
be made that (l) the controls are representative of the general population, (2) the
assembled cases are representative of all cases and (3) frequency of the disease in
the population is small.
However, as this is not usually the case, Odds ratio is calculated. Odds ratio is
a way of comparing whether the odds of a certain event is the same for the two
groups.
The odds ratio is obtained from ratio of the odds of exposure amongst the dis
eased (a / b) to that amongst non-diseased (c / d). This ratio [(a/b) / (c/d)] is alge
braically equal to the cross-product ratio: ad / bc.

113
 When OR is 1 there is no association with the risk factor, when OR is greater
than 1 there is an increased risk with the exposure, and OR less than 1 there is de
creased risk with the exposure, i.e. protection.

(cases with exposure) x (control without exposure)

OR =
(control with exposure) x (cases without exposure)

Table 9.2 Hypothetical data of case control study of lung cancer

Dr K Razy found another article where men with and without lung cancer were
asked if they had ever smoked. (Case Control design). It was found that 100 men
of the total of 1000 with lung cancer (cases) had smoked and 200 of the 4000 with
out lung cancer (Controls) had smoked. The odds of developing lung cancer in
those who smoke is as follows.
The odds ratio is calculated thus ad/bc or 100x3800 / 900x200 = 2
How does Dr. K. Razy explain this? He says there is a two fold greater odds of
getting lung cancer when a person smoked.

What is the difference between Odds and probability?

Odds is not a slang for probability. A probability is a number from 0 to 1 inclu
sive, usually expressed as a fraction, which is the ratio of the number of chances
of a specific event to the total number of chances possible.
For example, if one has 4 candies in a jar, 3 red and 1 yellow, then the probabil
ity of drawing the yellow is 1/4. There is one chance of a yellow candy of 4 total
chances.

114
 Odds are expressed as the number of chances for versus the number of chances
against. So, since there is 1 chance of yellow being picked , and 3 chances of pick
ing red, the odds are 3 to 1 AGAINST picking the yellow.
Note that this does NOT mean that the probability is 1/3 for or against in the
above example.
To convert odds to probability, we have to ADD the chances. So, if the odds
for a yellow candy to be picked is 1 to 3 the probability is 1 in 4 or 25%. The prob
ability ofa red is 3 by 4 or 75%.
Incidentally, odds of 1:1 would be read as “one TO one”, not “one OUT OF
one.” (The words “out of” seem to imply total chances, which is probability, not
odds.)

ATTRIBUTABLE RISK
Having understood that smokers were at a higher risk of developing cancer
than non smokers, Dr. K Razy now also understands that the relative risk and
odds ratio demonstrate the association of smoking with cancer and give a clue to
causation. Dr. K. Razy wonders how much of cancer could be avoided if people
stopped smoking. Can the difference in rates of cancer between smokes and non
smokers be attributed to smoking? Dr. K. Razy finds out this is true and is called
the attributable risk which is defined as the amount or proportion of the disease
incidence (or risk) that can be attributed to the specific exposure. The utility is
that it provides a measure of the amount of disease which potentially could be
avoided if the exposure was eliminated.
Both the odds ratio and the relative risk are computed by division and are rela
tive measures, in contrast absolute measures are computed as a difference rather
than a ratio.
We can calculate the attributable risk for the exposed persons or for the whole
population.

115
ATTRIBUTABLE RISK FOR THE EXPOSED GROUP

It is the difference between the frequency measures for exposed and unexposed
groups.
AR=a/(a+b)–c/(c+d)
When the amount of the disease attributable to a given cause or exposure is ex
pressed as a proportion (percentage) of the disease in the group, the result is called
the attributable risk percent. Its utility is to estimate the proportion of the disease
among the exposed that is attributable to the exposure or the proportion of the dis
ease in that group that could be prevented by eliminating the exposure.

Risk (exposed)
Risk (exposed)
- Risk (unexposed)
AR% = x 100

With reference to table 9.1 the attributable risk is = (10/1000 – 20/4000) / 10/
1000 0r 0.5 or 50% Meaning that 50% of the lung cancer cases among the ex
posed group is attributable to smoking. Dr. K. Razy understands that half the
cases of lung cancer would not occur if all people quit smoking.
AR = 1- 0.5 = 0.5

RELATIONSHIP OF RELATIVE AND ATTRIBUTABLE RISK

The relative risk is related to the attributable risk by the formula
AR= rate of disease in unexposed persons x (RR-1)
This is useful for estimating the attributable risk from case control studies as
OR is approximately equal to RR. As AR is more useful in risk management, cal
culations can be made for conditions with similar risk but different prevalence.

116
ATTRIBUTABLE RISK FOR THE TOTAL POPULATION
Having understood that stoppage of smoking can reduce the incidence of lung
cancer, Dr.K. Razy proposes to launch a health campaign to ask people to stop
smoking. The Mayor Ms. C. Nickle, is willing to spend money but asks Dr.K. Razy
to give a realistic estimate of the extent of reduction that could be achieved by the
programme. Having read about attributable risk Dr. K. Razy says that it would be
reduced by half. The mayor does not believe this as, she says, if women do not
smoke there will be no reduction among them. Dr. K. Razy is stumped. He realizes
that his programmes success would depend on the number of people smoking in
the community. If no one smoked no cancer attributable to smoking would occur.
The answer to Dr. K. Razy’s dilemma is the population attributable risk (PAR)
which estimates the excess rate of the disease in the total study population of ex
posed and non- exposed individuals that is attributable to the exposure, i.e. to de
termine which exposures have the most relevance to the health of a community.
PAR is calculated as the rate of disease in the population minus the rate in unex
posed group.
This measure can be calculated by multiplying the attributable risk by the pro
portion of exposed individuals in the population.
PAR = (Attributable Risk) x (Proportion Exposed in Population)
When the population attributable risk is expressed as a percentage of the dis
ease in the population, it is called population attributable risk percentage. This
tells us what fraction of disease in a population is attributable to exposure to a risk
factor. It is calculated by dividing the population attributable risk by the rate of the
disease in the population. This information helps set community health priorities
by contrasting the harmful effects of different exposures.

Risk (total) - Risk (unexposed)

PAR% = x100
Risk (total)

Dr. K. Razy tries to find out the total incidence of cancer in the population
with little success. However he finds out that about 40% of the population smoke.

117
Hence by calculating the incidence in smokers by the population who smoke
(40%) and the incidence among non smokers by the population who do not smoke
(60%) he can arrive at an estimate of the true incidence in the population.
So in this example the
Risk (Total) = 1%x 0.40+0.5%x 0.60 = 0.70%
And the PAR in the above example is
(0.70 – 0.50) / 0.70 x100 = 28.5%
Meaning that in a population wherein 40% of the people smoke, 28.5% of the
lung cancers can be attributed to smoking and this number would have been re
duced had there been no smoking.

EXPLANATION OF THE THREE TERMS

The relative risk refers to the likelihood of the disease in the exposed patients
relative to those who are not exposed. The relative risk is independent of and does
not require knowledge of the overall incidence of disease in the population as it
only compares the incidence among exposed and unexposed sample groups. Attrib
utable risk provides information about the excess risk of disease in the exposed
group as compared with the unexposed group. The population attributable risk
measures the potential impact of control measures of the exposure in a population,
and is relevant to decisions in public health.

118
 Summary Box Chapter 9

•Factors whose presence are associated with an increased possibility that

the disease will develop later, are called risk

•Risk is a probability. It can never be observed.

•Odds ratio is an estimate of the incidence ratio amongst the exposed and
non exposed persons in the dynamic population from which the controls
and cases are selected.

•The relative risk refers to the likelihood of the disease in the exposed
patients relative to those who are not exposed. The relative risk is
independent of and does not require knowledge of the overall incidence of
disease in the population as it only compares the incidence among exposed
and unexposed sample groups.

•Attributable risk provides information about the excess risk of disease in

the exposed group as compared with the unexposed group. The population
attributable risk measures the potential impact of control measures of the
exposure in a population, and is relevant to decisions in public health.

119
 CHAPTER 10

Bias & Error

The primary purpose of research studies is to extrapolate the results to the general
population. Unless the conclusions about the sample population are accurate and
in some ways can be applied to the more general population, the study and the re
sults are useless. The conclusions can be distorted by two factors— bias and error.

ERROR
Error occurs when an incorrect assumption or conclusion is made about data.
This can be either systematic or random as shown in Figure 10.1

Figure 10.1 Types of error

120
 RANDOM ERROR

Random error occurs because of chance and usually cannot be predicted. The
random errors that can be predicted or determined are due to sampling error, i.e.
sampling of the population is not representative and not probabilistic. The size
and design of a sample as well as the distribution of the variable of interest deter
mines the sampling error. As random error increases, the precision, i.e. the repro
ducibility or reliability, of the study decreases. Random error can be decreased by
increasing sample size or by making the study design more efficient so that the
sample is representative of the population.

SYSTEMATIC ERROR OR BIAS

A systematic error occurs in epidemiology when there is a tendency to produce

results that are different in a systematic manner from the true value e.g. a consis
tent error due to the malfunction of a measurement device. Systematic errors are
also called bias. The source of a systematic error can usually be determined. As
systematic error increases, the validity, i.e. the true intent, of the study decreases.
Reduction of bias and confounding can only be achieved by good study design.
Once bias is introduced into the study it is difficult to correct the effects analyti
cally.
Bias falls into three main categories selection, information and confounding.
No study can completely eliminate bias.

SELECTION BIAS

Selection biases are those that occur at the start of a study usually because the
method of selection of the study population is biased. Just as the selection of the
study population for a certain disease in a case control study is somehow related
to the exposure history, likewise in cohort studies the selection of an exposure is
somehow related to the disease.

121
Some types of Selection Bias
1. Berkson’s bias: bias that occurs when using a study population from the hos
pital. Case control and cross sectional studies carried out exclusively in hos
pital settings may produce a spurious association because the types of ill
ness may vary from the population and because of the services offered.
2. Self selection bias: occurs when individuals refer themselves for treatment
or volunteer for a study or refuse to participate because of an interest that is
connected to the outcome of the research. It includes both volunteers and
non respondent.
3. Healthy worker effect: caused by the fact that healthy individuals are the
ones that are working, the morbidity and the mortality rates are naturally
lower than in the general population.
4. Lead time bias: bias that occurs when comparing survival rates between a
group that is screened for a disease and a group that is not. The fact that the
disease is detected earlier because of screening makes the survival look
longer. This is due to the fact that the disease is diagnosed earlier.
5. Prevalence / incidence bias: mild or asymptomatic cases as well as fatal
and short disease episodes can be missed if a study is performed later in the
disease process.
6. Detection bias: if sign or symptoms of a specific disease that initiates a
search for that disease is caused by an innocuous exposure then the expo
sure itself may be falsely considered the cause of the disease.
7. Membership bias: persons who belong to a group (workers, athletes) may
experience a level of health that is systematically different than that of the
general population because the general population is composed of both
healthy and ill individuals.

122
INFORMATION BIAS
This bias occurs because the data that are collected or observed are incomplete
or incorrect.
1. Interviewer or observer bias: occurs when interviewers solicit data that is
biased in some way or have preconceived expectations of what they should
find in an examination.
2. Loss to follow-up bias: occurs because those individuals that drop out of a
study are different from those that stay in the study.
3. Misclassification: occurs by putting cases and control in the wrong cate
gory.
4. Recall bias: is the inability to recall events that happened in the past.
5. Inter- interviewer bias: is the systematic error that occurs when more than
one interviewer solicits records or interprets information from the study
subjects.
6. Questionnaire bias: is when leading questions or other flaws in the ques
tionnaire result in a differential quality (accuracy) of information between
compared groups.
7. Diagnostic suspicion bias: occurs when both the intensity and the outcome
of the diagnostic process are affected by the investigator’s knowledge of
the subject’s prior exposure to the putative cause.
8. Exposure suspicion bias: occurs when the intensity and outcome of the
search is affected by the investigators knowledge of the subject’s disease
status.
9. Hawthorne effect – an effect first documented at a Hawthorne manufactur
ing plant; people act differently if they know they are being watched.

10.Surveillance bias – the group with the known exposure or outcome may be
followed more closely or longer than the comparison group.

123
PREVENTION OF BIAS
The degree to which an investigator is aware of the possible sources of selec
tion bias in a proposed study determines the degree to which selection bias can be
avoided through proper study design. Information biases are easier to prevent than
selection biases. Case control design is affected by many sources of bias and has
less chances of defense as compared with cohort studies( Ref Tables 10.1 and
10.2).

CONFOUNDING
It is the distortion or the masking of an association between the exposure and
an outcome because of a third extraneous factor. When the effects of two expo
sures have not been separated and an incorrect conclusion that the risk is due to
only one risk factor is drawn, confounding is said to have occurred. For example,
if the effect of smoking on lung cancer is to be studied and no occupational his
tory is taken, then exposure to asbestos in miners may be missed (Fig 10.2).

EFFECTS OF CONFOUNDING
A confounding variable can increase, decrease or even change the direction of
the estimated association between an exposure and outcome; thus the conclusion
from the data is biased.

124
 Table 10.1 Prevention of Selection Bias in basic study designs
Source: Choi,. BCK. & Noseworthy, A.L. Classification, direction, and prevention of bias in epidemiological
research. Journal of Occupational Medicine, 1992 34(3) 265-271

125
 Table 10.2 Prevention of information bias in basic study designs.
Source: Choi, BCK. & Noseworthy, A.L. Classification. direction, and prevention of bias’in epidemiological
research. Journal of Occupational Medicine, 1992 34(3).26

Figure 10.2 Confounding smoking, asbestos mining and lung cancer

126
CONFOUNDER

It is a variable that is associated with exposure and at the same time is an inde
pendent risk factor for the disease. It is associated with but not a consequence of
the exposure. Even in unexposed individuals the confounder is related to the risk
of disease. They are not variables that are intermediate steps in a causal pathway
but rather direct risk factors of disease.
Criteria for a variable to be confounder are:
a) Must be a risk factor for the disease

b) Must be associated with the exposure

c) Must not be an intermediate step in causal pathway.

HOW TO DETERMINE CONFOUNDING IN A STUDY

Other than carefully analysing the conclusions and the biologic plausibility of
the association, one needs to look at the statistics such as the t-test or Chi-squared
test that are calculated in the analysis of the difference between the two groups be
ing studied. Another method to determine confounding is by stratifying the vari
ables and see the relative risk changes.

CONTROL OF CONFOUNDING

Table 10.3 Prevention of confounding bias in epidemiological studies

Source: Choi, BCK, & Noseworthy, AL. Classification, direction. and prevention of bias in epidemiological re
search. Journal of Occupational Medicine, 1992 34(3).265-271

127
Confounding can be controlled (Table 10.3) during the design of study by:
a) Randomisation

b) Restriction
c) Matching.

It can also be controlled during the analysis stage by

a) Stratification -by stratifying the groups by the confounder the effect can be
reduced or removed. The exposure can be evaluated within those strata.
b) Statistical analysis or regression - statistical analysis can be done at the end
of the study. A large number or sample is required for regression analysis.

Summary Box Chapter 10

• Bias and error can distort the findings of a study.

• Error occurs when an incorrect assumption is made about the data.
• Random errors are difficult to predict or eliminate.
• Systematic errors usually occur due to errors in measurement.
• There are several types of bias selection or information.
• Confounding is the masking of an association between exposure and
outcome due to a third extraneous factor.

128
 CHAPTER 11

Interventional Studies

In contrast to observational studies, where the epidemiologist takes no action but

only observes the natural cause of events or outcome, experimental studies in
volve some action, intervention or manipulation; such as deliberate application or
withdrawal of the suspected causes or changing one variable of the causative
chain in the experimental group while making no change in the control group and
observing and comparing the outcome of the experiment in both the groups.
This type of study can take one of three forms:
‣Randomised controlled trial
‣Field trial
‣Community trial.

AIMS OF EXPERIMENTAL STUDIES

‣To provide scientific proof of aetiological (or risk) factors which may permit
the modification or control of those diseases
‣To provide a method of measuring the effectiveness and efficiency of health
services for the prevention, control and treatment of disease and improve the
health of community
‣To evaluate the efficacy of medical management of disease in test and control
groups. These are commonly called clinical trials.

129
FIELD TRIALS
These are experiments on people without the disease in question to determine
the efficacy of a preventive or therapeutic agent or procedure. These trials in
volve people who are disease free but are presumed to be at risk. Data collection
takes place in the field usually among non-institutionalised people in the general
population. Phase III clinical trials which are usually carried out in hospitals and
clinics are sometimes done in the field. Trials for new vaccines and methods to
evaluate interventions aimed at reducing exposure are usually done as field trials
(Box 11.1).

Box 11.1 An Example of a field trial

A field trial on the comparative effectiveness of Malathion and Resigen by

ULV application on Aedes Aegypti.

Vythilingam I, Panart P.

Southeast Asian J Trop Med Public Health. 1991 Mar;22(1):102-7.

Field trials were conducted in two residential areas of Petaling Jaya

Municipality to test the adulticidal and larvicidal effects of Malathion
96% TG and Resigen on Aedes Aegypti. Malathion is the currently used
insecticide in Malaysia for the control of dengue. The Leco HD ULV
machine was used throught the trials. For Malathion the flow rate was 90
ml/minute at a vehicle speed of 8kph and for Resigen the flow rate was 200
ml/minute at the same vechicle speed. Malathion was more effective giving
higher mortality rates when compared with Resigen. The mortality rate of
adult Ae. Aegypti outdoor was higher than in the living room and kitchen.
Both insecticides did not show promising larvicidal effects

COMMUNTY INTERVENTION TRIAL

In this type of field trial the intervention is done on a community-wide basis
rather than on individuals. Due to practical difficulties only a small number of
communities are included. Random allocation of communities may not be feasi

130
ble. These studies are appropriate for diseases that have origins in social condi
tions which can most easily be influenced by intervention directed at group behav
iour as well as at individuals (Box 11.2).

Box 11.2 An example of a community interventional trial

Acute Respiratory Tract Infection: A Community Based Interventional

Study in Malaysia

Lye M. S. ; Nair R. C. ; Choo K. E. ; Kaur H. ; Lai K. P. F.;

Journal of Tropical Pediatrics 1996 42(3):138-143

A community-based intervention trial was conducted in Kelantan, Malaysia

with the aim of reducing severe acute respiratory tract (ARI) infection in
children. Interventions included health education of mothers on childhood
pneumonia and training of health staff on case management. In a house- to
house survey 1382 and 1107 children less than 5 years of age in the
intervention and control areas, respectively, were followed up every 2
weeks over a 62-week period. The reduction in the incidence of severe ARI
cases in the intervention area was significantly greater than in the control
area (P<0.05). The ARI mortality rates were low in both the intervention
and control areas (<0.1%). Our results indicate that with relatively
inexpensive methods and simple interventions, reduction of severe ARI
may be effectively achieved. This has important implications for an ARI
control programme in Malaysia and other developing countries.

RANDOMISED CONTROLLED TRIAL

Randomised controlled trials (RCT) are used for evaluating both the effective
ness and side effects of new types of intervention. Although the term randomised
control clinical trial is often used, the design is used for new treatment modalities,
including new health and medical care technology and organising and delivering
health services. The design follows stringent norms to reduce bias.

131
DESIGN OF A RANDOMISED CONTROL TRIAL

The Protocol
The study protocol comprises not only the explicit account of the study plan
but also a detailed diary of the execution of the plan. The purpose of the study pro
tocol is to:
a) Crystallise the project to the researchers themselves

b) Give referees the opportunity to review the project

c) Inform and educate all those taking part in the project

d) Ensure that the researchers do not forget any details of the plan, and to se
cure continuity
e) Document the procedures of the project for the future.

Selection of Subjects
The criteria for selecting the study subjects are stringent and often stated in
writing. The subjects are randomly chosen from the reference or target population.
They should be qualified or eligible for trial and must give informed consent. The
steps of a RCT are shown in Figure 11.1

Randomization
Participants are allocated into test and control groups at random. Random allo
cation permits chance to determine the assignment of subjects to various groups.
It eliminates selection bias and tends to create groups that are comparable in all
factors. It gives validity.
It is crucial that both groups are alike to ensure comparability between vari
ables that we recognize and are able to measure. However, there could be vari
ables that we may not recognize but which affect prognosis. Randomisation mini
mises the effect of these variables.

132
 Figure 11.1 Design of a Randomised Control Trial

Manipulation
In this step the investigator intervenes or manipulates the study group by delib
erate application or withdrawal, reduction of the suspected control factor being
tested, usually a drug. The control group may have a standard intervention, a pla
cebo (an inert substance that produces no effect) or no intervention at all.

Blinding (Masking)
The subjects need to participate without knowing which type of intervention is
being done on them. Similarly this information can be withheld from the assessor
and the person analysing the data. This process is called blinding. The types of
blinding are:
a) Single blind trial - participant is not aware of intervention received.

133
 b) Double blind trial– neither the doctor nor the participant is aware of interven
tion received.
c) Triple blind trial- Participant, investigator nor the person analysing the data
are aware of the intervention.

Follow Up
At defined intervals of time, both the groups are assessed for outcome of the in
tervention. The follow up is done with the same intensity and quality. A problem
of follow up is attrition due to death or drop outs. All must be included in the
analysis, i.e. in the denominator when comparing the proportions of successes or
failures.

Assessment of Outcome
The outcome must be specified in advance and expressed in term of positive re
sults which are beneficial to the subject. The negative or the severity and fre
quency of side effects and complications also need to be stated. Ideally the asses
sor should not know to which group the subject belongs. The blinding process
achieves this.

134
 Box 11.3 Example of a randomised control trial

Randomized, Controlled Intervention Trial of Male Circumcision for

Reduction of HIV Infection Risk: The ANRS 1265 Trial

A randomised, controlled, blindly evaluated intervention trial was carried out in

Orange Farm and surrounding areas, a semi-urban region close to the city of
Johannesburg. The recruitment of participants took place in the general
population from July 2002 to February 2004. At the end of the screen visit,
following screening and written consent, participants were divided into two
groups, using sealed envelopes. Each participant was invited by the manager
of the centre to choose an envelope containing the group name from a basket of
ten envelopes. After each randomisation, a new envelope was added to the
basket. This added envelope was taken sequentially from a set of envelopes
pre-prepared in such a way that each set of envelopes contained five for the
“Control” and five for the “Intervention” arm. Participants of the
intervention group were offered to be circumcised within a week.
Participants of the control group were asked to wait until the end of the trial
before being offered to be circumcised. A participant lost to follow-up was
defined as a participant who had not completed a planned visit in the 2 month
following the planned date of this visit and who did not complete any further
visit. A missing visit was defined as a visit not completed prior to a completed
visit. Each participant was invited to answer a face-to-face questionnaire, to
provide a blood sample, and to have a genital examination and an individual
counselling session. The counselling session (15-20 min) was delivered by a
certified counsellor and focused on information about STls in general and HIV
in particular and on how to prevent the risk of infection. To ensure
confidentiality, participants’ files were kept in a locked room at the centre and
each participant received a number that was used to identify all documents
related to that person. To ensure blinding of study personnel, the randomization
group information was not available to the personnel in charge of counselling
or collecting information in the centre during the participants’ visits.
Questionnaires were checked at the end of each interview. Participants failing
to turn up for any follow-up visit were visited at home by trial staff, who
encouraged them to come for the follow- up visits or ascertained the reasons for
dropping out. Laboratory results were stored in a database that was
independent of the one used to store the information related to each participant.
During the study, no HIV results were available to the investigation centre or to
the investigators, apart from the statistician in charge of the interim analysis.

Source PLOS Medicine

135
Stoppage
The trial is stopped when the desired period of observation is completed or be
fore the desired period when one treatment is clearly superior or adverse effects
are more than expected.

TYPES OF RANDOMISED CONTROL TRIALS

Concurrent Parallel Study Design

Comparisons are made between randomly assigned groups, one exposed and
another not exposed to specific treatment. Patient remains in the study or control
group for the duration of the investigation.

Figure 11.2 Concurrent Parallel Study Design

Crossover Type of Study Design

A crossover type of study design removes the individual variability between
subjects as each subject serves as his or her own control. All subjects, at some
time during the course of the investigation, receive the new therapy. The two
groups are switched; those who received the treatment under study are changed to
the control group, therapy or placebo, or vice versa. This design is not suitable for
illnesses of short duration.

136
 Figure 11.3 Crossover Type of study design study design

Clinical Trials
The randomised control design is used extensively for clinical trials. There are
many types of clinical trials. They are
(i) Prophylactic trial, e.g. immunisation, contraception
(ii) Therapeutic trial, e.g. drug treatment, surgical procedure
(iii) Safety trial, e.g. side effect of oral contraceptive
(iv) Effectiveness trial
(v) Risk factor trial, e.g. proving aetiology of a disease by inducing putative
agents in animals
(vi) Efficiency trial, e.g. efficiency of inserting IUCD.

PHASES

In the process of development of a new drug or vaccine the trials on humans

go through four phases. This is usually preceded by pre clinical studies.

137
Pre-clinical studies
These studies are conducted in vitro (test tube) and in vivo (animal) to estab
lish preliminary efficacy, toxicity and pharmacokinetic information of the candi
date drug.

Phase 0
Phase 0 is a recent designation for exploratory, first-in-human trials recom
mended by the Food and Drug Adminstration, USA, (FDA) in 2006. These trials
allow evaluation in humans of pharmacokinetic and pharmacodynamic prop
erty of the new drug by administration of low doses of drug for a short time.
There is no intent to diagnose or treat. Owing to the low doses administered and
the low risk of toxicity, the preclinical phase is reduced. They help in early selec
tion of candidate drug providing a potentially useful instrument for drug discov
ery, particularly in the field of oncology. Phase 0 studies are expected to reduce
costs and time for drug development. However, there are also concerns about the
utility and feasibility of Phase 0 studies.

Phase I
After considerable research on experimental animals, volunteers are institution
alised and receive a fraction of the anticipated dose of drug and are monitored for
effects. This phase requires high technology and various medical expertise.

Phase II
Once the initial safety of the study drug has been confirmed in Phase I trials,
Phase II trials are performed on larger groups (20-300) and are designed to assess
how well the drug works
The purpose of this phase is to assess the effectiveness of the drug or device, to
determine appropriate dosage and to investigate its safety. It is conducted on vol
unteers based on strict criteria.

138
Phase II studies are sometimes divided into Phase IIA and Phase IIB.
✴ Phase IIA is specifically designed to assess dosing requirements (how much
drug should be given).

✴ Phase IIB is specifically designed to study efficacy (how well the drug
works at the prescribed dose(s)).
Some trials combine Phase I and Phase II, and test both efficacy and toxicity.

Phase III
This phase of the trial is conducted on consenting hospital in-patients. The ran
domised control design is used in this stage. A large number of consenting pa
tients, depending upon the disease/medical condition are studied. This phase is
aimed at assessing how effective the drug is, in comparison with current ‘gold
standard’ treatment. These trials are the most expensive, time-consuming and diffi
cult.

Phase IV
Phase IV trial is also known as Post Marketing Surveillance Trial. Phase IV tri
als involve the safety surveillance (pharmacovigilance) and ongoing technical sup
port of a drug after it receives permission to be sold. The safety surveillance is de
signed to detect any rare or long-term adverse effects over a much larger patient
population and longer time period than was possible during the initial phases.

ETHICAL ISSUES
Ethics is concerned with the conduct of human beings. All scientific activities,
are conducted with the participation of human beings or have an impact on human
beings or on the wider society and environment. Laws directly tell us how to be
have (or not to behave) under various specific circumstances and prescribe reme
dies or punishments for individuals who do not comply with the law. Ethics is
what a body of professionals accept as a desirable or undesirable act.
139
 There are four universal principles of ethics
A. Beneficence – do good and maximise good outcomes
B. Non Malificience – do no harm and avoid unnecessary risk.
C. Respect & Autonomy– concern for autonomy of persons and courtesy
D. Justice & Equality– fair procedures and fair distribution of costs and bene
fits

HISTORY OF RESEARCH ETHICS

German medicine in the nineteenth and twentieth centuries served as a model

for modern medicine and medical practice, being linked to laboratory science.
Proven effectiveness based on rigorous experimentation usually on human sub
jects was the important requirement. However there were no guidelines for con
ducting these experiments. The German experiments on humans reached its peak
during the Second World War when several experiments were carried out on pris
oners in concentration camps.
After the war these experiments were judged as war crimes and crimes against
humanity, as the experiments were conducted without the consent of the partici
pants and most of the subjects died or were permanently crippled.
From these trials evolved the Nuremberg Code in 1948, which stated “The vol
untary consent of the human subject is absolutely essential,” for experiments, mak
ing it clear that subjects should give consent and that the benefits of research must
outweigh the risks.
Human research is now governed by the Declaration of Helsinki (1964) by the
World Medical Association. The Declaration is the basis for “Good Clinical Prac
tices” used in research today.

140
 ISSUES ADDRESSED IN THE DECLARATION OF HELSINKI INCLUDE:

‣Research with humans should be based on the results from laboratory and ani
mal experimentation
‣Research protocols should be reviewed by an independent committee
prior to initiation
‣Informed consent from research participants is necessary
‣Research should be conducted by medically scientifically qualified individu
als
‣Risks should not exceed benefits
‣Essentiality: Is the study required?
‣Principle of Uncertainty - a genuine doubt if the intervention is beneficial.

‣Maximization of public interest and of social justice

A clinical trial should conform to the above. During the course of the trial,
whenever it is known that one of two strategies is superior or detrimental the trial
should be stopped and the superior therapy be made available to all study partici
pants. Results should be reviewed as trial progresses.

QUASI –EXPERIMENTS
When one characteristic of a true experiment, i.e. manipulation, control or ran
domization, is missing it is a quasi-experimental study.

NON-RANDOMIZED TRIAL

It is not always possible for ethical, administrative and other reason, e.g. cost
and logistic to resort to a randomized control trial in human beings, in such a situa
tion one depends on a non-randomized trial. When there is no randomization, the
degree of comparability will be low and the chances of spurious results high.

141
Natural Experiment
When a naturally occurring event or situation is exploited by a researcher to
help answer a research question, it is called a natural experiment. The researcher
has little or no control over the situation that is being observed. A good example
of natural experiment is one by James Lind in 1747 on the prevention of scurvy
among sailors. He compared the effects of different acidic substances, ranging
from vinegar to cider, on groups of afflicted sailors, and found that the group who
were given oranges and lemons had largely recovered from scurvy after 6 days.

Clinical trial In Malaysia

Malaysia is emerging as a centre for conducting clinical trials for drugs
especially for drugs related to diabetes and hypertension. Doctors desirous of
conducting or participating in a clinical trial need to register with the
National Medical Research Register which is the web based tool designed to
support the implementation of the National Institute of Health (NIH)
guidelines on the conduct of research in the Ministry of Health Malaysia
(MOH). The doctor must undergo training for “Good Clinical Practice” and
follow the Malaysian Guidelines for Good Clinical Practice of the Ministry
of Health.

Summary Box Chapter 11

• Phases of a Randomised control trial (RCT)

• Phase I (safety, feasibility, acceptability)
• Phase II (preliminary effect size, side effects, dosing)
• Phase III (efficacy of new treatment compared to standard)
• Phase IV (application in clinical setting; long-term safety, fidelity)

142
ADDITIONAL READING
Field trials of medical decision-aids: potential problems and solutions. J. Wyatt
and D. Spiegelhalter Medical Informatics, Stanford University, CA 94305-5479.
http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=2247484&pageindex
=1#page
Nuremberg trials http://www.ushmm.org/research/doctors/
Helsinki declaration http://www.wma.net/e/ethicsunit/helsinki.htm
National Medical Research Register https://www.nmrr.gov.my

143
 CHAPTER 12

Screening

Screening is the practice of investigating apparently healthy individuals with the

objective of detecting unrecognized disease or its precursors in order that meas
ures can be taken to prevent or delay the development of disease or improve the
prognosis.
In many diseases the pathological process is established long before the appear
ance of the symptoms and signs which alert people of the need to seek medical ad
vice. By this time the disease process and the consequent damage may be irreversi
ble or difficult to treat.
Screening is the process of testing for infection or diseases in populations or in
individuals who are not seeking health care.

USES OF SCREENING
1. Case Detection (Prescriptive Screening)
The presumptive identification of unrecognized disease which does not arise
from a patient’s request, i.e. people are screened for their own good.

2. Control Of Disease (Prospective Screening)

People are examined for the benefit of others.

144
 3. Research Purposes
For many chronic diseases whose natural history is not fully known. Partici
pants should be informed that no follow-up therapy will be available.

4. Educational Opportunities
Opportunity for creating public awareness and for educating health profes
sionals.

TYPES OF SCREENING

‣Opportunistic Screening - is done only when the opportunity arises, it is usu

ally done in a clinical setting.

‣Mass Screening - is the screening of a whole population or a subgroup. There

is no reference to risk and usually no follow up. It is performed in a commu
nity setting.
‣High risk or Selective Screening - is done to detect a specific disease or pre
disposing condition in people who are known to be at high risk of having or
developing the condition.
‣Two stage Screening - is conducted on those who screen positive the first
time and are recalled for further testing. This process is less expensive as the
less invasive and less uncomfortable test is done first.
‣Multiphasic Screening - is the application of two or more screening
tests in combination to a large number of people at one time.

145
 Figure 12.1 HIV screening Coverage and HIV prevalence among ante-natal women in Malaysia,
1998 – 2002
Source: Consensus_report_MAA_2004.pdf

Figure 12.2 Process of Screening

146
 CRITERIA FOR SCREENING

1. The condition should be an important health problem (high mortality, dis

ability, discomfort, financial cost).
2. There should be an acceptable treatment for patients with recognised dis
ease.
3. Facilities for the diagnosis and treatment should be available (adequate fol
low up for positive cases).
4. There should be a recognisable latent or early symptomatic stage so that in
dividuals benefit by early detection.
5. There should be a suitable test or examination (quick, easy, inexpensive,
sensitive, specific).
6. The natural history of the condition, including development from latent to
declared disease, should be adequately understood.
7. There should be an agreed policy concerning whom to treat as patients.
8. The cost of case finding (including diagnosis and treatment of patients diag
nosed) should be economically balanced in relation to possible expenditure
on medical care as a whole.
9. The test should be acceptable to the population (safe and acceptable).
10. Case finding should be a continuous process and not a “one off’ procedure.
11. There should be a substantial burden of the disease in the community (not
too rare).

CHARACTERISTICS OF A SCREENING TEST

For a screening test to be useful it should have the following characteristics

147
VALIDITY

Validity is the ability of the test to measure what it intends to measure. It has
two components, sensitivity and specificity. The validity of a test is affected not
only by the characteristics of the test but by host factors such as stage of disease
and presence of other conditions.
The sensitivity and specificity characteristics of the tests help in making the de
cisions on whether or not to use the test. They are generally regarded as independ
ent of disease prevalence.

SENSITIVITY

It is defined as the ability of a test to identify correctly those who have the dis
ease i.e. the true positives. Sensitive tests are used in cases where there is an im
portant penalty for missing a dangerous but treatable disease, e.g. TB, syphilis,
and to rule out diseases in the early stages of diagnostic work up.
If the test is not sensitive it will fail to detect some of the people with the dis
ease. These are called false negative. The importance of false negative is that seri
ous diseases could be missed, and if the disease is curable in the early stages a
false negative result may mean a virtual death sentence to the individual. The
lower the sensitivity the larger will be the false negatives.

SPECIFICITY

It is defined as the ability of a test to identify correctly those who do not have
the disease, i.e. the true negatives. It is useful to confirm the diagnosis that has
been suggested by other data. If the test is not specific it will detect a large num
ber of people who are false positive, i.e. they are shown to have disease when they
do not have it. The disadvantage is that it causes a lot of anxiety and worry to the
individual and in some diseases, stigma. Further, it puts a burden on the health
care system.

148
ACCEPTABLE LEVELS OF SENSITIVITYAND SPECIFICITY

It is usually not possible to have a test that is both highly sensitive and highly
specific. Reducing the strictness of the criteria for a positive test increases sensitiv
ity but decreases specificity. Increasing the strictness of the criteria increases speci
ficity but decreases sensitivity. There is generally a trade off.

Another way to address trade offs is to use the results of several screening tests
together in either a parallel process or in series. Parallel screening increases sensi
tivity but lowers specificity (false positives are more likely); series screening tests
- initial screening test followed by an additional screening procedure - increase
specificity.
The determination of the sensitivity and specificity of the test is important as
for every person tested, regardless of whether he or she tests positive or negative,
a diagnosis of the disease is established or ruled out.

ESTABLISHING THE SENSITIVITYAND SPECIFICITY

Sensitivity : a/ (a+c) x 100

Specificty: d / (b+d) x 100
Percentage of false negative: c / (a+c) x 100
Percentage of false positive : b / (b+d) x 100

Note: Disease status is established by the currently available “gold standard”

149
ACCURACY
The proportion of true test results among all test results can be determined this
way: (a+d)/(a+b+c+d)

PREDICTIVE VALUE OF SCREENING TEST

The ability to predict the presence of disease from results is dependent on
prevalence of the disease in the population tested as well as the sensitivity and
specificity of the test. The higher the prevalence the higher the predictive value;
therefore, a screening programme is most productive and efficient if targeted at a
high-risk target population. Screening a total population for a relatively infrequent
disease is a waste of resources and will give a low yield of positives.
The measure referred to as the predictive value of a positive test is a proportion
of those with a positive test who have the disease. The predictive value of a nega
tive test is the proportion of non-diseased individuals among all those who have
negative test results.
Only when prevalence reaches 15-20 percent a respectable predictive value is
achieved. The result of any test must be interpreted in the context of the preva
lence of the disease in the population from which the subject originates.
Predictive value of a Positive test: a /(a+b) x 100
Predictive value of a negative test: d/ (c+d) x 100

RELIABILITY (PRECISION)
Reliability or repeatability is the factor that gives a consistent result when the
test is performed more than once on the same individual under the same condi
tions. Factors that affect consistency of results are either variations inherent in the
method or observer variations which can be interobserver variations or intraob
server variations.

150
YIELD
It is the extent of a previously unrecognised disease that is diagnosed and
brought to treatment as a result of screening. The yield is dependent on the sensi
tivity of the test, prevalence of the unrecognised disease, the individuals participa
tion in screening and follow up, and frequency of screening which would depend
on the disease’s own natural history, the incidence of disease and the individual
difference in risk.

THE PROBLEM OF BORDERLINE

A prior decision is made about the cut-of point on the basis of which individu
als are classified as normal or diseased (if we increase the sensitivity by lowering
the cutoff level we decrease the specificity. If we increase the specificity by rais
ing the cutoff level we decrease the sensitivity).

The following factors are taken into consideration:

• Disease prevalence - When prevalence is high in the community the screen
ing level is set at a lower level which will increase sensitivity.
• The disease - If disease is lethal and early detection markedly improves prog
nosis, a greater degree of sensitivity even at the expense of specificity is de
sired.

Evaluation of Screening Programmes

(i) Randomised control trials - should ideally be performed in the same setting
where screening programmes will be implemented and should employ the
same type of personal equipment and procedures that will be used in that
programme.
(ii) Uncontrolled trials - are conducted to see if people with disease detected ap
pear to live longer after diagnosis and treatment than patients not screened.
(iii)Other methods - include cohort studies and comparison in trends between
areas of different degrees of screening.

151
ERRORS IN SCREENING
Lead time bias - The interval between the time a condition is detected through
screening and the time it would normally have been detected by the reporting of
signs and symptoms is referred to as lead time bias. The bias occurs when screen
ing detects disease earlier in its natural history than would otherwise have hap
pened, so that the period of time from diagnosis to death appears to be lengthened
when in fact the survival is unaffected by treatment (Figure 12.1). Having addi
tional lead time may not alter the natural history of the disease and therefore may
not extend the length of life. This lead time bias tends to operate in screening for
cancers, no matter how aggressive the tumours are.

Figure 12.3 Lead time bias

Length time Bias – cases detected through a periodic early detection pro
gramme tend to have longer preclinical stages than those missed by screening but
are self –detected between examinations. Screen detected cases have a better prog
nosis than symptom detected cases as they may have a less severe disease and sur
vive longer, this length bias creates an apparent advantage for screen detected
cases that may not exist in reality (Figure 12.2).

152
 Figure 12.4 Length time bias
Source: Treatment Options for Prostate Cancer: Evaluating the Evidence. VIBHA BHATNAGAR. ROB
ERT M. KAPLAN. American Family Physician, 2005 15;71(10):1915-1922.

Length time bias occurs when the full spectrum of a particular tumour, such as
prostate cancer, is composed of cancers that range from very aggressive to indo
lent. Persons with less aggressive tumours are more likely to be discovered
through screening programme and therefore are likely to survive longer after de
tection regardless of the treatment given.
Patient Self-Selection Bias - Volunteers may differ from those who do not vol
unteer in characteristics, which may be related to survival.

INTERPRETATION ERROR

The Judges dilemma: The judge always starts with the premise “Not – guilty
unless proved otherwise”. However, he can make an error sometimes and free
the guilty or punish the innocent. He has committed an error. Similarly, in statis
tics we start with the premise “there is no difference”. When we conclude that
there is a difference between groups when actually there is none, we make a type

153
I error (false positive). When we conclude that there are no significant differences
when actually there are, we commit a type II error (false negative). In medicine la
beling a person sick when he is actually not is a type I error and the reverse is a
Type II error .
Practical examples using sensitivity, specificity, gold (reference) standard,
positive predictive value, and negative predictive value.

Table 12. 1 Calculation of Specificity and Sensitivity

Table 12.2 Calculation of positive predictive value.

154
 MyAlice company has developed a new ELISA test to screen for HIV infections.
To find out the numbers that the test might miss, the test was done on the serum
from 1,000 patients who were already proved positive by Western Blot (the gold
standard assay). 999 were found to be positive by the new ELISA. To find out the
false positives the manufacturers then used the ELISA to test serum from 1,000
nuns who denied any risk factors for HIV infection. The serum of the ten who were
found positive were tested by Western Blot and found negative. The other 999
were proved negative. (See Table 12.1)
MyAlice company is very happy to have a an excellent test with 99.9% sensitiv
ity and 99% specificity.

The company asked Dr. Skep Tic to use the test kit. With a sensitivity of 99.9%
and a specificity of 99%, the ELISA appears to be an excellent test. However, Dr.
Skep Tic wondered what would happen when the test is applied to a population of
a million people where 1% are infected with HIV. Of the million people, 10,000
would be infected with HIV (see Table 12.2)

DR SKEP TIC REALISES THAT

‣Sensitivity and specificity do not take into account the prevalence of the dis
ease.
‣It is more important to take into account the positive predictive value which
is related to the disease prevalence. If the prevalence was low the positive pre
dictivity would be low.
‣If a clinician made a wrong diagnosis due to a test which would be a more
grievous error: (a)to tell a patient that he or she is well, when, in truth the pa
tient has a serious illness (false-negative ), or (b)to tell the patient that he or
she is ill with a serious condition when in reality the patient is healthy (false
positive)?
‣Often clinicians would contend that missing a diagnosis is a much more se
vere mistake than a false-positive error. However, the latter has the negative
consequences of producing worry, concern, and unnecessary treatment.

155
ADDITIONAL READING
Editorial. Guidelines Based on Fear of Type II (False- Negative) Errors. Why
We Dropped the Pulse Check for Lay Rescuers. Resuscitation Volume 46, Issues
1-3, 23 August 2000, Pages 439-442

156
 CHAPTER 13

Association &
Causation
The essence of epidemiology is to determine the causation of disease, i.e. to find
out the specific cause or causes of the disease and to assist in its prevention and
control. In the quest for the cause we often encounter several factors which are not
actually the causative factor but are only associated with the disease condition.

ASSOCIATION
An association is said to exist between two variables when a change in one vari
able parallels or coincides with a change in another. This is also called “covaria
tion” or “correlation”. An association or covariation may be positive or negative
and may be proportionate or disproportionate.
An association is said to be causal when it can be proven that the presence of
an independent variable (exposure) produces a change on the dependent variable
(disease).
The association between two variables may be real or spurious.

157
 Figure 13.1 Association of water borne diseases and rainfall
Source: www.wpro.who.int/NR/rdonlyres/3FB0A304-554E-4637-A3A0-3443036E56BC/0/MAA.pdf

1.REAL ASSOCIATION

This exists when an independent variable (risk factor) causes changes in a de

pendent variable. This can further be classified as direct causal, indirectly causal
or interaction causal.
(a) Direct Causal Association - is inferred when the risk factor or independent
variable changes the dependent variable or condition directly without inter
vening variables.
Factor Disease
E.g. Tubercle bacillus Tuberculosis
Lead Lead poisoning
Iodine deficiency Goitre

158
 (b) Indirect Causal Association - is when the risk factor causes changes in the
dependent variable or condition through the mediation of other intermediate
variables or conditions.
E.g. Factor Step 1 Step 2 Disease
Salt intake HPT CHD
(c) Interaction (Including Conditional*) Causal Association - there may be in
teractions (positive or negative) between categories of independent vari
ables that produce changes in the dependent variable synergism (greater or
smaller effect)
E.g. Measles Death
Malnutrition Death
Measles + malnutrition Higher case fatality
* Is when two risk factors are incapable of producing a condition unless they
exist in the presence of each other.

NON-CAUSAL, SPURIOUS ASSOCIATION

This is when an association between two variables is statistically significant

but no causal relationship actually exists. These occur:
(i) Due to chance
(ii) Based on numerator analysis or ecological correlate
(iii) Due to bias.

CAUSATION
A cause of a disease is an event, a condition, characteristic or a combination of
these factors which plays an important role in producing the disease. A particular
cause may be necessary, sufficient, neither or both. A cause is termed sufficient
when it inevitably produces or initiates a disease and is termed necessary if a dis

159
ease cannot develop in its absence. A cause is not usually a single factor but often
comprises of several components.

FACTORS IN CAUSATION
Many factors are involved in causation of a disease. They may all be necessary
but are rarely sufficient, independently, to cause a particular disease or state.
Hence they can be classified as follows:
(a) Predisposing Factors - Factors that may create a state of susceptibility to a
disease agent like age, sex, previous illness.

Figure 13. 2 Factors associated with causation of tuberculosis

(b) Enabling Factors - are the circumstances that assist in causation or recov
ery from illness or in the maintenance of good health e.g. income, nutrition,
housing.
(c) Precipitating Factors - are factors that initiate the disease process e.g. expo
sure to a specific agent or a noxious agent.

160
 (d) Reinforcing Factors: factors that compound the effect such as repeated expo
sure and unduly hard work may aggravate the disease.

Figure 13.3 Assessing the relationship between a possible cause and on outcome

ESTABLISHING THE CAUSE OF ADISEASE

The process of determining whether observed association is likely to be causal

is called causal inference. In this process guidelines are used and judgments made.
Before concluding that any association observed is causal, other explanations such
as chance, bias and confounding have to be excluded.

161
EARLY SCHEMES IN ASSESSING CAUSALITY
Infectious diseases were initially the major problem affecting humans. In try
ing to identify the causes for these diseases the question arose as to what is the evi
dence required to prove that a particular organism was the causative one. These
are stated in the postulates put forth by Robert Koch.

KOCH’S POSTULATES:
‣The organism must be present in every case of the disease.
‣It must be possible for the organism to be isolated and grown in pure culture.
‣The organism must, when inoculated into a susceptible animal, cause the spe
cific disease.
‣The organism must then be recovered from the animal and identified.

As non-infectious diseases gained importance, the relevance of Koch’s postu

lates decreased. The question arose as to what would represent strong evidence of
their causation. The following are the guidelines for causation of disease.

GUIDELINES FOR CAUSATION

Temporal Relation
Cause must precede the effect. (Essential)

Plausibility
Is the association consistent with current knowledge? For example evidence,
from experimental animals. Lack of plausibility may simply reflect lack of medi
cal knowledge.

162
Consistency
Have similar results been shown in other evidences especially when a variety
of designs are used in different settings (meta analysis)? Lack of consistency does
not exclude a causal association because different exposure levels and other condi
tions may reduce the impact of the causal factor in certain studies.

Strength
What is the strength of association between the cause and effect? This is meas
ured by relative risk. The Higher relative risk, the stronger is the likelihood of the
factor being causative. A Relative Risk of greater than two can be considered
strong.

Dose - Response Relationships

Is increased exposure to the possible cause associated with increased risk?

Reversibility
Does the removal of a possible cause lead to reduction of disease ‘risk? If the
cause leads to rapid irreversible changes that subsequently produce disease,
whether or not there is continued exposure, then reversibility cannot be a condi
tion for causality. e.g. paralytic diseases.

Study Design
Is the evidence based on a strong study design? Table 13.1 gives the hierarchy
for the best designs to prove causality.

163
 Table 13.1 Types of studies and ability to prove causation

Judging the Evidence

In judging the different aspects of causation the correct temporal relationship is
essential, the greatest weight may be given to plausibility, strength of association,
consistency and dose-response relationship.
The likelihood of a causal association is heightened when many different types
of evidence lead to the same conclusion. Evidence from a well-designed study is
particularly important especially if they are conducted in a variety of locations.

Summary Box Chapter 13

• An association between two variables exists when change in one is

associated with a change in the other
• Association may be spurious or real.
• A causative factor is one which actually plays an important role in
producing the disease
• In addition to being associated with the disease, a causative factor must
satisfy other important criteria.
• Predisposing, enabling, precipitating or reinforcing factors are involved in
the causation of disease

164
ADDITIONAL READING
http://www.smokershistory.com/JFGlenn.htm
http://goliath.ecnext.com/coms2/gi_0199-4972440 Association-or-causation
evaluating- links.html
http://askdatasystems.com/EPIDEMI/ep/epimod1.htm

165
 CHAPTER 14

Communicable
Disease Epidemiology
PATTERNS OF DISEASE
Developing countries begin with a disease burden dominated by nutritional,
peri- natal, and infectious diseases and in the process of development, make the
transition to one dominated by non-communicable diseases, particularly Cardio
vascular Diseases (CVD). Four stages are identified in this transition.
For countries in the earliest stage of development, the predominant diseases are
due to infections, and nutritional deficiency-related disorders (Figure 14.1, Table
14.1). In the second stage, as infectious disease burdens are reduced and nutrition
improves, diseases related to hypertension, such as haemorrhagic stroke and hyper
tensive heart disease, become more common. In the third stage, as life expectancy
continues to improve, high-fat diets, cigarette smoking and sedentary life-styles be
come more common. Non-communicable diseases then predominate, with the
highest mortality caused by atherosclerotic CVD, most frequently ischaemic heart
disease and atherothrombotic stroke, especially at ages of less than 50 years. In
the fourth stage, increased efforts to prevent, diagnose, and treat ischaemic heart
disease and stroke are able to delay the impact of these diseases to more advanced
ages. Despite the shift described above, communicable diseases both new and old
continue to cause outbreaks from time to time (Figure 14.2).

166
 Figure 14.1 Global burden of disease 1990-2020 by disease group in developing countries
Source: World health organisation, evidence information and policy 2000

Table 14.1 Leading Causes of death in developed and developing countries – 2001
Source: WHO World Health Report 2002. Countries grouped by WHO Mortality Stratum, with Developing
Countries representing regions with High and very High Mortality. and Developed Countries representing re
gions with Low and Very Low Mortality

167
 Figure 14.2 Major outbreaks in western pacific region
Source: WHO/ WPRO

The increasing population and development puts a burden on the environment.

Some examples of the diseases caused by a deteriorating environment are:
‣Acute and chronic respiratory disorders, including asthma and allergies from
pollution
‣Toxic poisonings from water and food due to contamination from hazardous
wastes
‣Skin cancers and cataracts from the increased ultraviolet radiation caused by
the depletion of the ozone layer in the atmosphere, and
‣Increases in vector-borne diseases as the phenomenon of global warming
changes the distribution of diseases dependent upon climatic factors for their
spread.

168
PATTERNS OF DISEASES IN MALAYSIA
Malaysia is in the epidemiological transition and the focus of public health is
shifting from infectious to non-communicable diseases. The rising trend in smok
ing will cause an increased proportional mortality due to diseases related to smok
ing such as heart disease and lung cancer. Infectious diseases such as malaria have
come down; however dengue remains a major problem. New diseases such as the
Nipah virus, Hand, Foot and Mouth disease, and the possibility of Avian Influenza
epidemic are a cause of concern. HIV infections and AIDS cases are also on the
rise.
The factors that play a role in the changing patterns of disease are multiple.
They include:

‣Changing life styles and living standards,

‣Demographic factors, urbanisation and industrialisation,
‣Medical interventions, maintenance of people with transmissible genetic de
fects
‣Widespread effects of technology on ecology.

THE REASON FOR DISEASE VARIANCE

CHANGES IN POPULATION
(a) Consequence of Population Growth

The heaviest burden of ill health falls on countries with fastest growing popula
tions and an unsustainable economic development. Population growth and rapid
urbanization cause overcrowding, unhygienic conditions, lack of clean water and
inadequate sanitation which are breeding grounds for infectious diseases. Respira
tory and food and water-borne disease transmissions are very common in high den
sity populations. Other causes are human encroachment on tropical forests, and en
vironmental change.

169
(b) Movement of Populations

Internal and international population movement provides opportunities for

spread of disease. Faster means of travel and globalisation facilitate the spread of
disease.

CHANGES IN AGENT

The development of new and resistant strains of certain agents also contribute
to disease variance e.g. leprosy bacillus to dapsone, typhoid bacillus to chloram
phenicol, plasmodia resistance to chloroquine, penicillinase producing gonococci
etc.

INCOME DISTRIBUTION

Poor living conditions cause exposure to infections and a direct cause of malnu
trition is poverty. Due to poverty, health systems are unable to afford protection
against diseases.

OTHER FACTORS

‣Social and behavioural change

‣Cultural practices
‣Sanitation standards
‣Nutrition and socio economic status
‣Climatic and geographical factors
‣Global food trade
‣Availability of preventive and curative facilities

170
COMMUNICABLE DISEASE
A communicable disease is an illness due to a specific infectious agent or its
toxic products that arises through transmission of that agent or its products from
an infected person, animal or reservoir to a susceptible host, either directly or indi
rectly through an intermediate plant or animal host, vector or inanimate environ
ment.
Infection is the invasion by and multiplication of pathogenic micro organisms
in a bodily part or tissue, which may produce subsequent tissue injury and pro
gress to overt disease through a variety of cellular or toxic mechanisms. Infections
are classified as:
‣Sub-clinical - Infection is present but there are no signs or symptoms. The
changes are at the tissue or bio-chemical level.
‣Acute - Symptoms become manifest and occur soon after infection.
‣Chronic - Even after the symptoms pass and the host becomes apparently nor
mal, the organism continues to be present. Often viral infections persist at
low levels and are difficult to control. Often it is hard to tell infected from un
infected.

171
 Figure 14. 3 The modes of transmission of diseases

There are several ways that a disease organism affects the host. This is called
the Pathogeneic Mechanism. The different modes are:
‣Direct tissue invasion
‣Toxin production
‣Persistent or latent infection
‣Enhanced susceptibility to drugs
‣Immune suppression
‣Immunologic enhancement or allergic reaction leading to damage to host.

172
 Box 14.1 Common Terminologies in Disease Transmission

Infection is said to have occurred if an infectious agent has entered and

established itself in a host.

Infectivity is the ability of an agent to invade and multiply in a host.

Pathogenicity is the ability to produce clinically apparent illness.

Virulence is the proportion of clinical cases resulting in severe clinical

manifestation. Virulence may depend on dose, route of infection and host
factors such as age or race.

Immunogenicity is the ability to produce specific immunity, primarily

humoral, cellular immunity, or a mixture of both· in the host.

A carrier is a person without apparent disease who is capable of transmitting

the agent to others; carriers may be asymptomatic, i.e. who never show
symptoms during the time they are infected. Incubatory or convalescent
carriers are capable of transmission before or after they are clinically ill.

Chronic carrier is one who harbours an agent for an extended time

following the initial infection.

Generation time is the period between the receipt of infection by the host
and the maximal communicability of that host.

Incubation time is the time interval between the receipt of infection and the
onset of illness.

Herd immunity is the resistance of a community to a disease. Is not

necessary to achieve 100 percent immunity in a population in order to halt an
epidemic or control a disease.

173
 Box 14.2 Characteristics of Time

Endemic – diseases which are regularly and continuously present.

Epidemic – a significant excess over that expected on basis of past

experience, an unusual clustering over time.

Pandemic – an epidemic occurring across several geographic regions

Short time variation – outbreaks, point epidemics

Periodic variations:
• seasonal changes – changes in disease occurrence with seasons,
• cyclical variations - changes in disease occurrence every two to three
years. This occurs due to accumulation of susceptible or changes in
environmental conditions (Figure 7.1)
•secular variations – changes over decades. A secular trend can be
influenced by intervention strategies e.g. introduction of immunisation
(Figure 7.2).

Knowledge of seasonal and cyclical trend of disease is useful for

surveillance, planning, preparedness and control of outbreaks and
epidemics

CHAIN OF INFECTION
Transmission occurs when the agent leaves its reservoir or host through a por
tal of exit and is conveyed by some mode of transmission and enters through an
appropriate portal of entry to infect a susceptible host.

THE INFECTIOUS AGENT

Any microorganism that is capable of producing an infection. These include

bacteria, viruses, parasites and fungi.

174
 RESERVOIRS

The habitat in or on which an infectious agent normally lives, grows and multi
plies is called the reservoir. Reservoir may be humans, animals and the environ
ment. However, it may or may not be the source from which an agent is trans
ferred to a host. The source of infection is the person, animal or object from which
the host acquires the infective agent.

Figure 14.4. Portals of Entry and Exit in Mammals

Source. Modified from Mims CA. Dimmock NJ. Nash A el al. Pathogenesis of infectious diseases. ed 4. London.
Academic Press. 1995. P1O

175
 Figure 14.5 The Chain of Transmission for Tuberculosis

PORTAL OF EXIT

Is the path, by which an agent leaves the reservoir or host, it usually corre
sponds to the site at which the agent is located in the host (Figure 14.4).

176
 Figure 14.6 Methods to break the chain of transmission in tuberculosis.

MODES OF TRANSMISSION
The mechanism of spread of infection through the environment or through an
other person is called mode of transmission. Transmission may be direct or indi
rect.
(a) Direct Transmission: There is an immediate transfer of the agent from a
reservoir to a susceptible host by direct contact, e.g. sexual intercourse, fae
caloro or through droplet spread.
(b) Indirect Transmission: When transmission occurs through another me
dium such as air, water, soil, inanimate objects or through arthropods it is
called indirect transmission.
(i) Airborne transmission: The commonest mode of transmission. Drop
lets are generated from the respiratory system when a person coughs,
sneezes, talks or has a medical procedure. These droplets carry the infec
177
 tious agent to the susceptible person. Droplet nuclei are very tiny particles
that represent the dried residue of droplets.
(ii) Arthropods transmitting diseases are called vectors. The vector may be
truly biological if the agent multiplies in the arthropod before it is transmit
ted. If the agent only multiples in the vector it is called a propagative cy
cle eg. Dengue virus, if the agent has part of the life cycle in the vector eg,
filariasis, it is called cyclo developmental, if the agent has part of its life
cycle and also multiplies in numbers e.g. in Malaria it is called cyclopropa
gative type of transmission. The host wherein the sexual cycle occurs is
called the definitive host and the one where the asexual part of develop
ment takes place, is the intermediate host. Vectors can also transmit dis
ease by mechanical means.
(iii) Another common medium of transmission is water.

Portal of Entry
The agent enters a susceptible host through a portal of entry, which must pro
vide access to tissues in which the agent can multiply or a toxin can act. The infec
tious agent can enter through the mouth, through a wound or be breathed in.

SUSCEPTIBLE HOST

Susceptibility of a host depends on genetic factors, specified acquired immu

nity and other general factors which alter an individual’s ability to resist infection
or to limit pathogenicity.

PRINCIPLES OF CONTROL
An understanding of the dynamics of disease transmission is necessary to con
trol it.
Control: It is the reduction of disease prevalence in the community. The trans
mission still occurs but is no longer a major public problem.

178
 Elimination: When the pathogen ceases to exist in the human host it is called
elimination, e.g. poliomyelitis in the western hemisphere.
Eradication: it is the complete removal of the pathogenic organism in all its
forms from both the human and the environmental reservoirs, e.g. smallpox.

METHODS OF CONTROL

a. Elimination of Reservoir
Where man is the reservoir, one could find and treat all patients and carriers.
For zoonosis, it can be done by elimination of the suspect hosts, e.g. culling of
poultry for avian influenza.

b. Interruption of Transmission
This is achieved by improving environmental sanitation, personal hygiene, and
control of vectors by insecticides and pesticides. Isolation of the patient, quaran
tine of those suspected to being infected and limiting exposure to infected animals
are also strategies for interrupting transmission.

c. Protection of Susceptible Host

The susceptible host can be protected by immunization, routine chemoprophy
laxis especially during epidemics, or by preventing contact with vectors.

d. Notification and Legislation

When the occurrence of disease is reported to a specified authority it is called
notification. Laws are also enacted to control certain diseases, e.g.
‣Notifiable communicable diseases - Communicable Disease Act 1998
‣Powers of health officer - Food Act 1998, Destruction of Disease Bearing In
sect Act.

179
 ‣International Health regulations.

e. Surveillance
Public health surveillance refers to an ongoing and systematic collection, analy
sis and interpretation of health data in the process of describing and monitoring a
health event.
This information is used for planning and evaluating public health intervention
and public health programmes. The basic idea behind this system is observing, re
cording and collecting facts, analysing them and considering a reasonable course
of action.

EPIDEMIC INVESTIGATION AND MANAGEMENT

The biggest challenge in epidemiology is investigating and controlling disease
outbreaks and epidemics

REASON FOR INVESTIGATING A POSSIBLE OUTBREAK

‣Control/prevention: Prevent the occurrence of additional cases and outbreaks

in the future.
‣Research opportunities: To study natural history of the disease. Gain knowl
edge by assessing control measures and usefulness of new epidemiologal and
lab technique’s
‣Training of health personnel.
‣For public, political or legal concerns.
‣Programme considerations - may reveal weakness in a health programme.
‣May identify populations overlooked, failures in interventional strategy,
changes in agents or events beyond scope of the programme.

180
STEPS IN OUTBREAK INVESTIGATION

In the investigation of an ongoing outbreak, quick response and a systematic

approach are important. The steps below are in conceptual order. In practice the
steps may not be in the same order and several may need to be done simultane
ously.
1. Preparing For Field Work
(a) Investigation - Adequate scientific knowledge of the problem on hand is
needed and necessary supplies and equipment are collected. If needed, ex
perts on the subject may have to be called in or consulted.
(b) Administration - Administrative procedures such as approval for travel
plans and budget are obtained.
(c) Consultation - Contacts with local staff and officials should be established.
This includes addresses and phone numbers and details of meeting sched
ules. Every person involved in the investigation must know their roles espe
cially when there are teams from outside the immediate area.

2. Establishing the Existence of an Outbreak

Epidemic is the occurrence of more cases of disease than expected in a given
area or among a specific group of people over a particular period of time.

181
 Box 14.3 Case Definitions for Infection with Influenza A (H1N1) Virus

A confirmed case of swine influenza A (H1N1) virus infection is defined as a

person with an acute febrile respiratory illness with laboratory confirmed
influenza A (H1N1) virus infection at CDC by one or more of the following
tests:
1. real-time RT-PCR
2. viral culture

A probable case of swine influenza A (H1N1) virus infection is defined as a

person with an acute febrile respiratory illness who is:

* positive for influenza A, but negative for H1 and H3 by influenza RT- PCR,
or
*positive for influenza A by an influenza rapid test or an influenza
immunofluorescence assay (IFA) plus meets criteria for a suspected case

A suspected case of swine influenza A (H1N1) virus infection is defined as a

person with acute febrile respiratory illness with onset

* within 7 days of close contact with a person who is a confirmed case of

influenza A (H1N1) virus infection, or
* within 7 days of travel to community either within the country or
internationally where there are one or more confirmed influenza A(H1N1)
cases, or
* resides in a community where there are one or more confirmed influenza
cases.

Cluster is an aggregation of cases in an area over a particular period without

regard to whether the number of cases is more than expected. It is a non-uniform
distribution of cases in a study area. In an outbreak, we assume the cases are re
lated or they have a common cause, and clustering gives an important clue to the
investigation. It is important to decide whether a cluster of cases is an outbreak
and to establish if the cases are in excess of expected (a rapid comparison with
available records is done). Mapping of cases can identify clustering. Excess num
bers may not necessarily be an outbreak; it may be because of changes in local re
porting procedures, etc.
182
3. Verifying the Diagnosis
Diagnosis should be verified to ensure that the problem has been properly diag
nosed and to rule out laboratory error. The clinical findings and laboratory results
should be reviewed for inconsistencies. Clinical findings should be summarised as
frequency distributions and by age, sex and other factors. This step is useful for
characterising the spectrum of illness, verifying diagnosis and developing case
definitions. Efforts should be made to get the time frame of each of the clinical
symptoms. Several patients should be interviewed to verify the symptoms and
elicit information on possible exposure.

4. Establishing a Case Definition

A case definition is a set of standard criteria for deciding whether a person has
a particular disease or health related condition. It does not make a clinical diagno
sis.
It is the first step in establishing consistent criteria before definitive diagnosis
can be made. It is especially important for unknown diseases. The definition
should be optimal as too broad a definition will yield a large number of false posi
tives and too narrow a definition will yield false negatives.

Figure 14.7 An epidemic curve suggestive of point source

183
 Figure 14.8 Epidemic curve of propagated epidemic

The cases are often classified as definitive, probable and possible. An example
is given in Box 14.3. As investigation proceeds and the hypotheses become fo
cused, the possible definition may be dropped.

Figure 14.9 Dengue casein Subang Jaya, Malaysia 2002

Source: Malaysian Centre for Remote Sensing (MACRES)

Most of the dengue incidence cases occurred in residential areas, construction sites influence the number of
dengue incidences and the distribution outbreak areas.

Low incidence was reported in the industrial and commercial sector.

184
 5. Identifying and Counting Cases
Cases need to be identified using as many sources as possible. The following
information should be collected and classified as in the step above. Initial efforts
should be directed at places such as hospitals and health centres where diagnosis
is likely to be made.
‣Identifying information - name, address and telephone number.
‣Demographic information - age, sex, race and occupation.
‣Clinical information.
‣Risk factor information - tailor risk factor information to the problem.
‣Reporter information - identify the person who provided the case report.

6. Perform Descriptive Epidemiology

Once data is collected a description of the outbreak by time place and person is
done. This may be repeated several times during the course of the investigation.
Time: Plot an epidemic curve, which is a graph of the time on the x-axis and
number of cases on the y-axis. This curve provides several important clues about
what is happening in an outbreak and helps identify the type of exposure and the
route of spread.
An epidemic curve with a steep slope and a more gradual down slope indicates
a point source epidemic in which persons are exposed to the same source over a
relative brief period (Figure 14.7). Any sudden rise in number of cases suggests
sudden exposure to a common source.
If the duration of exposure is prolonged, the epidemic is called a continuous
common source and the epidemic curve will have a plateau. Person-to-person
spread (propagated epidemic) should have a series of progressively taller peaks
one incubation period apart (Figure 14.8).
The initial case in the outbreak is called the index case or the one which intro
duced the organism into the population. It is important to identify the index case.

185
 Place: the geographic extent of the problem is described. The cases are identi
fied on a map as confirmed and presumptive cases. Other ecological data of inter
est such as water supply points, wind currents or other exposures are added in.
This is called a spot map and helps identify clustering and relationship with eco
logical factors (Figure 14.9).

Figure 14.10 Phases for pandemic alert for Influenza

186
 Box 14.4 Current WHO phases of pandemic alert for influenza
WHO has a six-phased approach for easy incorporation of new
recommendations and approaches into existing national preparedness and
response plans for epidemic disease especially Influenza. The grouping and
description of pandemic phases have been revised to make them easier to
understand, more precise, and based upon observable phenomena. Phases 1–3
correlate with preparedness, including capacity development and response
planning activities, while Phases 4–6 clearly signal the need for response and
mitigation efforts. Furthermore, periods after the first pandemic wave are
elaborated to facilitate post pandemic recovery activities.

During the 2009 Pandemic influenza World Health Organization classified the
pandemic into 6 phases

In nature, influenza viruses circulate continuously among animals, especially

birds. In Phase 1 no viruses circulating among animals have been reported to
cause infections in humans.

In Phase 2 an animal influenza virus circulating among domesticated or wild

animals is known to have caused infection in humans, and is therefore considered
a potential pandemic threat.

In Phase 3, an animal or human-animal influenza reassortant virus has

caused sporadic cases or small clusters of disease in people, but no human-to
human transmission sufficient to sustain community-level outbreaks have
occurred.

Phase 4 is characterised by verified human-to-human transmission.

“Community-level outbreaks” can occur. This phase indicates a significant
increase in risk of a pandemic but does not necessarily mean that a pandemic is a
forgone conclusion.

Phase 5 is characterised by human-to-human spread of the virus into at least

two countries in one WHO region.

Phase 6, the pandemic phase, is characterized by community level outbreaks in

at least one other country in a different WHO region in addition to the criteria
defined in Phase 5. Designation of this phase will indicate that a global
pandemic is under way. During the post-peak period, levels in most countries
with adequate surveillance will have dropped below peak observed levels.
Seasonal outbreaks or second wave pandemic occur.

187
 Person: Information on the host characteristic such as age, race and sex or by
exposure, occupation, use of medication, drugs, etc. may help clarify the problem
and its cause. Incidence and attack rates are calculated.

7. Developing a Hypotheses
In reality this step can start with the first notification of an outbreak. However
only after the data collection does it get crystallised. The hypotheses should ad
dress the source of agent, mode of transmission and exposure that caused the dis
ease, and it should be testable.

8. Evaluating Hypothesis
Once the hypothesis is made it is tested. This is commonly done by comparing
rates such as attack rates or by doing a case control study. Laboratory studies are
also important to test hypotheses. If need be, further studies need to be done to re
fine the hypotheses further.

9. Implementing Control and Prevention Measures

This is the primary goal of the investigation. Steps may be initiated as soon as
possible during the course of the investigation. There are four common interven
tions - sanitation, prophylaxis, diagnosis and treatment, and control of disease vec
tors.

10. Communicating the Finding

Filing a written report to the health authorities is important as it enables them
to assist in disease control, coordinate with different agencies if required, make
policy decisions and add to professional and public knowledge. The report should
be in a scientific format of introduction, background, methods, results, discussion
and recommendation.

188
 Table 14.2 List of Notifiable Diseases. Those in Bold Italics need to be notified by telephone within 24 hrs.

CONTROL OF COMMUNICABLE DISEASES IN MALAYASIA

The communicable disease control programme was started in 1971 with the es
tablishment of the epidemiological unit under the health division of the Ministry
of Health and became fully operational in 1996. The general aim was to closely
monitor the epidemiological pattern of common communicable diseases existing
in Malaysia.

Objectives
‣To closely monitor the epidemiological patterns of disease in the country by
locality, time and person; to effect appropriate and timely preventive and con
trol measures so that the mortality and morbidity due to such diseases are de
creased and unnecessary suffering is avoided.

189
 ‣To reduce morbidity and mortality due to communicable diseases so that they
do not pose public health problems.
‣To promote healthy living, healthy and safe working conditions, clean envi
ronment, appropriate preventive measures, early detection and treatment,
regular surveillance and appropriate rehabilitative services.
‣To promote active community participation and inter-sectoral collaboration
towards the development of a caring and healthy society.

Strategies
‣Identification and definition of national concern
‣Conducting applied research studies to clarify the epidemiological pictures,
testing feasibility
‣Programme monitoring, research and evaluation.

Achievement
At present there are 25 infectious diseases required by the Prevention and Con
trol of Infectious Diseases Act 1988 to be notified and kept under surveillance. Of
these, 9 need to be reported by phone or other means (Table 14.2).

NOTIFIABLE COMMUNICABLE DISEASES IN MALAYSIA

Under ACT 342 PREVENTION AND CONTROL OF INFECTIOUS DIS

EASES ACT 1988 Incorporating latest amendment - P.U. (A) 374/2006 all medi
cal practitioners who treat or have information about the existence of any commu
nicable disease in any premises should notify of the existence of these communica
ble diseases to the nearest medical officer without fail and delay. The appropriate
forms should be used under the guidelines of the act.

190
 Summary Box Chapter 14

• Infectious diseases can be transmitted directly or indirectly.

• There are three principles for controlling diseases—control, elimination,
eradication.
• Surveillance is an important activity for controlling disease spread.
• Outbreak investigation is the biggest challenge for an epidemiologist as it
puts all his/her skills as an epidemiologist and a health professional into
practice.
•Several important steps are to be followed while investigating any
epidemic. These actions may run parallelly.

ADDITIONAL READING
http://ricn.on.ca/chainoftransmission/
http://www.who.int/csr/disease/avian_influenza/phase/en/index.html

191
References

1. Beaglehole R., Bonita R., Kjellstrom T. 2006. Basic epidemiology.

A.I.T.B.S:WHO
2. James F. Jekel. David L. Katz. Joann G. Elmore. 1996. Epidemiology, Biostas
tistics and Preventive Medicine. 2nd edition. W. B. Saunders Company.
3. Mausner Judith S. Kramer Shira. Epidemiology -An Introductory Text. 2nd Edi
tion. Philadelphia. W. B. Saunders Company.
4. Park K. Parks Textbook of Preventive and Social Medicine. 17th edition. In
dia: 2000. Mis Banarsidas Bhanot Publishers.
5. Gordis, L. 6. 2004 Epidemiology. Philadelphia: W. B. Saunders Company
6. Farmer, Richard. 1997. Lecture notes on epidemiology and public health medi
cine. 4th edition. Blackwell Science
7. Phoon, W.O., & Chen P. C. Y. 1987. Text book of community medicine in
SEA. Singapore.John Wiley & Sons
8. Sang Tsie Han. Health research methodology. A guide/or training in re-
search methods. WHO Regional Publications - Western Pacific in Action se
ries 5
9. CDC. Principle of epidemiology. An introduction to applied epidemiology
and biostatistics. 2nd edition. Atlanta: U. S. Department of Health and Human
Services
10. Executive Summary. Persidangan Rancangan Malaysia Kelapan

cxcii
11. McKenzie James F. Pinger Robert R. Kotecki Jerome K. 2002. An Introduc
tion to Community Health. 4th Edition. Jones and Bartlett Publishers.
12. Wallace Robert B. Public Health & Preventive medicine. 14th edition. Apple
ton and Lange
13. Turnock Bernard J. Public Health: What it is and how it works’. 2nd Edition.
Jones and Bartlett Publishers.
14. Hennekens, C. H., &Burning 1. E. 1987. Epidemiologv in medicine. Bos
ton: Little, Brown and Company
15. Ministry of Health Malaysia. Laman rasmi bahagian pendidikan kesihatan,
KKM. www.infosihat.gov.my
16. Zenz, C. 1994. Occupational Medicine. 3rd edition. Mosby
17. Novick, L. F. Mays G. 2001. Public health administration. Principles for popu
lation based management. Gait ersburg Maryland. An Aspen Publication
18. Choi, B. C. K., & Noseworthy, A. L. 1992. Classification, direction and pre
vention of bias in epidemiologic research. Journal of occupational medicine.
34(4),265 - 271
19. Malaysia: Achieving the Millennium Development Goals. Successes and Chal
lenges. United Nations Country Team. Malaysia.
20. Chye L. C. Yahya H. Second Report of the National Cancer Registry. Cancer
incidence in Malaysia. National Cancer Registry. MOH
21. M C Gupta. B K Mahajan. 2003. Textbook of Preventive and Social Medicine.
3rd edition. New Delhi. Jaypee Brothers Medical Publishers (P) Ltd.
22. Chirstie D. 1988. A guide to occupational epidemiology. CCH Australia Lim
ited.
23. Olsen J. Merletti F. Snashall D. Vuylsteek K. 1991. Epidemiology in practise.
searching for causes of work related diseases: an introduction to epidemiology
at the work site (part II, pp 46-65). Oxford: Oxford University Press.

cxciii
24. Harper AC. Holman C D’AJ. Dawes VP. 1994. Risks II: Epidemiological
methods of measurement. The health of populations: an introduction (2nd ed.
Ch 5, pp77- 118). Melbourne, Churchill Livingstone.
25. Torrence ME. 1997. Bias and errors. Understanding epidemiology (Ch 6, pp
117- 132). St Louis, MO Mosby.
26. Gordis L. 1996. Case control and cross sectional studies. Epidemiology (Ch 9,
99124-140). Philadephia. WB Saunders Company.

cxciv