VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 OF PRIMARY HEALTH CARE ‘It is not possible to divorce the practice of medicine from the society in which it is practised.’ See page 223 Original Scientific Paper Calculating cardiovascular risk in people with Type 2 diabetes See page 181 Original Scientific Paper Pre-calling babies to improve immunisation timeliness See page 189 Original Scientific Paper PSA testing in asymptomatic males See page 199 Original Scientific Paper Acute otitis media in the under-fives See page 205 Original Scientific Paper Brief mental health intervention for Maori See page 231 Viewpoint The platform model of pain management See page 254 CONTENTS VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 ISSN 1172-6164 (Print) ISSN 1172-6156 (Online) OF PRIMARY HEALTH CARE 178 Editorials 231 From the Editor 178 Guest Editorial 180 Fiona Mathieson, Kara Mihaere, Sunny Collings, Anthony Dowell, James Stanley A practical issue Felicity Goodyear-Smith Maori cultural adaptation of a brief mental health intervention in primary care Short Report 239 Commentary: risk prediction models for people with Type 2 diabetes The anatomical placement of body organs by Australian and New Zealand patients and health professionals in general practice Marjan Kljakovic Kamlesh Khunti Improving Performance 181 Original Scientific Papers 242 Quantitative Research 181 New Zealand Diabetes Cohort Study cardiovascular risk score for people with Type 2 diabetes: validation in the PREDICT cohort Tom Robinson, C Raina Elley, Sue Wells, Elizabeth Robinson, Tim Kenealy, Romana Pylypchuk, Dale Bramley, Bruce Arroll, Sue Crengle, Tania Riddell, Shanthi Ameratunga, Patricia Metcalf, Paul Drury 189 Early connections: effectiveness of a pre-call intervention to improve immunisation coverage and timeliness Felicity Goodyear-Smith, Cameron Grant, Tracey Poole, Helen Petousis-Harris, Nikki Turner, Rafael Perera, Anthony Harnden 199 205 The incidence of acute otitis media in New Zealand children under five years of age in the primary care setting Barry Gribben, Lesley Salkeld, Simon Hoare, Hannah Jones 213 Sally Abel, Bob Marshall, Donny Riki, Tania Luscombe 249 Continuing Professional Development 249 Does the order of presentation and number of online resources affect the frequency of access by learners? Steven Lillis, Samantha Murton 249 String of PEARLS about musculoskeletal conditions 250 Vaikoloa: Hearing loss among Pacific peoples Ofa Dewes 251 Addressing patient alcohol use: a view from general practice Thomas Mules; Jennifer Taylor; Rachel Price; Logan Walker; Baneet Singh; Patrick Newsam; Thenmoli Palaniyappan; Toby Snook; Mahfuzah Ruselan; John Ryan; Jaishree Santhirasegaran; Phoebe Shearman; Petronella Watson; Richard Zino; Louise Signal; Geoff Fougere; Helen Moriarty; Gabrielle Jenkin Factors influencing diagnostic decision-making Kathleen Callaghan Nuggets of Knowledge: Statins and memory loss Linda Bryant 254 Viewpoint 254 From ladder to platform: a new concept for pain management Lawrence Leung 259 Frequently asked questions on measurement of bone mineral densitometry Joseph Lee, Nelson Loh 262 Letters to the Editor 263 Film Review 263 Mixed Method Research 223 Potion or Poison? Colloidal silver David Woods 253 Qualitative Research 217 Cochrane Corner: Topical antibiotics are probably better than placebo for acute conjunctivitis but most get better anyway Bruce Arroll PSA testing in general practice Fraser Hodgson, Zuzana Obertová, Charis Brown, Ross Lawrenson Evaluation of Tu Meke PHO’s Wairua Tangata Programme: a primary mental health initiative for underserved communities A good death: a film about end-of-life care and advance care planning—produced by Prof. D Robin Taylor and Paul Trotman Reviewed by Prof. Rod MacLeod 264 About the Journal of Primary Health Care VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 177 EDITORIALS FROM THE EDITOR A practical issue Felicity GoodyearSmith MBChB, MD, FRNZCGP, Editor J PRIM HEALTH CARE 2012;4(3):178–179. CORRESPONDENCE TO: Felicity Goodyear-Smith Professor and Goodfellow Postgraduate Chair, Department of General Practice and Primary Health Care, The University of Auckland, PB 92019 Auckland, New Zealand f.goodyear-smith@ auckland.ac.nz 178 T his issue has a wide range of research on a variety of topics, but the common thread is that these studies address practical issues relevant to New Zealand primary health care. The lead paper by Robinson et al. reports on the locally developed PREDICT-based risk model which is more accurate in predicting cardiovascular risk than the adjusted Framingham equation.1 In his accompanying commentary, Dr Kamlesh Khunti, Professor of Primary Care Diabetes and Vascular Medicine at the University of Leicester, points to the need for studies on the impact of use of risk prediction models with outcomes such as patient adherence to medication or increased understanding before the use of these models becomes routine practice.2 A study of routine pre-call of infants at four weeks to alert parents to the need to present their babies to general practice at six weeks to start their immunisation schedule has found that this increases the coverage and timeliness of the immunisation series.3 However, the most significant finding of this study is the importance of enrolment of newborn babies with a general practice. Making an early connection with a practice means that an infant is much more likely to be immunised fully and on time, as well as the other likely health care gains that may result from engagement with their general practitioner (GP) or practice nurse. Controversy around prostate-specific antigen (PSA) screening continues. A study of GPs in the Waikato found a high likelihood of them PSA testing asymptomatic men including those aged 70 years or older, despite there being no evidence of benefit in this activity.4 A qualitative study suggests that, in Wellington at least, GPs may not be screening and intervening sufficiently with patients around alcohol misuse.5 While some GPs may be ignoring Ministry of Health recommendations regarding prostate and alcohol screening, it appears that best practice guidance on management of acute otitis media in children is being followed. A cohort study of nearly 20 000 New Zealand children shows a significant decline in the use of antibiotics in treatment, in concordance with accepted best practice.6 There are two studies on primary mental health initiatives for Maori and other underserved populations. Mathieson and colleagues report on a Maori adaptation of a brief intervention involving cognitive behavioural therapy and guided self-management,7 and a research team in Hawkes Bay describe an integrated, holistic tikanga Maori–based programme targeting Maori, Pacific and quintile 5 populations aimed at reducing mental health inequalities.8 Callaghan explores factors that might influence GPs’ decision making and finds that clinical information and the probability of disease are rated as highly important and desirable by ‘standard setters’ in general practice academic departments and The Royal New Zealand College of General Practitioners.9 Lillis and Murton explore the provision of online resources to GP registrars in training and conclude that they are only likely to access the top four in the list, underlining the importance of prioritising and possibly limiting provided resources.10 A study by Kljakovic finds that patients from both Australian and New Zealand general practices performed poorly in correctly locating body organs in line drawings, and health professionals achieved this only moderately better than their patients.11 On an even more practical bent, two viewpoint papers offer specific clinical tips. Leung extends the concept of the World Health Organization analgesic ladder to that of a platform, providing a broad range of pain relief interventions in a VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE EDITORIALS FROM THE EDITOR stepped-up manner,12 and Lee and Loh review the facts around bone mineral densitometry.13 Our regular columns provide practical tips on the use of topical antibiotics in acute conjunctivitis (Cochrane Corner), best practice evidence for managing a number of musculoskeletal conditions (String of PEARLS), the potential harm and lack of evidence of benefit for use of colloidal silver (Potion or Poison?), strategies to identify cognitive impairment from statin use (Nuggets of Knowledge) and possible actions to address hearing loss among Pacific peoples (Vaikoloa).14 Callaghan writes, ‘it is not possible to divorce the practice of medicine from the society in which it is practised’.9 The papers in this issue are about primary health care practice in our New Zealand communities with our own patient populations. No longer do we need to rely on international research conducted in secondary care settings to inform our practice—the breadth and depth of New Zealand primary care research means our discipline has come of age. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. intervention to improve immunisation coverage and timeliness. J Prim Health Care. 2012;4(3):189–198. Hodgson F, Obertová Z, Brown C, Lawrenson R. PSA testing in general practice. J Prim Health Care. 2012;4(3):199–205. Mules T, Taylor J, Price R, Walker L, Singh B, Newsam P, et al. Addressing patient alcohol use: a view from general practice. J Prim Health Care. 2012;4(3):217–222. Gribben B, Salkeld L, Hoare S, Jones H. The incidence of acute otitis media in New Zealand children under five years of age in the primary care setting. J Prim Health Care. 2012;4(3):205–212. Mathieson F, Mihaere K, Collings S, Dowell A, Stanley J. Maori cultural adaptation of a brief mental health intervention in primary care. J Prim Health Care. 2012;4(3):231–238. Abel S, Marshall B, Rikki D, Luscombe T. Evaluation of Tu Meke PHO’s Wairua Tangata Programme: a primary mental health initiative for underserved communities. J Prim Health Care. 2012;4(3):242–248. Callaghan K. Factors influencing diagnostic decision-making. J Prim Health Care. 2012;4(3):223–230. Lillis S, Murton S. Does the order of presentation and number of online resources affect the frequency of access by learners? J Prim Health Care. 2012;4(3):213–216. Kljakovic M. The anatomical placement of body organs by Australian and New Zealand patients and health professionals in general practice. J Prim Health Care. 2012;4(3):239–241. Leung L. From ladder to platform: a new concept for pain management. J Prim Health Care. 2012;4(3):254–258. Lee J, Loh N. Frequently asked questions on measurement of bone mineral densitometry. J Prim Health Care. 2012;4(3):259–261. Dewes O. Hearing loss among Pacific peoples. J Prim Health Care. 2012;4(3):250–251. References 1. Robinson T, Elley R, Wells S, Robinson E, Kenealy T, Pylypchuk R, et al. New Zealand Diabetes Cohort Study cardiovascular risk score for people with type 2 diabetes: validation In the PREDICT COHORT. J Prim Health Care. 2012;4(3):181–188. 2. Khunti K. Commentary: risk prediction models for people with Type 2 diabetes. J Prim Health Care. 2012;4(3):180. 3. Goodyear-Smith F, Grant C, Poole T, Petousis-Harris H, Turner N, Perera R, et al. Early connections: effectiveness of a pre-call A tribute to Professor Marjan Kljakovic Just as this issue is going to press we have received the very sad news that Marjan, the author of one of the papers in this issue (Kljakovic M. The anatomical placement of body organs by Australian and New Zealand patients and health professionals in general practice. J Prim Health Care. 2012;4(3):239–241), died today, having suffered a major myocardial infarction on 29 July. A New Zealand academic general practitioner, Marjan relocated across the ditch seven years ago as Professor at the Academic Unit of General Practice and Community Health, Australian National University Medical School, Canberra. However, he remained a Kiwi at heart. Marjan was highly regarded as an inspirational thinker and an innovative teacher who contributed extensively to philosophical debate in the general practice arena. He will be sadly missed. Editor, 14 August 2012 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 179 EDITORIALS GUEST EDITORIAL Commentary: risk prediction models for people with Type 2 diabetes Kamlesh Khunti PhD, MD, FRCGP, FRCP Department of Health Sciences, University of Leicester, United Kingdom J PRIM HEALTH CARE 2012;4(3):180. CORRESPONDENCE TO: Kamlesh Khunti Professor of Primary Care Diabetes and Vascular Medicine, Department of Health Sciences, University of Leicester, Gwendolen Road, Leicester LE5 4PW, UK kk22@le.ac.uk 180 P eople with Type 2 diabetes are at increased risk of cardiovascular disease, the determinants of which are multifactorial.1 A number of international guidelines recommend calculating future cardiovascular disease risk for management of patients with Type 2 diabetes. There has been a plethora of cardiovascular disease risk prediction models for Type 2 diabetes and a recent systematic review identified 45 prediction models, of which 12 were developed for patients with Type 2 diabetes.2 Less than onethird of these were externally validated in a diabetes population and overall the discriminative value for most prediction models was moderate.2 Another systematic review confirmed limited evidence of impact on patient management and outcomes with the use of prediction models.3 References 1. Sarwar N, Gao P, Seshasai SR, Gobin R, Kaptoge S. Diabetes mellitus, fasting blood glucose concentration, and risk of vascular disease: a collaborative meta-analysis of 102 prospective studies. Lancet. 2010;375:2215–22. 2. van Dieren S, Beulens JWJ, Kengne AP, Peelen LM. Prediction models for the risk of cardiovascular disease in patients with Type 2 diabetes: a systematic review. Heart. 2012;98:360–9. 3. Willis A, Davies MJ, Yates T, Khunti K. Primary prevention of cardiovascular disease using validated risk scores: a systematic review. Journal of the Royal Society of Medicine. 2012. 4. Robinson T, Elley C, Wells S, et al. New Zealand Diabetes Cohort Study cardiovascular risk score for people with Type 2 diabetes: validation in the PREDICT cohort. J Prim Health Care. 2012;4(3): 181–189. In this issue of the journal, Robinson and colleagues have conducted a validation study of the Diabetes Cohort Study (DCS) CVD Risk Predictive model in people with Type 2 diabetes in New Zealand.4 The strengths of this study are the large numbers of people included, the long follow-up with 12.8% of people having a cardiovascular outcome, and the validation being conducted in a population in which the score was derived. The study found that the DCS model had marginally better discrimination than the currently used New Zealand Framingham risk equation. Overall the discriminative value was moderate. Many of the previous new scores have been compared with the well-established United Kingdom Prospective Diabetes Study risk score and, therefore, this is one limitation of this study. In addition, robust evidence on the impact of use of risk prediction models on patient outcomes in terms such as adherence to medications, patient understanding or improvements in harder outcomes is lacking. Until such evidence is available, the use of risk prediction models in routine clinical practice will not be adopted. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH New Zealand Diabetes Cohort Study cardiovascular risk score for people with Type 2 diabetes: validation in the PREDICT cohort Tom Robinson MBChB, MPH, FNZCPHM, FRNZCGP;1,3 C Raina Elley MBChB, FRNZCGP, PhD;1 Sue Wells MBChB, FRNZCGP, FNZCPHM, PhD;2 Elizabeth Robinson MSc;2 Tim Kenealy MBChB, FRNZCGP, PhD;1 Romana Pylypchuk MPH, MSc;2 Dale Bramley MBChB, MPH, FNZCPHM;3 Bruce Arroll MBChB, FRNZCGP, PhD;1 Sue Crengle MBChB, FNZCPHM, FRNZCGP, PhD;4 Tania Riddell MBChB, MPH, FNZCPHM;2 Shanthi Ameratunga MBChB, FRACP, FNZCPHM, PhD;2 Patricia Metcalf PhD;5 Paul L Drury MA, MB BChir, FRCP, FRACP6 1 Department of General Practice and Primary Health Care, The University of Auckland, Auckland, New Zealand 2 Section of Epidemiology and Biostatistics, The University of Auckland 3 Waitemata District Health Board, Auckland ABSTRACT INTRODUCTION: New Zealand (NZ) guidelines recommend treating people for cardiovascular disease (CVD) risk on the basis of five-year absolute risk using a NZ adaptation of the Framingham risk equation. A diabetes-specific Diabetes Cohort Study (DCS) CVD predictive risk model has been developed and validated using NZ Get Checked data. AIM: To revalidate the DCS model with an independent cohort of people routinely assessed using PREDICT, a web-based CVD risk assessment and management programme. METHODS: People with Type 2 diabetes without pre-existing CVD were identified amongst people who had a PREDICT risk assessment between 2002 and 2005. From this group we identified those with sufficient data to allow estimation of CVD risk with the DCS models. We compared the DCS models with the NZ Framingham risk equation in terms of discrimination, calibration, and reclassification implications. 4 Te Kupenga Hauora Maori, The University of Auckland 5 Department of Statistics, The University of Auckland 6 Auckland Diabetes Centre, Auckland District Health Board RESULTS: Of 3044 people in our study cohort, 1829 people had complete data and therefore had CVD risks calculated. Of this group, 12.8% (235) had a cardiovascular event during the five-year follow-up. The DCS models had better discrimination than the currently used equation, with C-statistics being 0.68 for the two DCS models and 0.65 for the NZ Framingham model. DISCUSSION: The DCS models were superior to the NZ Framingham equation at discriminating people with diabetes who will have a cardiovascular event. The adoption of a DCS model would lead to a small increase in the number of people with diabetes who are treated with medication, but potentially more CVD events would be avoided. KEYWORDS: Cardiovascular disease; diabetes; prevention; risk assessment; reliability and validity Introduction Globally there is an epidemic of Type 2 diabetes.1,2 It was estimated that in 2010 there were over 195 000 people in New Zealand (NZ) with diabetes—5.6% of the adult population.3 People with diabetes are at increased risk of dying of cardiovascular disease (CVD) which accounts for almost 50% of all deaths amongst people with diabetes.4,5 There is considerable evidence that energetic management of risk factors such as blood pressure, dyslipidaemia, and glycaemia reduces the risk of CVD in people with diabetes.6–11 However, it is accepted that rather than treating risk factors separately, clinicians should use absolute CVD risk to guide patient management.12,13 NZ guidelines for cardiovascular risk assessment use a predictive risk equation adapted from VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):181–188. CORRESPONDENCE TO: Tom Robinson Department of General Practice and Primary Health Care, The University of Auckland, PB 92019 Auckland, New Zealand t.robinson@auckland.ac.nz 181 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH the United States Framingham Heart Study.14,15 This equation has a number of disadvantages for predicting risk of CVD amongst people with diabetes in NZ. In particular, the Framingham cohort was from the United States, did not include ethnic groups that are important in NZ, and only included a small number of people with diabetes.16 In addition, the equation does not include a number of diabetes-specific variables— such as duration of diabetes, glycaemic control, and albuminuria—that are predictive of cardiovascular outcomes.17–20 The NZ adaptation of the Framingham model does include adding a single additional five-year risk of 5% for these factors.14 In 2010, Elley et al. reported two predictive CVD equations based on the New Zealand Diabetes Cohort Study (DCS). This was a prospective open cohort that used data from a national primary care diabetes programme (Diabetes Get Checked), which commenced in 2000.21 Full details of the derivation and validation of the equation are described in the original article.21 Briefly, data from 36 127 people with Type 2 diabetes, but without pre-existing CVD, were matched to national hospitalisation and mortality databases. Predictor variables for the first equation (DCS-A) included age at diagnosis, duration of diabetes, sex, ethnicity, smoking status, systolic blood pressure, HbA1c, total cholesterol: HDL cholesterol ratio (TC/HDL), and the presence of microor macroalbuminuria. A second equation (DCS-B) also included current antihypertensive treatment. The performance of both equations was tested on 10 030 individuals from a different geographic area in NZ with discrimination and calibration superior to the original Framingham equation.21 Before using a prognostic model in clinical practice it is important to validate it using data from other independent populations of patients.22 This study aimed to validate the DCS models using data from a cohort of people routinely assessed in NZ general practice with PREDICT, a CVD risk assessment and management programme. Methods Design This validation study uses data from primary care to assess the discrimination, calibration and 182 reclassification implications of the DCS equations in predicting CVD events, compared with actual events over five years. Study population PREDICT is a web-based, real-time decision support programme that has been integrated with most practice management software in use in NZ primary care.23 General practitioners and practice nurses enter required clinical data to create a risk profile. This profile is sent by a secure internet connection to a central server that returns the patient’s NZ Framingham five-year cardiovascular risk score with management recommendations. At the same time, an electronic profile is stored and linked to an encrypted National Health Index (NHI) number. These were anonymously linked to national hospitalisation, pharmaceutical dispensing and mortality outcomes and also to regional laboratory data. Individuals identified as having Type 2 diabetes and no known pre-existing CVD with a PREDICT assessment between 27 August 2002 and 31 December 2005 were included. Individuals were said to have diabetes if they were identified by their primary care physician as having diabetes at first risk assessment, or if they had been identified as having diabetes in the national hospitalisation database, or had been prescribed insulin or an oral hypoglycaemic agent prior to or on their first PREDICT assessment date. If the type of diabetes was unclear, we assigned them as Type 2 if they were never on insulin, if they had been on an oral hypoglycaemic agent, or if their age of onset was over 30 years in Maori and Pacific or over 50 in other ethnic groups. Preexisting CVD was identified from the primary care physician’s risk assessment record. Risk variables Risk factor variables required for the DCS equations were extracted for each individual. Data on some of the variables were missing from early PREDICT risk assessments. Duration of diabetes was included if it could be calculated from any subsequent PREDICT risk assessment record. Missing laboratory data were obtained from laboratory records where results from up to five years prior to the baseline assessment or two VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH weeks after the assessment were available. After the addition of these data, only individuals with a complete minimum dataset were included in the final study cohort. Ethnicity was derived from both the primary care practitioner records and the encrypted NHI database and was prioritised in the order: Maori, Pacific, South Asian, East Asian, ‘Other’ and European. Outcome measures Primary care data were linked to national hospitalisation and mortality databases by NHI number to identify all CVD events over the five years following baseline for each individual. CVD events included hospital admission or death from ischaemic heart disease, cerebrovascular disease or peripheral vascular disease. These were identified from national hospital and mortality databases coded according to ICD-9 and ICD-10 (see the appendix in the web version of this paper).21 Five-year risk was calculated for each individual according to both the NZ Framingham and the DCS equations.21 Analyses We compared predicted risk with observed outcomes. To assess discrimination, the ability of the models to distinguish between individuals who do or do not have a subsequent CVD event, we calculated the area under the receiver operating characteristic (ROC) curve (C statistic).13,22,24 Calibration was assessed by comparing the observed and predicted probabilities of CVD events in the pre-specified deciles of DCS model risk, and performing a Hosmer–Lemeshow test for equivalence. The effect of reclassification of risk from the NZ Framingham model to the DCS models was measured using a 15% five-year cardiovascular risk threshold. NZ guidelines recommend drug treatment with five-year risks above 15%. A scatter plot of risks predicted by the two models with these pre-determined risk categories was also produced.24 All analyses were undertaken using Stata® 11.2. Ethical approval This validation study was approved by the Multiregion Ethics Committee (WGT/04/09/077) as part of the Diabetes Cohort Study. The PREDICT cohort study and research process was ap- WHAT GAP THIS FILLS What we already know: People with Type 2 diabetes are at high risk of a cardiovascular event. A locally derived Diabetes Cohort Study CVD risk equation—http://www.nzssd.org.nz/cvd/—has been found to be more valid for those with diabetes in New Zealand than the currently used Framingham equation. What this study adds: Before incorporating the new equation into national recommendations for management, further validation was required using an independent cohort. The Diabetes Cohort Study CVD risk equation predicted risk more accurately than the currently used adjusted Framingham equation among people with diabetes in the New Zealand PREDICT cohort. proved by the Northern Region Ethics Committee Y in 2003 (AKY /03/12/314) with subsequent annual approval by the National Multi-region Ethics Committee since 2007 (MEC/07/19/EXP). Results Study population The derivation of the study cohort and subsequent CVD events is shown in Figure 1. We classified 3044 (13.3%) people on the database as people with Type 2 diabetes without pre-existing CVD. Of these, 1829 (60.1%) had the minimum dataset of risk variables required and formed the final cohort for the study. About two-thirds of these individuals (65.9%) were included after having data added from sources other than the first risk assessment record. These data were diabetesspecific variables (HbA1c, urinary albumin/creatinine ratio, diabetes duration, and age of onset of diabetes). All individuals were followed for five years from their initial CVD risk assessment. During that time, 235 had first CVD events (12.8%), in which 45 (2.5%) were fatal and 190 (10.4%) were non-fatal events. Baseline characteristics of participants are presented and compared with those of the 1215 people excluded due to missing variables in Table 1. Compared with those included, a higher proportion of the excluded group were European and a lower proportion Pacific. Excluded participants had slightly higher systolic blood pressures and TC/HDL ratios and were less likely to be recorded as being past smokers. Although the two groups had similar risks of CVD events VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 183 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Discrimination Figure 1. Flow diagram of participants through study 24 446 people with first PREDICT risk assessment between 27 August 2002 and 31 December 2005 851 people with Type 2 diabetes and pre-existing CVD 20 551 people with Type 2 diabetes 3044 people with identified Type 2 diabetes 1215 people without minimum dataset 1829 (60.1%) people with minimum* datasets The ROC curves for the DCS-B equation (with antihypertensive status included) and the modified NZ Framingham equation are shown in Figure 2. The area under the ROC curve (C statistic) was 0.678 (95% CI 0.642–0.714) for the DCS-A equation (without antihypertensive status included) and 0.684 (95% CI 0.648–0.720) for the DCSB equation. The C statistic for the NZ Framingham equation in this study was 0.648 (95% CI 0.612–0.684) and the unadjusted Framingham equation was 0.649 (95% CI 0.613–0.685). Both DCS equations had significantly higher C statistics than the NZ Framingham equation (p=0.04 DCS-A equation and p=0.01 DCS-B equation). The DCS-B equation C statistic was also significantly higher than the DCS-A equation (p=0.04). The C statistics for the DCS equations in this study were similar to those found in the original DCS validation study (0.69).21 34.1% complete data from first PREDICT assessment 65.9% duration of diabetes derived from subsequent assessments Calibration 65.4% HbA1c from urinary ACRs variables derived from other laboratory data 0.5% derived from subsequent PREDICT assessments within 1 year 26 died from non-CVD causes 235 (12.8%) people had first CVD event within five years of date of their first PREDICT risk assessment (45 fatal and 190 non-fatal) Reclassification * Minimum dataset = age at diagnosis, gender, duration of diabetes, smoking status, systolic blood pressure, HbA1c, fasting serum total cholesterol and HDL, urine albumin-creatinine ratio, ethnicity, antihypertensive medication according to the NZ Framingham equation and observed events during follow-up, the excluded group had higher fatal CVD and other-cause mortality. Differences in duration, HbA1c, and albuminuria between the ‘Included’ and ‘Excluded’ groups may not reflect a true underlying difference, as information on these variables was missing for many in the ‘Excluded’ group. 184 Figure 3 compares the mean predicted risk with the mean observed five-year event rate for each decile of predicted risk for DCS-B and NZ Framingham equations. The DCS equations predicted higher risks than the NZ Framingham equation for people in the higher deciles of risk. The Hosmer–Lemeshow test showed that estimated risks based on the baseline risk profile tended to be higher than the real event rate for all equations (p<0.001 for DCS-A, p=0.001 for DCS-B, and p=0.02 for NZ Framingham). The effect of reclassification of risk from the NZ Framingham model to the DCS models was measured using a 15% five-year cardiovascular risk threshold. Figure 4 plots the predicted five-year risk of a CVD event for each individual using the DCS-B and NZ Framingham equations. Horizontal and vertical lines represent the 15% five-year risk cut-offs above which drug therapy is usually recommended in NZ. Area B in the graph represents individuals that are classified as being at low risk under the DCS equation, but high risk under the NZ Framingham equation. Area D on the graph represents people who were classified as VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Table 1. Characteristics at baseline for study cohort compared with those excluded due to missing variables Characteristic Included Excluded Mean (SD) or Median (IQR)* or n (%) Data available, n (%) Mean (SD) or Median (IQR)* or n (%) 1829 1215 Age (years) 57.3 (10.7) 1215 (100) 57.0 (12.5) Age at diagnosis (years) 51.7 (11.1) 360 (30) 53.4 (11.3) 0.01 4 (2-8) 360 (30) 1 (0-4) <0.001 N Diabetes duration (years)* 0.6 1215 (100) Gender 926 (50.6%) Men 668 (55.0%) 1215 (100) Ethnicity 550 (30.1%) 575 (47.3%) Maori 245 (13.4%) 155 (12.8%) Pacific 756 (41.8%) 296 (24.4%) South Asian 162 (8.9%) 93 (7.7%) East Asian 82 (4.5%) 80 (6.6%) Other 25 (1.4%) Total cholesterol:HDL ratio HbA1c (%)* 0.02 <0.001 European Systolic blood pressure 16 (1.3%) 134.8 (15.6) 1215 (100) 137.8 (19.0) <0.001 4.1 (1.2) 1214 (100) 4.3 (1.3) <0.001 7.4 (6.5-8.6) 803 (66) 6.8 (6.2-7.9) <0.001 1215 (100) Smoking status Current smoker 342 (18.7%) Previous smoker 386 (21.1%) <0.001 211 (17.4%) 95 (7.8%) 458 (38) Albuminuria Microalbuminuria 301 (16.5%) Macroalbuminuria 88 (4.8%) <0.001 100 (21.8%) 36 (7.9%) 1215 (100) Taking medications Antihypertensives 213 (11.7%) 186 (15.3%) 0.003 Lipid-lowering meds 185 (10.1%) 143 (11.8%) 0.1 15.0 (10.1-20.7) 0.6 † 5-year CVD risk * 14.8 (10.6-19.3) 1213 (100) 1215 (100) Outcomes * † P value Any CVD event 235 (12.8%) 154 (12.7%) 0.9 Fatal CVD event 45 (2.5%) 70 (5.8%) <0.001 Non-CVD deaths 26 (1.4%) 45 (3.7%) <0.001 Median and inter-quartile range (IQR) given as distribution is skewed. NZ-adjusted Framingham equation high risk under the DCS equation, but low risk under the NZ Framingham equation. Areas A and C represent people who are given the same classification by both models. There are more people who subsequently have CVD events in area D than in area B, indicating that, although there were both successes and failures in reclassification, sensitivity for the DCS equation was superior. However, there are also more people who did not have events who were reclassified to high risk. Moving from the NZ Framingham equation to the DCS-B equation improves sensitivity from 63.8% to 77.0% (Table 2), but decreases specificity from 54.0% to 51.4%. DCS-A behaves similarly. The positive predictive values (PPVs) and the negative predictive values (NPVs) are higher for the two DCS equations than the NZ Framingham equation. These results are based upon our study cohort five-year prevalence of CVD events of 12.8% (95% CI, 11.3–14.4%). Sensitivity analysis around the 95% confidence intervals of CVD VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 185 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Figure 2. Receiver operating curves for the DCS-B equation and the NZ Framingham CVD equation (five-year risk) 1.00 care in NZ. The latter model in this cohort had an area under the ROC curve of 0.65 while the DCS equation gave an area of 0.68. Strengths and limitations Sensitivity 0.75 The use of diabetes duration, HbA1c levels, and degrees of albuminuria are likely to contribute to better discrimination, particularly for higher risk groups. In addition, the DCS equation was developed in a cohort of much greater relevance to the NZ diabetes population than the Framingham equation, even though adjustments were made by the NZ guidelines.14 0.50 0.25 0.00 0.00 0.25 0.50 0.75 1.00 1-Specificity DCS with antihypertensives NZ Framingham Figure 3. Five-year risks observed in the study cohort compared with those predicted by the DCS-B equation and the NZ Framingham equation for different deciles of risk event prevalence does not change the relative performance of the models (results not shown). Discussion Main finding For this cohort of people with Type 2 diabetes, the DCS equations were better at discriminating risk of a first CVD event than the NZ Framingham equation currently being used in primary 186 There are a number of important methodological limitations in this validation study. One is that these equations were developed and validated on a cohort of people who may have being treated for their CVD risk. Indeed, the PREDICT programme, from which this cohort is gathered, is specifically intended to lower risk of CVD events by encouraging appropriate treatment. Historically, these equations have been used to predict the outcomes for patients if they were not treated. It is therefore likely that the risk of CVD events in this scenario would be higher than the observed risk in this cohort. All three equations overestimated risk when compared to actual events in the cohort. This was in contrast to the initial validation study, where the DCS equations underestimated risk of first CVD event.21 The first possibility is the effectiveness of the PREDICT programme in reducing risk, as discussed above. A second possibility is that the PREDICT cohort is at lower risk than the original DCS cohort, possibly because people with higher risk and comorbidities were enrolled earlier into the Get Checked programme. It is also possible that the difference is due to random error due to the relatively small number of observed events in this study. The overall five-year CVD incidence estimation is somewhat imprecise (12.8%, 95% CI 11.3–14.4%). A further limitation is that an equation should ideally be validated on a population that is representative of the population on which it will eventually be used. It is a strength of this study that the PREDICT cohort is much more repre- VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH sentative of NZ people with diabetes than the original Framingham equation cohort. Nevertheless, the baseline characteristics table suggested that this cohort was not completely representative of NZ people with diabetes. For example, Pacific people were over-represented compared to the overall diabetes population. It is also possible that our inability to obtain data on all patients in the original cohort has introduced a selection bias to our final validation cohort. As discussed, included subjects may have been at a slightly lower risk than those excluded. However, it is important that equations discriminate across diverse as well as representative populations. In validating any equation, it is important to have accurate data on the cohort being studied. The PREDICT data are collected by primary care physicians in routine care rather than by researchers and therefore may contain some inaccuracies. As previously mentioned, some data were not available from the baseline risk assessment and had to be obtained from either laboratory data or subsequent risk assessment records. It is possible that other data, such as smoking status or diabetes duration, may be inaccurately recorded. Such deficiencies would lead to poorer discrimination of the equations. The degree to which they might also lead to systematic under- or overestimation of the predicted risks is more difficult to judge. However, these limitations also reflect the ‘real-life’ clinical situation in which these tools are used. Implications for clinical practice Whilst the current NZ guidelines for managing cardiovascular risk provide guidance for management across a broad range of risks, an important decision point is when to begin treatment of risk with medications.14 At a population level, changing from the current NZ Framingham equation to one of the DCS equations would have a substantial impact on the way diabetes is managed in NZ. If this cohort were representative of people with diabetes in NZ, then changing to the DCS-B equation would result in recommending treating 53% of people with diabetes with lipidlowering and antihypertensive medication instead of 49%. Using the Ministry of Health estimates of the number of people with Type 2 diabetes in NZ, and excluding the estimated 22% of people with diabetes who have pre-existing CVD, then 6503 extra people should be offered treatment. However, the benefit would be that an additional 2587 people, who would have had a first CVD event over the next five years, would be correctly identified and offered preventive therapy. As researchers interested in the population manage- Figure 4. Scatter plot of five-year risk of CVD events predicted by the DCS-B equation against the NZ Framingham equation with 15% five-year risk cut-offs shown Chart area A B C D NZ Framingham classification Low High High Low DCS classification with antihypertensives Low Low High High 634 185 549 226 1594 40% 12% 34% 14% 100% 43 11 139 42 235 18% 5% 59% 18% 100% Total Outcome No event Proportion CVD event Proportion All patients Proportion 677 196 688 268 1829 37% 11% 38% 15% 100% Table 2. Sensitivity, specificity, positive predictive value, and negative predictive value for predicted vs actual CVD events using the DCS and Framingham equations with a 15% risk cut-off (12.9% event prevalence) DCS-A DCS-B Sensitivity 79.6% 77.0% 63.8% Specificity 45.3% 51.4% 54.0% Positive predictive value 17.7% 18.9% 17.0% Negative predictive value 93.8% 93.8% 91.0% VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE NZ Framingham 187 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH ment of diabetes, we believe this would be a very worthwhile trade-off. At an individual level, the person with diabetes or their clinician needs to know, firstly, that on average over a five-year period, 13 people out of 100 will have a CVD event. Secondly, of these 13 people, the DCS equation will correctly identify 10 of these people so they can be offered treatment, whereas the current NZ Framingham equation will successfully predict eight. However, to achieve this, 53 of the 100 people will need to take medication (or five more than currently would be the case). Since it is impractical to offer individual patients the choice of model, we believe it is important to consult both clinicians and patient representatives as to which model is preferable. Should the DCS equation be approved, then it can be seamlessly added to the PREDICT decision support engine and available at the point of care. ACKNOWLEDGEMENTS We would like to acknowledge the patients, practices and primary care organisations who contributed data both to the Diabetes Cohort Study and the PREDICT study, and Diagnostic MedLab Ltd who provided anonymous matched laboratory data. FUNDING We would like to acknowledge the NZ Health Research Council and Auckland Medical Research Foundation for supporting the collection of data for the original cohort studies and for supporting this validation study. One author (T Riddell) is the recipient of a National Heart Foundation fellowship. COMPETING INTERESTS None declared. 188 In conclusion, we have validated the previously developed DCS equations using a different cohort of people with diabetes from the PREDICT dataset and compared its performance with the currently used NZ Framingham equation. We have shown that it has advantages over the current equation at a technical and population health level. From a patient perspective, it will lead to an increased chance of treatment, but also the opportunity to prevent more first CVD events. References 1. Danaei G, Finucane MM, Lu Y, Singh GM, Cowan MJ, Paciorek CJ, et al. National, regional, and global trends in fasting plasma glucose and diabetes prevalence since 1980: systematic analysis of health examination surveys and epidemiological studies with 370 country-years and 2.7 million participants. Lancet. 2011;378(9785):31–40. 2. Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care. 2004;27(5):1047–53. 3. Wright C. Estimated prevalence of Type 2 Diabetes in New Zealand. In: Robinson T, editor. Wellington: Ministry of Health; 2011. 4. Seshasai SR, Kaptoge S, Thompson A, Di Angelantonio E, Gao P, Sarwar N, et al. Diabetes mellitus, fasting glucose, and risk of cause-specific death. N Engl J Med. 2011;364(9):829–41. 5. Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality from coronary heart disease in subjects with Type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med. 1998;339(4):229–34. 6. SIGN. Management of diabetes: a national clinical guideline: Scottish Intercollegiate Guidelines Network. Edinburgh: Scottish Intercollegiate Guidelines Network; 2010. 7. Marso SP, Kennedy KF, House JA, McGuire DK. The effect of intensive glucose control on all-cause and cardiovascular mortality, myocardial infarction and stroke in persons with Type 2 diabetes mellitus: a systematic review and meta-analysis. Diab Vasc Dis Res. 2010;7(2):119–30. 8. Selvin E, Bolen S, Yeh HC, Wiley C, Wilson LM, Marinopoulos SS, et al. Cardiovascular outcomes in trials of oral diabetes medications: a systematic review. Arch Intern Med. 2008;168(19):2070–80. 9. Brugts JJ, Yetgin T, Hoeks SE, Gotto AM, Shepherd J, Westendorp RG, et al. The benefits of statins in people without established cardiovascular disease but with cardiovascular risk factors: meta-analysis of randomised controlled trials. BMJ. 2009;338:b2376. 10. Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA. 10year follow-up of intensive glucose control in Type 2 diabetes. New Eng J Med. 2008;359(15):1577–89. 11. Holman RR, Paul SK, Bethel MA, Neil HA, Matthews DR. Long-term follow-up after tight control of blood pressure in type 2 diabetes. New Eng J Med. 2008;359(15):1565–76. 12. Jackson R, Lawes CM, Bennett DA, Milne RJ, Rodgers A. Treatment with drugs to lower blood pressure and blood cholesterol based on an individual’s absolute cardiovascular risk. Lancet. 2005;365(9457):434–41. 13. D’Agostino RB Sr, Vasan RS, Pencina MJ, Wolf PA, Cobain M, Massaro JM, et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008;117(6):743–53. 14. New Zealand Guidelines Group. Evidence-based best-practice guideline. The assessment and managment of cardiovascular risk. Wellington: New Zealand Guidelines Group; 2003. 15. New Zealand Guidelines Group. New Zealand cardiovascular guidelines handbook: a summary resource for primary care practitioners. 2009 edition. Wellington: New Zealand Guidelines Group; 2009. 16. Anderson K, Odell P, Wilson P, Kannel W. Cardiovascular disease risk profiles. Am Heart J. 1991;121:293–8. 17. Elley CR, Kenealy T, Robinson E, Drury PL. Glycated haemoglobin and cardiovascular outcomes in people with Type 2 diabetes: a large prospective cohort study. Diabet Med. 2008;25(11):1295–301. 18. Selvin E, Marinopoulos S, Berkenblit G, Rami T, Brancati FL, Powe NR, et al. Meta-analysis: glycosylated hemoglobin and cardiovascular disease in diabetes mellitus. Ann Intern Med. 2004;141(6):421–31. 19. Fox CS, Sullivan L, D’Agostino RB Sr, Wilson PW. The significant effect of diabetes duration on coronary heart disease mortality: the Framingham Heart Study. Diabetes Care. 2004;27(3):704–8. 20. Newman DJ, Mattock MB, Dawnay AB, Kerry S, McGuire A, Yaqoob M, et al. Systematic review on urine albumin testing for early detection of diabetic complications. Health Technol Assess. 2005;9(30):iii–vi, xiii–163. 21. Elley C, Robinson E, Kenealy T, Bramley D, Drury P. Derivation and validation of a new cardiovascular risk score for people with Type 2 diabetes: The New Zealand Diabetes Cohort Study. Diabetes Care. 2010;33(6):1347–52. 22. Altman DG, Vergouwe Y, Royston P, Moons KG. Prognosis and prognostic research: validating a prognostic model. BMJ. 2009;338:b605. 23. Bannink L, Wells S, Broad J, Riddell T, Jackson R. Web-based assessment of cardiovascular disease risk in routine primary care practice in New Zealand: the first 18 000 patients (PREDICT CVD-1). N Z Med J. 2006;119(1245):U2313. 24. McGeechan K, Macaskill P, Irwig L, Liew G, Wong TY. Assessing new biomarkers and predictive models for use in clinical practice: a clinician’s guide. Arch Intern Med. 2008;168(21):2304–10. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Early connections: effectiveness of a pre-call intervention to improve immunisation coverage and timeliness Felicity Goodyear-Smith MBChB, MD, FRNZCGP;1 Cameron Grant MBChB, FRACP, PhD;2 Tracey Poole;1 Helen Petousis-Harris BSc, PhD;1 Nikki Turner MBChB, MPH, FRNZCGP;1 Rafael Perera Sc, DPhil;3 Anthony Harnden MBChB, MSc, FRCGP, FRCPCH3 1 ABSTRACT INTRODUCTION: Children who have missed or delayed immunisations are at greater risk of vaccinepreventable diseases and getting their first scheduled dose on time strongly predicts subsequent complete immunisation. Developing a relationship with an infant’s parents and general practice staff soon after birth followed by a systematic approach can reduce the number of delayed first immunisations. AIM: To assess the effectiveness of a general practice–based pre-call intervention to improve immunisation timeliness. METHODS: Clustered controlled trial of general practices in a large urban district randomised to either delivery of pre-call intervention to all babies at aged four weeks or usual care. Department of General Practice and Primary Health Care, Faculty of Medical and Health Science, The University of Auckland, Auckland, New Zealand 2 Department of Paediatrics, Faculty of Medical and Health Science, The University of Auckland 3 Department of Primary Health Care, University of Oxford, Oxford, England RESULTS: Immunisation timeliness for infants receiving the primary series of immunisations among their nominated Auckland general practices was higher than expected at 98% for the six week event. The intervention was statistically but not clinically significant. Coverage was significantly lower among infants with no nominated practice which reduced overall coverage rate for the district. DISCUSSION: Pre-call letters with telephone follow-up are simple interventions to introduce into the practice management system and can be easily implemented as usual standard of care. Early identification of newborn infants, primary care engagement and effective systems including tracking of infants not enrolled in general practices has the greatest potential to improve immunisation coverage rates even further. KEYWORDS: Randomized controlled trial; immunization; vaccination; general practice; intervention studies Introduction New Zealand (NZ) historically has mediocre immunisation coverage of children1 and relatively high rates of vaccine-preventable disease.2 The risk of vaccine-preventable diseases is greater if childhood immunisations are incomplete, which includes both missed and delayed immunisations. Delays in immunisation puts infants at significant risk of contracting and being hospitalised for diseases such as Haemophilus influenzae type b3–6 and pertussis4,6–9 and increases the potential reservoir of disease in unvaccinated infants.5,10 Receiving the first dose on the vaccination schedule on time is one of the strongest and most consistent predictors of subsequent complete immunisation11–13 and delays are significantly more likely to result in lower overall coverage.12,14–16 In our previous study of 124 practices in the Auckland and Midland regions of NZ, median coverage at six weeks for the diphtheria, tetanus and acellular pertussis immunisation was 93%, while timely receipt of this dose was only 40%.17 Factors that impact upon immunisation receipt and timeliness are now well established. These VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):189–198. CORRESPONDENCE TO: Felicity Goodyear-Smith Professor, Department of General Practice and Primary Health Care, Faculty of Medical and Health Science, The University of Auckland, PB 92019 Auckland, New Zealand f.goodyear-smith@ auckland.ac.nz 189 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH include the knowledge and attitudes of caregivers, particularly antenatally,18–22 attitudes of health professionals23,24 and aspects of health care systems such as cost, recall and reminders and cost to provider.25 The early establishment of a relationship between general practices and the infant’s parents can reduce the number of infants whose first immunisation is delayed.12 Our prior study demonstrated that early enrolment with a primary care provider was associated with a higher level of immunisation completeness.17,26,27 The NZ immunisation schedule from June 2008 for the first six months of life consisted of two combination vaccines: INFANRIX® hexa and Prevenar ® given at ages six weeks, three months and five months. INFANRIX® hexa contains antigens from diphtheria, tetanus, and pertussis (DTaP), polio, Haemophilus influenzae type b and hepatitis B. Prevenar ® (PCV7) is a conjugate pneumococcal vaccine that contains antigens from seven pneumococcal serotypes that are predominant causes of invasive pneumococcal disease.28 In addition, Bacillus Calmette-Guérin (BCG) is given to infants living in households with people who have, or have had, tuberculosis (TB) or are in immigrant families from countries where TB is common, and HBvaxPRO® (hepatitis B vaccine) is given with hepatitis B–specific immunoglobulin to newborns of mothers who are hepatitis B carriers. NZ practices typically have recall systems set up in their electronic practice management systems (PMSs) to remind parents that their child’s immunisations are overdue. All childhood scheduled immunisations should be recorded on the National Immunisation Register (NIR) at the time of delivery, with data directly transmitted from the practices. The NIR sends reminder messages to practices if information about immunisation events is delayed. The overdue times for NIR are set outside those for the PMS so that the practices have time to follow up before they start receiving overdue messages from the NIR. The PMS will generate a recall for a child at eight weeks if the six week doses have not been given, whereas the NIR will consider that immunisation event overdue when the child is aged 10 weeks. Similarly, 190 PMS and NIR timeliness ‘windows’ for the three month vaccines are aged four and four and a half months respectively, and again for the five month event aged six and six and a half months. From our previous work, we hypothesised that the enrolment of children with a general practice soon after birth, and parents being actively invited when their baby is four weeks old to attend the practice for their first (six week) set of vaccines, followed up with early phone calls if they do not respond, would improve immunisation timeliness. We therefore aimed to conduct a randomised controlled trial (RCT) of a general practice–based precall intervention. Our objectives were to assess the effectiveness of this enhanced practice system on coverage and timeliness of the six week, three month and five month immunisations. Methods Study design This was an RCT of a multicomponent intervention compared with usual care. Randomisation was at the level of the general practice. The study was registered with Australia New Zealand Clinical Trials Registry (00082892) and ethical approval was obtained from the Ministry of Health Auckland Regional X Ethics Committee (Reference NTX/08/08/072). Setting and study population The setting was practices in the Auckland District Health Board (ADHB) catchment area and the study population was babies born in the ADHB region and/or those whose parents nominated practices in the ADHB region as their general practice. The study took place between 1 November 2008 and 20 April 2010. The NIR is notified of the nominated general practice for all newborns and the practice then is informed by the NIR that they are the baby’s nominated practice. Children are tracked using the unique National Health Index (NHI) number assigned at birth. Intervention Our intervention consisted of a brief letter of welcome and invitation to attend when the baby VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH was six weeks, plus simple information on immunisation. This was sent by the practice to the baby’s caregiver when the baby was four weeks old (see appendix in the web version of this paper for this pre-call material). A follow-up phone call was made to the baby’s caregiver when the baby was five weeks old if an appointment had not already been made for the six week vaccinations and, if the caregiver did not present the baby for immunisation, a further attempt at contact (early recall) was made to the caregiver when the baby was seven weeks old. Phone calls were the preferred method of pre-call/recall although a text message, email or letter could also be used. Practices were given a $15 shopping mall voucher per baby to acknowledge the time and effort required to administer the intervention, to be claimed irrespective of whether or not it resulted in immunisation of the baby. Inclusion/exclusion criteria and practice allocation A database of all 148 general practices operating in the ADHB region was created by combining general practice databases held by the Immunisation Advisory Centre and the Department of General Practice and Primary Health Care. Telephone directories were cross-referenced to ensure all practices in the ADHB region were included. All practices were assigned a number (practice code). Practices identified as not involved in delivering infant immunisations were excluded. Following consent, block randomisations were conducted of recruited practices using a computer random number generator to assign each practice to either the intervention or control group. The research team other than the project manager were blind to the identity of practices in the intervention and control groups. Intervention practices were assisted to adjust their PMS to automatically send out pre-call and recall letters to their patients in the study. Outcome measures Our primary outcome measure was receipt of six week immunisations and age at which these were delivered as recorded on the NIR for all the babies in our study in intervention, control and non-participating practices. Secondary measure- WHAT GAP THIS FILLS What we already know: Children who have delayed or missed immunisation events are at greater risk of vaccine-preventable diseases. Getting their first scheduled dose on time strongly predicts subsequent complete immunisation, and developing a relationship between the general practice staff and a baby’s parents soon after birth can reduce the number of delayed first immunisations. What this study adds: Immunisation coverage and timeliness for infants receiving the primary series of immunisations among their nominated Auckland general practices is extremely high, with no clinically relevant room for improvement. A pre-call intervention made a statistically significant improvement in timeliness of immunisation, but only by one day. Coverage was significantly lower among infants with no nominated practice and this reduced the overall coverage rate for a district. Targeting both the systems and services that can identify and track infants who are not engaged with primary care at birth has the greatest potential to improve immunisation coverage rates even further. ments were receipt of three month and five month immunisations and age at which these were delivered. Pre- and post-trial surveys were also conducted for participating practices to establish their practice population and their pre-call/ recall practices before, during and after the trial. Power calculations This was a clustered randomised trial with each enrolled practice being a cluster. Because receipt of all three primary series doses is important to most effectively reduce risk of vaccine-preventable diseases, the trial was powered on receipt of three month and five month immunisations. Our previous study had shown that the inflation effect could be between 20 and 40 (see Table 1).17 A significant contributor to this large inflation effect was the fact that at that time some practices were not enrolling pre-school children in their practices for pragmatic reasons which led to a large inter-practice variability in immunisation timeliness and coverage rates. At the time of this current study, this problem had largely been addressed with the introduction of the NIR plus the introduction of children being fully vaccinated by their second birthday as a performance indicator as part of the Primary Health Organisation Performance Management Programme. With most or all children being enrolled with practices VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 191 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Table 1. Summary of sample size calculations Immunisation doses Timely immunisation coverage Current % Desired % Sample size required DE = 20 Sample size required DE = 10 6 weeks 88 98 4000 2000 3 months 75 94 2000 1000 5 months 60 85 2000 1000 DE = Design effect soon after birth, the design effect was expected to be significantly smaller. We planned to deliver 1000 interventions. This sample size was calculated to be sufficient to have 80% power at the 0.05 level of statistical significance to detect an increase in the timely immunisation coverage from 75% to 94% at the three month and from 60% to 85% at the five month immunisations. Analysis The variables in the NIR dataset consisted of NHI number (converted to a unique identifier), date of birth, dates when six week, three month and five month immunisations given (vaccines coded v1 for INFANRIX® hexa and v2 for Prevenar ®), nominated practice at birth, practice(s) giving immunisation for all children born in Figure 1. Recruitment of practices in ADHB All general practices in ADHB 148 Eligible practices 128 Recruited practices 63 Intervention practices (A) 31 192 Non-eligible practices 20 Non-recruited practices (C) 65 Control practices (B) 32 the ADHB catchment area for the study period (1 November 2008 to 20 April 2010). At this date all intervention babies were aged six months or older. Immunisation events identified as BCG vaccine were deleted. Survival curve analysis was used to measure delay in immunisation. For each infant participant the days from their ideal immunisation date (i.e. 42 days for six week vaccine) to the actual day they received vaccine were counted. For this analysis, second and third dose assumptions were made, i.e. if the three- and/or five month vaccines had been given it was assumed that the child had previously received the earlier doses. We compared total scores (i.e. number of delayed days) using survival analysis (Kaplan-Meier and Cox proportional hazards) for (1) intervention versus control group (intention to treat analysis), (2) pre-call versus non-pre-call in the intervention group, and (3) pre-call in the intervention group versus control group to test if there was any significant differences. This allowed analysis by continuous rather than dichotomous data (i.e. defining an immunisation event as either on time or delayed). This provided greater statistical power and allowed for graphic representation of results plotting number of delayed days over time for both groups. Adjustment for clustering effect was conducted. Results Practice recruitment is presented in Figure 1. From 128 eligible practices, 63 were recruited with 31 randomised to the intervention group (A) and 32 to the control group (B). Groups A, B and C (non-recruited practices) were similar with respect to the socioeconomic status of the practice locations and the average practice size. The num- VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Figure 2. Babies born in practice groups during the study period ADHB 11 555 babies born Intervention (A) practices 2842 (24.6%) babies Pre-call delivered 1198 (10.4%) babies Control (B) practices 2414 (20.9%) babies Non-participating (C) practices 5827 (50.4%) babies No nominated practice 472 (4.1%) babies Pre-call not delivered 1644 (14.2%) babies from the NIR in time, practice nurses stopping the intervention for periods of time (for example during the summer period when there were locum nurses), plus one A practice failed to deliver any interventions at all. In many instances details were not recorded as to whether or not the intervention included a five-week pre-call or seven-week recall as well as the four-week precall mail-out, so sub-analyses of how often these were required was not possible. ber of babies born in the ADHB between 1 November 2008 and 20 April 2010 (a one year 5.75 month period) was 11 555 (see Figure 2). This is 7834 babies per calendar year, which was close to our estimate of 8000 babies per year for the 128 eligible general practices. Half of the eligible practices (63) were enrolled in the study, and close to half (46%) of the babies were nominated to these practices. There were slightly more babies born to the control (B) than the intervention (A) practices. A small number of infants (n=472, 4%) had no nominated practice. The overall coverage rate for the six week vaccine ‘1’ (INFANRIX® hexa) for A, B and C practices delivered on time by eight weeks of age was 98%—see Table 2. This was also the case for vaccine 2 (Prevenar ®). Scatter plots of the difference in timing between v1 and v2 for all cases Only 1198 of the 2842 babies in the A practices received pre-call interventions (42%). The reasons for these not being delivered included practices not receiving or being aware of the notification Table 2. Overall vaccination rate for six week event for vaccine 1 for intervention, control, non-participating and no nominated practice Type of practice Six week vaccine 1 received by age eight weeks Opted off / Declined No NIR data Total % completed by practice type A 2743 53 46 2842 97 B 2388 26 0 2414 99 C 5744 83 0 5827 99 D1 32 16 0 48 67 D2 0 0 424 424 0 10 907 177 470 11 555 % by A, B or C only 98% % overall 94% Key: A = Intervention practice B = Control practice C = Non-participating practice D1 = No nominated practice, six week immunisation data available D2 = No nominated practice, six week immunisation data not available v1 = INFANRIX® hexa VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 193 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Figure 3. Differences in timing between vaccines 1 and 2 for six week vaccination event and then for those in practices A and B showed that these were rarely given separately; hence, separate analyses for v1 and 2 were not required. See Figure 3 for scatter plots for the six week events which shows a straight line when time of delivery of vaccine 1 is plotted against time of delivery of vaccine 2. When the 472 children with no nominated practice (D) were included in the total, the overall vaccination rate completed by eight weeks of age for the six week vaccination dropped from 98% to 94%. Table 3 shows the average age of receipt of INFANRIX® hexa at six weeks, three months and five months by practice type for children receiving immunisations. It can be seen that non-participating practices (C) consistently had a slightly longer average delay than recruited practices (A and B), but children without a nominated practice who were vaccinated had a much greater average delay. While the majority (88%) of the babies attended the practice their parents nominated at birth, 12% were either vaccinated or declined vaccination at a different practice for vaccine 1 at the six week event. We analysed data both using the second and third dose assumptions (e.g. if the three month dose was recorded in the NIR, we assumed the six week event had been given) and not making this assumption but categorising these children as having no information on the six week vaccination event. This made a slight difference in coverage rates. When applied to immunisation registers, the third-dose assumption results in an over-estimate of immunisation coverage that is smaller than the underestimate produced by assuming all those with missing data have not been immunised.29 When the vaccination times of A and B practices for receipt of the six week vaccine were compared (intention-to-treat analysis), there was no indication of a difference between the groups (Log Rank (Mantel-Cox): Chi-square 0.268, df=1, p=0.605). There also was no difference in days to vaccination for the three month vaccine event (Log Rank (Mantel-Cox): Chi-square 0.540, df=1, p=0.46) nor for the five month vaccine (Log Rank (Mantel-Cox): Chi-square 0.281, df=1, p=0.60). Table 3. Average age of receipt of INFANRIX® hexa at six weeks, three months and five months by practice type Practice type Average age in weeks of receiving six week v1 Average age in months of receiving three month v1 Average age in months of receiving five month v1 A 7.06 3.40 5.64 B 7.09 3.43 5.65 C 7.26 3.48 5.7 D 8.92 4.49 6.32 Key: A = Intervention practice B = Control practice C = Non-participating practice D = No nominated practice v1 = INFANRIX® hexa 194 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH However, this analysis did not take into consideration the fact that the 58% of children in the A practices did not receive the intervention. We therefore conducted a second analysis looking at vaccination in group A, stratifying by actual delivery of the intervention. This showed that babies receiving the intervention were statistically more likely to receive their six week vaccination event earlier (Log Rank (Mantel-Cox): Chi-square 19.187, df=1, p<0.001) with mean days to six week vaccination event 49.6 days for those who received interventions compared with 51.2 days for those who did not—see Figure 4. An advantage of this analysis is that it takes into account types of censored data, for example where six week data is missing but the third dose assumption is used. The same pattern was repeated for vaccination coverage stratified by delivery of intervention for A practices for the three month vaccination event (Log Rank (Mantel-Cox): Chi-square 16.527, df=1, p<0.001) with mean days 100.1 compared with 103.8, and for the five month event (Log Rank (Mantel-Cox): Chi-square 11.621, df=1, p=0.001) with mean days 166.2 compared with 170.1. A third analysis comparing only those babies in the A practices who had received the intervention with the group B babies found a similarly significant result for the six week vaccination event results (Log Rank (Mantel-Cox): Chi-square 5.969, df=1, p=0.015), the three month (Log Rank (Mantel-Cox): Chi-square 10.722, df=1, p=0.001) and the five month (Log Rank (MantelCox): Chi-square 6.753, df=1, p=0.009). While there was no statistical difference in timeliness between A and B groups, we expected that recruited practices (A and B) would have less delay in immunisation overall than C practices which declined to participate (and may be less focused on vaccination). We therefore repeated the timeliness analysis including group C for the six week vaccine. We found that group C practices had a significantly more delayed vaccine rate for the six week vaccine (Log Rank (Mantel-Cox): Chi-square 14.705, df=1, p=0.001). Lastly, analysis of practice surveys found that 13 A and 12 B practices used some form of pre-call prior to the trial. All but one A practice (which had failed to use the pre-call intervention) intended to continue using our pre-call intervention post-trial. Discussion At the commencement of our study in 2008, NZ immunisation rates were estimated from the NIR data to be 88% for the six week vaccine event, 75% for the three month event and 60% for the Figure 4. Days to six week vaccination event in Group A stratifying by actual delivery of the intervention Comparison of time to vaccination comparing intervention delivered or not delivered Log of comparison of time to vaccination comparing intervention delivered or not delivered to demonstrate differences The window for ‘timely’ six week vaccination is 6–10 weeks (42–70 days) VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 195 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH five month event. We calculated our sample size based on these data. However, we found the coverage rate for enrolled children in all practices in the study was much higher than expected, at 98% for the six week vaccine and dropped off very little for the three month and five month doses. We did find a statistically significant improvement in timeliness of vaccine receipt at the six week, three month and five month events comparing children who received the intervention with those who had not, but because the coverage rates were so high, this only translated into children receiving the vaccine on average one day earlier (from 49.5 days to 50.5 days for the vaccine due at 42 days), which is not clinically significant. Given that there were much higher coverage rates across all practices to start with, the lack of clinical significance is not surprising. While the average coverage among all the ADHB practices was 98%, there was a small percentage of infants without nominated practices (n=472, 4%). Thirty-three percent of infants with no nominated practice for whom there was an entry for the six week vaccination event were not vaccinated. When these infants were added to the total, the overall coverage rate dropped to 94%. This is an important finding. Infants with no nominated general practices are much less likely to get immunised and, if they do, are much more likely to be delayed. This reduces the overall coverage rate significantly for the region (making the general practices look as though they are performing less well than they are). Children who are not getting general practice services including immunisation pre-call and/or recall can be considered to be ‘falling through the cracks’: Attention needs to be given to this group of children, both encouraging earlier general practice enrolment for them and targeting outreach services for those unenrolled. We identified a number of places where errors occur in the collection of data in the NIR: the practitioner (usually the practice nurse) may enter incorrect data in the PMS, there may be technical problems with the transfer of data from the PMS to the NIR, the NIR might send back error messages to the practitioners’ inboxes that they do not know how to action, a practitioner may not 196 know how to manage entry of complex ‘catchup’ schedules when a child is presenting late or has received a different immunisation regime in another country, or the PMS might not have the facility to record these complex immunisation entries accurately. An unvaccinated child presenting at three months may have this immunisation event recorded as the three month rather than the six week vaccination event. Furthermore when patients transfer practices it is usually not possible for their full clinical records to be transferred electronically between practices. This means that the new practice needs to manually enter previous vaccines from paper records or the child well health book, which will not always be available. How our findings relate to what is already known Timeliness of children’s vaccination varies widely between and within countries.30–39 There are a range of factors associated with timeliness reported in these studies. Ethnicity, area of residence21,40,41 and negative media coverage are all associated with delay in vaccination.36 Integrated systems including outreach and recall has been shown to be effective. For example, an extensive programme in Chicago which combined immunisation education at birth with ongoing reminder–recall achieved over 90% on-time adherence for recommended immunisations among inner-city children aged 0–35 months.42 Having a nurse vaccine manager who is in charge of tracking inventory, training staff, and developing vaccination protocols can improve timeliness of vaccine delivery.43 Our study shows that enrolment and early engagement with a general practice is resulting in excellent coverage and timeliness. However the children not enrolled with a general practice fare poorly and represent a small but significant group for whom outreach services should be targeted, and assistance given for the infant to join a general practice. Strengths and limitations We piloted the intervention set-up, received early datasets for testing, performed dummy runs, dou- VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH ble-entered data and performed meticulous data cleaning to ensure we had an accurate dataset to analyse. However, we recruited fewer practices to the study (63/128, 50%) than the 60% we had intended and we had delays in some practices implementing the intervention. The intervention was only delivered to 42% of the eligible children in the intervention practices, plus some of the practices reported that they were already doing some form of recall; hence, the comparison between intervention and control group had no significant difference. We had incomplete records of whether or not the delivered intervention included five-week pre-call or seven-week recall contacts. Recommendations As a result of this study we recommend that NIR notify general practices so that they are aware of all newborns for whom the practice is the nominated provider as soon as possible after birth, and that DHBs follow up newborns with no nominated providers to ensure registration with a provider as early as possible. Different regions will use different strategies to achieve this. We further recommend that practices send a pre-call letter with accompanying information about immunisation when infants are aged four weeks, as a fully automated prompt system followed by telephone contact if the family does not make contact seems to be a cost-efficient and sensible strategy. Because many Auckland practices are already doing some form of pre-call, having a standardised four-week pre-call letter prompt system that all practices could use may have a positive effect on timeliness. It could assist those practices who are not pre-calling and act as a reaffirmation to those practices who are pre-calling that their current commitment is worthy of continuing. Conclusion We found immunisation coverage and timeliness for infants receiving the primary series of immunisations among their nominated Auckland general practices to be extremely high, with no clinically relevant room for significant improvement. The intervention trialled in this study made a statistically significant improvement to timeliness of vaccination; however, only by one day. However, coverage was significantly lower among infants with no nominated practice and this reduced the overall coverage rate. Non-enrolment of babies at birth with a general practice is a significant factor in delayed or missed immunisations. Targeting both the systems and services that can identify and track these infants has the greatest potential to improve immunisation coverage rates even further. References 1. Ministry of Health. The National Childhood Immunisation Coverage Survey 2005. Wellington; 2007. 2. Ministry of Health. Immunisation handbook 2006. Wellington: Ministry of Health; 2006. 3. Grant C, Roberts M, Scragg R, Stewart J, Lennon D, Kivell D. Delayed immunisation and risk of pertussis in infants: unmatched case-control study. BMJ. 2003;326:852–3. 4. Kolos V, Menzies R, McIntyre P. Higher pertussis hospitalization rates in indigenous Australian infants, and delayed vaccination. Vaccine. 2007;25(4):588–90. 5. Singleton R, Bulkow LR, Levine OS, Butler JC, Hennessy TW, Parkinson A. Experience with the prevention of invasive Haemophilus influenzae type b disease by vaccination in Alaska: the impact of persistent oropharyngeal carriage. J Ped. 2000;137(3):313–20. 6. Grant C, Scragg R, Lennon D, Ford R, Stewart J, Menzies R. Incomplete immunisation increases the risk of pertussis in infants. BMJ. 2003;326:852–3. 7. Somerville RL, Grant CC, Grimwood K, et al. Infants hospitalised with pertussis: estimating the true disease burden. J Paed Child Health. 2007 Sep;43(9):617–22. 8. Greenberg DP, Doemland M, Bettinger JA, et al. Epidemiology of pertussis and Haemophilus influenzae type b disease in Canada with exclusive use of a diphtheria-tetanus-acellular pertussis-inactivated poliovirus-Haemophilus influenzae type b pediatric combination vaccine and an adolescent-adult tetanus-diphtheria-acellular pertussis vaccine: implications for disease prevention in the United States. Pediatr Infect Dis J. 2009;28(6):521–8. doi: 10.1097/INF.0b013e318199d2fc. 9. Crowcroft NS, Pebody RG. Recent developments in pertussis. Lancet. 2006 Jun 10;367(9526):1926–36. 10. Anonymous. Pertussis epidemic in New Zealand, 1990/1991. Commun Dis NZ. 1991;91(9):88–9. 11. Bobo JK, Gale JL, Thapa PB, Wassilak SG. Risk factors for delayed immunization in a random sample of 1163 children from Oregon and Washington. Pediatrics. 1993;91(2):308–14. 12. Guyer B, Hughart N, Holt E, et al. Immunization coverage and its relationship to preventive health care visits among innercity children in Baltimore. Pediatrics. 1994;94(1):53–8. 13. Williams IT, Milton JD, Farrell JB, Graham NM. Interaction of socioeconomic status and provider practices as predictors of immunization coverage in Virginia children. Pediatrics. 1995;96(3 Pt 1):439–46. 14. Farizo KM, Cochi SL, Zell ER, Brink EW, Wassilak SG, Patriarca PA. Epidemiological features of pertussis in the United States, 1980–1989. Clin Infect Dis. 1992 Mar;14(3):708–19. 15. Wood D, Donald Sherbourne C, Halfon N, et al. Factors related to immunization status among inner-city Latino and African-American preschoolers. Pediatrics. 1995;96(2 Pt 1):295–301. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 197 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH ACKNOWLEDGEMENTS Thanks to the general practices involved in the study for their valuable cooperation and assistance in making this study possible and, in particular, the practice nurses precalling the children at the intervention practices. We appreciate the valuable advice provided to us by our Stakeholder Advisory Board on design and delivery of the intervention. Thanks also to Mr Charles Misquitta and Ms Janine Peters at Auckland District Health Board (ADHB) for their assistance with the ADHB National Immunisation Register (NIR), Mr Mishra Suryaprakash at the Ministry of Health (MoH) for communication with the National Immunisation Register Governance Group, members of the NIR Governance Group for approving this study, Rachel Guthrie, Population Event Registers Team Lead, MoH, for assistance with accessing NIR data, Deirdre McMahon and Dipan Ranchhod (Data Quality Analyst), National Collections Data Management Sector Services, MoH, for assistance with NIR data, and to Dr Lynn Taylor, Operations Manager of IMAC for information on potential errors occurring between practice and NIR messaging. FUNDING This study was funded by the Health Research Council of New Zealand and the Ministry of Health. COMPETING INTERESTS None declared. 198 16. Strine TW, Luman ET, Okoro CA, McCauley MM, Barker LE. Predictors of age-appropriate receipt of DTaP Dose 4. Am J Prev Med. 2003;25(1):45–9. 17. Grant CC, Turner NM, York DG, Goodyear-Smith F, PetousisHarris HA. Factors associated with immunisation coverage and timeliness in New Zealand. Br J Gen Pract. 2010 Mar;60(572):e113–20. 18. Samad L, Butler N, Peckham C, Bedford H. Incomplete immunisation uptake in infancy: maternal reasons. Vaccine. 2006;24(47–48):6823–9 19. Fields V, Sumpter C, Seagraves L, Laney A, Lott A, McBride K. Determining factors affecting parental non-compliance with vaccination schedules of children ages 6 months to 2 years. La Grange College; 2007. Available from: http://www. lagrange.edu/resources/pdf/citations/2007/nursing/nursing%20-%20fields.pdf 20. Tickner S, Leman PJ, Woodcock A. Factors underlying suboptimal childhood immunisation. Vaccine. 2006;24(49– 50):7030–6. 21. Trauth JM, Zimmarman RK, Musa D, Mainzer H, Nutini JF. Do beliefs of inner-city parents about disease and vaccine risks affect immunization? J Nat Med Assoc. 2002;94(9):820–32. 22. Wroe A, Turner N, Salkovskis PM. Understanding and predicting decisions about early childhood immunisations. Health Psych. 2004;23(1):33–41. 23. Pattison H, Pareek M. Health professionals’ attitudes to MMR vaccine. Advice in primary care affects parents’ decision to take up MMR vaccination.[comment]. BMJ. 2001;322(7294):1121. 24. Petousis-Harris H, Goodyear-Smith F, Soe B, Turner N. Family practice nurses perspectives on barriers to childhood immunisation. Vaccine. 2005;23:2725–30. 25. Briss PA, Rodewald LE, Hinman AR, Shefer AM, Strikas RA, Bernier RR, Carandekulis VG, Yusuf HR, Ndiaye SM, Williams SM. Reviews of evidence regarding interventions to improve vaccination coverage in children, adolescents, and adults. Am J Prev Med. 2000 January 2000;18(1):97–140. 26. Goodyear-Smith F, Grant C, York D, et al. Determining immunisation coverage rates in primary health care practices: A simple goal but a complex task. International J Med Inform. 2008;77:477–85. 27. Desmond N, Grant C, Goodyear-Smith F, Turner N, PetousisHarris H. Nurses make a difference in immunisation service delivery. Aust J Adv Nurs. 2011;28(4):31–5. 28. Immunisation Advisory Centre. New Zealand National Immunisation Schedule from 1 September 2008. Auckland, 2008. 1 p. 29. Hull B, Lawrence G, MacIntyre C, McIntyre P. Estimating immunisation coverage: is the ‘third dose assumption’ still valid? Commun Dis Intell. 2003;27:351–61. 30. Clark A, Sanderson C. Timing of children’s vaccinations in 45 low-income and middle-income countries: an analysis of survey data. Lancet. 2009 May 2;373(9674):1543–9. 31. Luman ET, Barker LE, McCauley MM, Drews-Botsch C. Timeliness of childhood immunizations: a state-specific analysis. Am J Pub Health. 2005 Aug;95(8):1367–74. 32. Luman ET, McCauley MM, Stokley S, Chu SY, Pickering LK. Timeliness of childhood immunizations. Pediatrics. 2002 November 1, 2002;110(5):935–9. 33. Cotter JJ, Bramble JD, Bovbjerg VE, et al. Timeliness of immunizations of children in a Medicaid primary care case management managed care program. J Nat Med Assoc. 2002;94(9):833–40. 34. de Nuncio MLZ, Nader PR, Sawyer MH, De Guire M, Prislin R, Elder JP. A prenatal intervention study to improve timeliness of immunization initiation in Latino infants. J Community Health. 2003;28(2):151–65. 35. Hull BP, McIntyre PB. Timeliness of childhood immunisation in Australia. Vaccine. 2006;24(20):4403–8. 36. Dannetun E, Tegnell A, Hermansson G, Torner A, Giesecke J. Timeliness of MMR vaccination—influence on vaccination coverage. Vaccine. 2004;22(31–32):4228–32. 37. Luman ET, Barker LE, Shaw KM, McCauley MM, Buehler JW, Pickering LK. Timeliness of childhood vaccinations in the United States: days undervaccinated and number of vaccines delayed. JAMA. 2005 Mar 9;293(10):1204–11. 38. Bailie RS, Si D, Dowden MC, et al. A systems approach to improving timeliness of immunisation. Vaccine. 2009 Jun 2;27(27):3669–74. 39. O’Grady K-A, Krause V, Andrews R. Immunisation coverage in Australian indigenous children: time to move the goal posts. Vaccine. 2009 Jan 7;27(2):307–12. 40. Dombkowski KJ, Lantz PM, Freed GL. The need for surveillance of delay in age-appropriate immunization. Am J Prev Med. 2002;23(1):36–42. 41. Ehresmann KR, White KE, Hedberg CW, et al. A statewide survey of immunization rates in Minnesota school age children: Implications for targeted assessment and prevention strategies. Ped Infect Dis J. 1998;17(8):711–6. 42. Vora S, Verber L, Potts S, Dozier T, Daum RS. Effect of a novel birth intervention and reminder-recall on on-time immunization compliance in high-risk children. Hum Vac. 2009 Jun;5(6):395–402. 43. Joyce C. Steps to success: getting children vaccinated on time. Ped Nursing. 2007 Nov–Dec;33(6):491–6. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH PSA testing in general practice Fraser Hodgson MBChB, DipObst, FRNZCGP, PGCertTravMed, PGDipPHC; Zuzana Obertová PhD; Charis Brown BCom, MIndS, PhD; Ross Lawrenson MBBS, MD, FRCGP, FAFPHM, FFPH. Waikato Clinical School, The University of Auckland, Hamilton, New Zealand ABSTRACT INTRODUCTION: In New Zealand, prostate-specific antigen (PSA) testing has increased significantly (275 000 tests/year). Controversy exists around PSA testing as part of an unorganised screening programme. AIM: To look at the use of PSA testing in a sample of general practices and investigate the reasons GPs undertake PSA testing. METHODS: Five Waikato general practices investigated looking at PSA laboratory tests of men ≥40 years in 2010 compared against GP notes. Testing rates, reasons for testing, histology and referral/s were examined for different age groups. A questionnaire was sent to the GPs to determine their views on PSA testing. RESULTS: One in four men aged 40+ years had a PSA test in 2010. Of these men, 71% were asymptomatic. More than half of men tested aged 70+ years were asymptomatic. Ten percent of all PSA tests were elevated. Twenty-one of 23 prostate cancers were diagnosed following an elevated PSA test: more than 80% of these men had histories of prostate pathology or lower urinary tract symptoms. The questionnaire confirmed that GPs believe in the benefits of PSA screening and it also showed they had difficulty in providing patients with information about pros and cons of PSA testing. DISCUSSION: All GPs in this study tested asymptomatic men. GPs in this study value PSA screening and believe that it reduces mortality rates. However, although PSA tests were most frequently done on asymptomatic patients, the majority of patients subsequently diagnosed with prostate cancer had been tested because of symptoms or had previous prostate problems. KEYWORDS: Prostate specific antigen (PSA); PSA testing; screening; prostate cancer; general practitioners Introduction Prostate cancer is a common cause of male cancer in New Zealand with approximately 3000 new cases diagnosed each year and 560 deaths.1 The natural history of prostate cancer is that it usually occurs in older men. It is slowly progressive with a long lead-time to diagnosis and symptoms. Five-year survival rates are also high at more than 80%.2–4 The long lead-time prior to symptoms suggests prostate cancer to be a good candidate for screening. The PSA test is helpful as a management tool in patients with established prostate cancer; despite its limitations it is also used as a screening tool, being relatively cheap and simple to use.5 General practitioners (GPs) face conflicting messages about the need to screen. The Urological Society of Australia and New Zealand believe GPs should offer asymptomatic men a PSA test.6 This advice is partly based on the results of two randomised control trials, the European Randomized Study of Screening for Prostate Cancer (ERSPC) and Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial.7,8 The ERSPC study has shown reduced mortality in men who were screened between the ages of 50 and 69 years.8 However, because of the acknowledged global issues of over diagnosis and harm caused by screening, the NZ Ministry of Health has fol- VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):199–204. CORRESPONDENCE TO: Fraser Hodgson Waikato Clinical School, The University of Auckland, PB 3200 Hamilton, New Zealand f.hodgson@auckland.ac.nz 199 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Table 1. The number and percentage of men who received a PSA test, had an elevated PSA and had been referred to a specialist in a 12-month period in five general practices by age Age group Had PSA test Years n % PSA raised % n % Total n 40–49 241 12.4 6 2.5 2 33.3 1938 50–59 481 26.7 38 7.9 16 42.1 1802 60–69 415 35.7 48 11.6 21 43.8 1161 70–79 244 35.2 32 13.1 11 34.4 694 80+ 99 30.7 23 23.2 5 21.7 323 Total 1480 25 147 9.9 55 37.4 5918 lowed the trend in other OECD countries and has avoided making recommendations supporting a national screening programme.5,9 What are GPs to do? Patients typically have a very straightforward understanding of the PSA test. A negative test means they do not have cancer, while a positive test will identify cancer early enough to allow curative treatment. They believe the risks of having a simple blood test are minimal.10 However, several studies confirm that patients’ perception of risk are not accurate.11,12 In New Zealand, it is known that PSA tests are used widely, with 275 000 tests a year being carried out principally ordered by GPs.13 In general, GPs would test asymptomatic men, but many tests are also being undertaken in men with lower urinary tract symptoms (LUTS) or previous history of prostate problems.14–19 What has not been recorded is why men are tested and what happens to the men who have been tested and are found to have a raised PSA level. This study was designed as a pilot for a larger project looking at costs and complications of screening. This pilot was to be run within the Waikato District Health Board. The Waikato region is a large geographical area covering 34 890km2 or approximately 13% of New Zealand’s land mass, with a population of 353 000. There are a number of main urban areas in the region, including Hamilton. A characteristic of this region is the number residing in rural and isolated areas (23.8% compared with 14.3% for the total population).20 The aim of this study is to examine the age-specific rate of PSA testing in five general practices 200 Referred to specialist n in the Waikato region during a 12-month period and to understand why they are being tested. We also examined the outcomes of testing. Methods This study was carried out in five Waikato-based general practices with a total population of approximately 25 000 registered patients. The practices were purposefully selected to be representative of both rural and urban practice. Ethics approval was obtained (reference number NTY/11/02/019). Practices were approached by the lead author, provided information on the content of the study and then invited to be part of the project. GP permission was sought to access PSA test results for all enrolled men aged 40 years and older from their practice who had received a PSA test result during 2010. Once permission was received, the laboratory provided all PSA results for the given period attached to the GP and the practice. In-house GP patient records were reviewed to determine the reason the PSA test was performed. These were coded under four broad categories: 1. Opportunistic testing (e.g. done with blood tests for acute or non-related chronic problems, e.g. cardiovascular risk assessment (CVRA), flu injection, non-specific check-up etc.) 2. Previous raised PSA or prostate problems (including prostate cancer, prostatitis, benign prostatic hyperplasia) 3. Patient request (patient with no symptoms) 4. Patient had evidence of lower urinary tract symptoms (LUTS), including retention, reduced flow, nocturia, urgency, frequency, haematuria, dribbling, and erectile dysfunction. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Each practice provided baseline data of their population of enrolled men aged 40 years and over, date of birth, ethnicity, and National Health Index (NHI) number. The age and PSA levels of those tested were recorded. We analysed the data looking at testing rates by age and reason for testing. We also analysed referral rates to specialists for those with a raised PSA test. A raised PSA was defined according to Pathlab recommendations: • • • • • 40–49 years >2.5ug/L 50–59 years >3.5ug/L 60–69 years >4.5ug/L 70–79 years >6.5ug/L 80+ years >7ug/L. WHAT GAP THIS FILLS What we already know: Despite ongoing debate regarding the benefits of PSA screening in general practice, PSA testing is frequently done, and the majority of tests can be considered screening rather than as an aid to diagnosis in men with symptoms. What this study adds: While many studies have asked GPs for their views regarding screening of asymptomatic patients, this study reviewed not only GPs’ testing activity but also looked at their subsequent management. An important finding was that a considerable proportion of PSA screening is performed in men aged 70+ years despite the fact that international advisory groups discourage screening in older men since it is known to cause harm and there is no evidence of reduced mortality in men aged over 70 years. Results Data were entered into an Excel spreadsheet that included date of birth and reason for testing for all men 40 years and older. In addition to reasons for PSA testing, details such as digital rectal examination, referrals, biopsies, complications, treatment, diagnosis date, stage/Gleason/extent were all captured. A questionnaire survey was later sent to the GPs in the practices to ascertain their views regarding PSA testing. Each question had a choice of five responses from ‘strongly agree’ through to ‘strongly disagree’. It included questions such as ‘I believe that PSA screening will improve mortality rates for prostate cancer’; ‘I am concerned about the harm caused to men due to PSA screening for prostate cancer’; ‘I believe that the benefit of screening outweighs any harm’; and ‘It is difficult to give a balanced view to patients regarding the pros and cons of PSA testing.’ Demographic information was also collected. We identified 5918 resident male patients aged 40 years and older in the five practices. During the 12 months 1480/5918 (25%) had been tested with at least one PSA test. The range varied 10–37% over the five practices. Testing was least likely in the 40–49 year age group and declined slightly in the 80+ year age group (see Table 1). Overall 147/1480 (10%) PSA tests were elevated (Table 1). In those with an elevated PSA, 55 out of 147 (37%) were referred to a specialist. Of the 55 referrals, 39 had a biopsy, 21 out of these had prostate cancer, and 18 were benign (Table 2). However, there were 10 referrals to specialists with normal PSA test results. These had reasonable clinical grounds (LUTS and/or abnormal DRE) to do so (Figure 1). Two men who were biopsied had prostate cancer. When we looked at the notes to ascertain the reason patients had a PSA test overall, 71% of the Table 2. Number of men with raised PSA test, referral, biopsy, and diagnosed with prostate cancer grouped by reason for PSA test Reasons for PSA test Raised PSA test Referral Biopsy Diagnosis of prostate cancer Opportunistic testing 28 6 5 4 Previous prostate problems 98 40 29 14 Patient requested test 2 1 0 0 LUTS 19 8 5 3 Total 147 55 39 21 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 201 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Figure 1. Patients with normal PSA level referred to a specialist and their outcomes Reasons for referral • 6 with lower urinary tract symptoms > 1 biopsy > 1 diagnosed with prostate cancer • 1 with abnormal digital rectal examination > 1 biopsy > 1 diagnosed with prostate cancer • 2 with previous prostate cancer and rising but normal PSA level > no biopsy • 1 with elevated PSA in 2009 and upper normal range PSA in 2010 > no biopsy Figure 2. Proportion (%) of reasons for PSA test in five general practices by patients’ age 70 years. Only 44% agreed that all men over 40 should have at least one PSA test. Seventytwo percent of GPs said they did a digital rectal examination and PSA test when checking for prostate cancer. Certain questions looked at consultation restraints surrounding PSA explanations. Thirtynine percent of the GPs felt they needed more knowledge to advise patients. Fifty-six percent felt it was difficult to give a balanced view to patients regarding PSA testing. The majority felt patients had difficulty understanding the issues despite the GP’s best efforts, and 61% said that patients elected to get the test anyway. However, the majority did not feel pressured by patients to perform PSA screening. GPs were evenly divided in their views regarding such things as medicolegal concerns, time restraints and whether patients could in fact make up their own minds about PSA testing. Discussion time GPs did this opportunistically; 14.3% when there was a history of prostate problems; 3.9% on patient request; 10.8 % with lower urinary tract symptoms (Figure 2). GP questionnaire results Of the 26 GPs in the five practices, 18 questionnaires were returned (69%): 10 male and 8 female. The majority worked five or more sessions in joint practices and were aged in their 40s and 50s. Sixty-one percent agreed or strongly agreed that PSA screening reduced mortality rates. Fifty-five percent were concerned regarding harm caused by PSA testing, but the majority felt the benefits outweighed this. All bar one GP did PSA screening: mostly selectively and/or opportunistically, mainly focused on men aged between 50 and 202 All doctors in the five general practices involved in the study were testing asymptomatic patients. This was confirmed by the GP questionnaire, in which all bar one GP said they practised screening. Testing of asymptomatic men is common in New Zealand and this seems to be consistent with findings in other countries.15–17 Many studies have asked GPs and primary care doctors for their views regarding screening of an asymptomatic patient; we have gone one step further in this study and identified why men are tested. In our study, 25% of men 40 years and over were tested in 2010. Testing rates increased with age and GPs focused most of their screening on men aged between 50 and 69 years. This may in fact be worthwhile as the ERSPC study showed up to 20% reduction in mortality in men in this age group.8 What is of concern in our study is the number of PSA tests done on patients aged 70+ years. A large number of these tests (56%) were done opportunistically when there is no evidence to support a mortality benefit in this age group. The National Screening Advisory Committee identified that nearly 50% of men 40 years and over had had a PSA test at some time, which compared to only 18% during 2008.21 In a 2011 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Health Committee inquiry into the early detection and treatment of prostate cancer, 50% of men 50 years and over were estimated to have a PSA test over a two-year period.22 The New Zealand Health Survey suggests that approximately 50% of men aged 50 and over had a PSA test in the previous five years, of which 80% were asymptomatic.23 This represented a 40% coverage rate in New Zealand in any one year. Our study found that there was a 25% chance of having a PSA test in 2010 if 40 years and over. If we limit it to 50 years and over the chance is 31%. These figures are consistent with the above-mentioned reports. Our GPs generally believed that the benefits of PSA testing outweighed harm and resulted in reduced mortality rates. Ten percent (147/1480) of men who had a PSA tests in 2010 showed an elevated level of PSA. Management of these patients reflect the uncertainty over the value of PSA testing as a screening tool. Fifty-five out of 147 (37%) men were referred to a specialist (mostly urologist), which at face value seems to be a low referral rate. It seems that many men with raised PSA tests are initially managed with repeated PSA tests, rather than proceeding directly to referral. As this study had only a limited follow-up period we cannot comment on how many patients may eventually require a referral and biopsy. 1000 PSA tests done for patients with previous prostate problems (including previously elevated PSA), approximately 66 new prostate cancers would be identified. There would be 18 new cancers for patients presenting with LUTS, while only four new cancers for every 1000 tests done opportunistically (i.e. those without symptoms or previous prostate problems). Therefore, when diagnosing prostate cancer it seems worthwhile if GPs focus their PSA testing on men with previous history of prostate problems or LUTS. GPs within our study reported that they find explanation to patients about PSA screening difficult. Most seem convinced that screening is beneficial although evidence supporting improved mortality rates in the 70+ years age groups is lacking. More education for GPs may prove worthwhile. One of the strengths of this study was that it was population based and that the researchers were able to link patient data with laboratory data. Collection of detailed data directly from clinical notes by the same researchers, one of whom was a clinician, is considered a strength. Seeking GP views via a questionnaire and relating these findings to practice was likewise believed to be a positive strength. In our sample, we observed that younger patients (40–69 years) were more likely to be referred (42.4%) and to have a biopsy (79.5%) than older men (29.1% referred, and 31.3% biopsied, respectively). In total, 39 out of 55 men with elevated PSA (71%) underwent a biopsy; 21 new cases of prostate cancer were identified. In addition, there were 10 referrals to a specialist in the absence of raised PSA. There seemed to be reasonable clinical grounds (LUTS and/or abnormal DRE) to do so. Two of these 10 men were biopsied, both of whom proved to have cancer. Weaknesses of this study could include the selection of practices which might not reflect a true representation of current practice. We also had lower numbers of Maori men than we were hoping to find, precluding analysis by ethnicity. GP questionnaire numbers were small and may not be representative of the New Zealand population. This study did not look at follow-up after biopsy and any risk or benefits from the diagnosis of prostate cancer. These aspects are being followed up by the larger, three-year Midland Prostate Cancer Study of which this study represents the pilot stage. We looked at the reasons for having the PSA test done and checked whether this was helpful in identifying men with prostate cancer. Having a previously raised PSA or history of a prostate problem together with a raised PSA proved to be the most productive in identifying prostate cancer. From our study it is estimated that for every In conclusion, this study looked at PSA testing by GPs in five practices in the local Waikato setting. GPs in this study believed in the benefits of screening and were opportunistically testing, focusing their screening on men aged 50–69 years. They were more likely to refer and there was a greater chance of identifying prostate cancer VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 203 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH when PSA testing was done on men with previous history of prostate problems or LUTS. Using their clinical judgment, GPs identified further prostate cancer cases even when PSA levels were within normal ranges. This study has provided useful findings that are informing a larger study on the management of men with prostate cancer. References FUNDING The authors would like to thank their funders, the Waikato Medical Research Foundation (project number: 003626510) for their support of the project. COMPETING INTERESTS None declared. 204 1. Ministry of Health. New registrations and deaths 2006. Revised edition. Wellington; 2010. 2. Hardie C, Parker C, Norman A, Eeles R, Horwich A, Huddart R, Dearnaley D. Early outcomes of active surveillance for localized prostate cancer. BJU Int. 2005;95:956–60. 3. Roemeling S, Roobol MJ, Postma R, Gosselar C, van der Kwast TH, Bangma CH, Schröder FH. Management and survival of screen-detected prostate patients who might have been suitable for active surveillance. Eur Urol. 2006;50:475–82. 4. Roemeling S, Roobol MJ, De Vries SH, Wolters T, Gosselaar C, van Leenders GJLH, Schröder FH. Active surveillance for prostate cancers detected in three subsequent rounds of a screening trial: characteristics, PSA doubling times, and outcome. Eur Urol. 2007;51:1244–50. 5. Greene KL, Albertsen PC, Babaian RJ, Carter HB, Gann PH, Han M, Kuban DA, Sartor AO, Stanford JL, Zietman A, Carroll P. Prostate Specific Antigen Best Practice Statement: 2009 Update. J Urol. 2009;182:2232–41. 6. Urological Society of Australia and New Zealand PSA Testing Policy 2009. [cited 2011 Nov 9]. Available from: www.usanz. org.au/uploads/29168/ufiles/USANZ_2009_PSA_Testing_Policy_Final1.pdf 7. Andriole GL, Crawford ED, Grubb RL, Buys SS et al. Mortality results from a randomized prostate-cancer screening trial. N Engl J Med. 2009;360:1310–9. 8. Schröder FH, Hugosson J, Roobol MJ, Teuvo LJ et al. Screening and prostate-cancer mortality in a randomized European study. N Engl J Med. 2009;360:1320–8. 9. New Zealand Guidelines Group. Testing for prostate cancer. Information for men and their families. Ministry of Health; 2008. 10. Evans R, Edwards AGK, Elwyn G, Watson E, Grol R, Brett J, Austoker J. ‘It’s a maybe test’: men’s experiences of prostate specific antigen testing in primary care. Br J Gen Pract. 2007;57(537):303–10. 11. Arroll B, Pandit S, Buetow S. Prostate cancer screening: knowledge, experiences and attitudes of men aged 40–79 years. N Z Med J. 2003;116(1176):U477. 12. Avery KN, Blazeby JM, Lane JA, Neal DE, Hamdy FC, Donovan JL. Decision-making about PSA testing and prostate biopsies: a qualitative study embedded in a primary care randomised trial. Eur Urol. 2008;53(6):1186–93. 13. Best Practice Advocacy Centre (BPAC). Annual Pharmaceutical & Laboratory Report; 2009. 14. Drummond FJ, Carsin AE, Sharp L, Comber H. Factors prompting PSA-testing of asymptomatic men in a country with no guidelines: a national survey of general practitioners. BMC Fam Pract. 2009;10:3. 15. Durham J, Low M, McLeod D. Screening for prostate cancer: a survey of New Zealand general practitioners. NZ Med J. 2003;116(1176). 16. Gavin A, McCarron P, Middleton RJ, Savage G, Catney D, O’Reilly D, Keane PF, Murray LJ. Evidence of prostate cancer screening in a UK region. BJU Int. 2004;93:730–4. 17. Hicks RJ, Hamm RM, Bemben DA. Prostate cancer screening. What family physicians believe is best. Arch Fam Med. 1995;4(4):317–22. 18. Jonler M, Eddy B, Poulsen J. Prostate-specific antigen testing in general practice: a survey among 325 general practitioners in Denmark. Scand J Urol Nephrol. 2005;39:214–8. 19. Little B, Ho KJ, Gormley G, Young M. PSA testing in general practice. Prostate Cancer Prostatic Dis. 2003;6(2):154–8. 20. Pool I, Baxendine S, Cochrane W, Lindop J. New Zealand regions, 1986–2001: population geography. Population Studies Centre Discussion Papers 54; October 2005. 21. Sneyd DMJ. Prevalence of opportunistic prostate cancer screening in New Zealand. A report for the National Screening Advisory Committee; 2010. 22. Inquiry into early detection and treatment of prostate cancer. Report of the Health Committee; 2011. 23. Ministry of Health. A portrait of health: key results of the 2006/07 New Zealand Health Survey. Wellington; 2008. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH The incidence of acute otitis media in New Zealand children under five years of age in the primary care setting Barry Gribben MBChB, MMedSci, FRNZCGP;1 Lesley J Salkeld MBChB, FRACS, FRCSC, DipABO;2 Simon Hoare MBChB, DCH, MD, FRACP, FRCPCH;3 Hannah F Jones MBChB, Dip Paed4 1 CBG Health Research Limited, Auckland, New Zealand ABSTRACT INTRODUCTION: Acute otitis media (AOM) is a common childhood infection. Baseline data are required to evaluate potential changes in the epidemiology of AOM with new public health measures. 2 Paediatric Otolaryngology, Starship Hospital, Auckland 3 AIM: To estimate the incidence of AOM in children under five years of age in primary care in New Zealand. METHODS: Using a cohort study design, consultation notes from 1 November 2008 to 31 October 2009 from 63 primary care facilities were analysed for new and recurrent episodes of AOM, complications, antimicrobial use and outcome. Paediatrics, North Shore Hospital, Auckland 4 Paediatrics, Starship Hospital, Auckland RESULTS: There were 19 146 children in the sample. The raw incidence of AOM was 273 per 1000 children (27.3%; 95% CI 216–330). Of the 3885 children, 2888 (74%) had one episode of AOM and 152 (4%) of these children developed recurrent AOM. Incidence declined with age. There was no difference in incidence between Maori, Pacific and ‘Other’ ethnicities. Antibiotics were used to treat 2653 (51%) AOM episodes and 113 (4.3%) of these children re-presented within three days of antibiotic therapy for persistent symptoms. Tympanic membrane perforation was the only complication noted, observed in 62 (1%) episodes. DISCUSSION: These data indicate that AOM is an important and frequent childhood infection in New Zealand. The show a significant decline in the use of antibiotics to manage AOM in concordance with accepted best practice. The complication rate of AOM is likely under-represented. This study enables future research into the effectiveness of current and future immunisations and changing management practices in New Zealand. KEYWORDS: Otitis media; incidence; child, preschool; New Zealand, antibiotic Introduction Acute otitis media (AOM) remains a frequent global infection of childhood, with up to 80% of children having at least one episode by three years of age.1 Ten to 30% have recurrent episodes,2,3 and 2–25% will have persistent middle ear effusion extending beyond three months,1 many of which require tympanostomy tube insertion.4,5 Persistent tympanic membrane perforation and otorrhoea are significant issues, particularly in specific populations.5–7 Otitis media and eustachian tube disorders are the third leading diagnosis in GP consultations between 0 and 14 years,8 and myringotomy and tube insertion the most frequent childhood surgical procedure in the USA.9 AOM is also the commonest reason for prescribing antibiotics to children in developed countries,2 contributing to concerns of emerging antibiotic resistance.6 The current incidence of AOM in New Zealand is not known. Tilyard et al. reviewed otitis media (OM), including AOM, treatment in general prac- VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):205–212. CORRESPONDENCE TO: Hannah F Jones Paediatric Registrar, Starship Hospital, PB 92024, Auckland 1142, New Zealand. hannahj@adhb.govt.nz 205 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Table 1. Geographic distribution and proportion of subjects in the study population by District Health Board Study population n Population under five years of age N Proportion of the DHB population included in study (%) Auckland 1480 28680 5 Bay of Plenty 1697 14410 12 Canterbury 780 33290 2 District Health Board Capital & Coast 136 19440 1 Counties Manukau 2686 41200 7 Hawkes Bay 1824 11290 16 Hutt Valley 877 10620 8 Lakes 466 8050 6 MidCentral 354 11470 3 Nelson Marlborough 461 8545 5 Northland 1357 11240 12 Otago 1771 10550 17 South Canterbury 267 3285 8 Southland 0 7890 0 Tairawhiti 1015 3905 26 Taranaki 294 7675 4 Waikato 1097 27070 4 Waitemata 1950 37290 5 Wairarapa 0 2680 0 West Coast 395 2125 19 239 4370 5 19146 305075 Whanganui All tice in New Zealand from 1993 to 1994 across all age groups,10 and Ryan et al. surveyed the management of AOM by GPs further in 2002.11 Historical studies show rural Maori children have some of the highest rates of chronic suppurative otitis media (CSOM) in the world.12–14 A more recent New Zealand study demonstrated an incidence of 3871 hospital admissions for OM per 100 000 of the population of children under five years of age, with higher rates of medical admissions amongst Maori and Pacific children.15 The incidence of AOM may have been affected by the introduction of a pneumococcal conjugate vaccine (PCV7) to New Zealand in June 2008. There has been a modest reduction in incidence of all-cause AOM in clinical trials of this vaccine in other countries.16,17 A new pneumococcal conjugate vaccine that targets the two predominant bacterial pathogens of AOM (Streptococ- 206 cus pneumoniae and non-typeable Haemophilus influenzae) has been correlated with a greater effect based on an AOM clinical efficacy study with a precursor formulation.18 This new vaccine was due to replace the PCV7 immunisation in the NZ immunisation schedule in 2011. The term ‘acute otitis media’ as used throughout this paper can be interpreted as symptomatic, rapidly forming purulent fluid in the middle ear cleft which occasionally results in rupture of the previously intact eardrum causing discernable otorrhoea. ‘Otitis media with effusion’ (OME), meaning inflammation associated with fluid behind an intact eardrum without signs and symptoms of an acute infection, is not addressed in this study. ‘Chronic suppurative otitis media’, where alluded to, describes dry perforations, draining cholesteatomas or chronic discharge via a perforation. ‘Otitis media’ is used generally to VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH describe inflammation of the middle ear without regard to aetiology. In this study, we have examined the current incidence of AOM in New Zealand preschool children presenting to primary care. We also assessed clinical features (including secondary complications and recurrent disease), antimicrobial use, and treatment failures. These data are important for resource management and establishing the impact of current and future vaccines. Methods This cohort study reviewed children less than five years of age enrolled in a sample of general practitioner (GP) practices in New Zealand from 1 November 2008 to 31 October 2009. The study used the random sample of GP practices enrolled in HealthStat, an established primary care surveillance system in New Zealand (See Appendix 1 in the web version of this paper).19 The 93 GP practices participating in the HealthStat surveillance panel at the time of the study were invited to take part. Sixty-seven of these practices responded, 63 practices accepted the invitation and four declined. These 63 practices had 261 835 patients enrolled at the time of the study. Children were eligible for the study if they were under five years of age on 1 November 2008 and were continually enrolled in a participating practice from 1 November 2008 to 31 October 2009; 19 146 eligible children were retrospectively enrolled in the study. The geographical distribution of the study population by District Health Boards (DHBs) is detailed in Table 1. Nineteen of the 21 DHBs were represented in the study. Anonymous data are automatically downloaded weekly from the HealthStat practice databases, using an encrypted messaging system. These data include dates of all consultations, diagnostic coding (and other classifications if recorded by clinicians), prescriptions, laboratory results and demographic data. Although up to three ethnicities were collected for each patient, only one ethnicity was coded per patient with precedence given for Maori, then Pacific, then ‘Other’. WHAT GAP THIS FILLS What we already know: Acute otitis media (AOM) represents significant disease burden amongst children worldwide. Antibiotics are frequently prescribed for AOM although in most clinical scenarios this is no longer considered best practice. New immunisations targeted at common AOM pathogens have been associated with a decreased incidence of AOM in international studies. What this study adds: This study presents an estimate of the incidence of AOM in children in New Zealand, and the demographics of the affected population and the management of AOM in primary care. It provides a baseline for future analyses of the impact of existing and new immunisations on the incidence of AOM and of changes in the management of AOM in primary care in New Zealand. Data on the incidence of AOM were collected by a review of the text of clinical notes written during the study period. All notes with terms that either related to otological disease (e.g. ‘AOM’, ‘OM’, ‘bulging’) or to an otological history or examination (e.g. ‘ear’, ‘drum’, ‘tymp’) were electronically selected for analysis: a list of the search terms is given in Appendix 2 in the web version of this paper. These notes were then reviewed by a team of research nurses. Ambiguous terms were interpreted in context, e.g. ‘tymps’ was either understood to mean tympanic membrane(s) or tympanogram. In addition to these records, a random sample of 100 records from each practice was reviewed for idiosyncratic abbreviations that indicated a possible otological problem. When a new abbreviation was found, all notes from all the practices were re-examined for the new term. Episodes of AOM were coded for diagnostic certainty according to the clinical evidence provided (Table 2). If a research nurse had difficulty coding a consultation in terms of diagnosis or diagnostic certainty, the consultation record was examined by the senior research nurse and, if necessary, a GP. All consultation notes reviewed by the research nurses were examined to determine whether the case related to a new episode of AOM, recurrent AOM, ‘treatment failure’ or complicated AOM. A new episode of AOM was defined VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 207 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Table 2. Ranking of criteria for the identification of acute otitis media based on terms used in the consultation notes Level* Criteria Level 1 AOM subjects defined as children (<5 yrs) with diagnosis of AOM/OM. Level 2 AOM subjects defined as children (<5 yrs) with diagnosis of AOM/OM and at least two clinical signs or symptoms: fever, otalgia, bulging tympanic membranes, or otorrhoea. Level 3 AOM subjects defined as children (<5 yrs) with diagnosis of AOM/ OM and/or clinical signs or symptoms (fever, otalgia, bulging tympanic membrane, otorrhoea) and microbiological results (S.pneumoniae culture positive, H. influenzae culture positive, serotype results if available). assessed no improvement in symptoms after 48 hours of antibiotic therapy. From the collected data, the raw incidence of new episodes of AOM was calculated using the following equation: Incidence = Number of new AOM episodes in children <5 years of age during the study period Total number of children <5 years of age in the study cohort from 2008 to 2009 *The higher the level, the greater the degree of clinical evidence supporting the diagnosis Table 3. Age, gender and ethnicity demographics of patients in study sample, with comparison to the New Zealand population Study population NZ population 2009 N (%) N (%) 23 109 (8.8)* 362 858 (8.4) Male 9730 (50.9) 156 465 (51.3) Female 9416 (49.2) 148 610 (48.7) p Further data analyses were performed using SAS 9.2. The confidence intervals were calculated using SAS proc surveymeans (using the Wald method) to estimate the average number of AOM per child, allowing for clustering of children within practices. Age 0–5 Results Gender 0.1982 Ethnicity Maori 6358 (33.2) 86 930 (28.5) Pacific 1485 (7.8) 33 735 (11.1) Other 11 303 (59.0) 18 4410 (60.4) 0.0001 * Includes children five years old at 1 November 2008 who were excluded from the final study sample. as an episode of AOM following a 30-day symptom-free interval. Therefore consultation notes from 1 October 2008 to 31 October 2008 were also reviewed to ensure these patients had not been consulted for middle ear symptoms in the 30 days prior to diagnosis. Recurrent AOM described three or more new episodes of AOM within six months or four or more new episodes of AOM within one year. ‘Complicated AOM’ was limited to complications which arose during the study period. Perforations were only included as a complication if there was AOM and a visible perforation on the same day. A tympanostomy tube was not regarded as a perforation. Otorrhoea alone or in the presence of a tympanostomy tube was not considered a complication. The prevalence of tympanostomy tube insertions was beyond the scope of this study. A case was considered ‘treatment failure’ if the physician 208 Between 1 November 2008 and 31 October 2009, 19 146 children younger than five years old were identified in participating practices representing 19 out of 21 DHBs. The demographics of the study versus the New Zealand population are shown in Table 3. The number of consultations for AOM identified was 6261. Of these, 5225 related to new episodes of AOM. Hence, the incidence of AOM was 273 episodes per 1000 (27.3%) children under five years of age per annum (95% CI 216–330). The incidence definition includes multiple episodes of new AOM for the same child. The majority of children (74%) affected by AOM had only one episode of AOM during the study period (Table 4). Approximately 20% of children had two cases of AOM and nearly 7% had three or more cases in the 12-month analysis. These episodes were stratified by age, gender and ethnicity (Table 5). The incidence was highest in children less than one year old and significantly declined with age. There were no significant ethnic group–related differences. Eighty-two percent of episodes met level 2 diagnostic criteria. Less than 1% of cases fulfilled level 3 diagnostic criteria, which required a microbiological result obtained from otorrhoea or tympanocentesis. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Table 4. Number of new episodes of acute otitis media per child from 1 November 2008 to 31 October 2009 New episodes of AOM Number of children (N) Total number of episodes (n) Percentage of children with number of episodes (%) 1 2888 2888 74.3 2 731 1462 18.9 3 199 597 5.1 4 58 232 1.5 5 8 40 0.2 6 1 6 0.0 3885 5225 All Recurrent AOM was identified in 152 children, comprising 4% of all the children who had at least one episode of AOM (Table 6). Recurrent AOM was most frequently diagnosed in younger subjects; 6.5% children under one year of age developed recurrent AOM in comparison to less than 1% of four-year-old children. Fifty-one percent of the AOM episodes were treated with antibiotics. In 4.3% of treated episodes, the child re-presented with persistent symptoms within three days of their initial presentation. The only complication noted in this study was visible perforation, present in 62 (1%) episodes. Discussion This retrospective study is the first major analysis of AOM incidence in young New Zealand children. It emphasises the major frequency, impact and socioeconomic burden of AOM. Two hundred and seventy-three episodes per 1000 children per annum is an incidence comparable Table 5. Incidence of AOM by age, gender and ethnicity Number of children (N) Number of new AOM episodes (n) Percentage incidence of episodes (%) 95% CI (%) Age* <1 3485 1740 49.9 39.8-60.0 1 4091 1377 33.7 25.8-41.4 2 3951 836 21.2 16.0-26.2 3 3733 688 18.4 13.7-23.1 4 3886 584 15.0 12.1-18.0 All 19146 5225 27.3 21.6-33.3 9416 2496 26.5 21.0-31.9 Gender Female Male 9730 2729 28.0 22.1-32.8 All 19146 5225 27.3 21.6-33.3 11303 3187 28.2 21.1-35.3 Ethnicity Other Maori 6358 1721 27.1 20.7-33.4 Pacific 1485 317 21.3 12.9-29.8 All 19146 5225 27.3 21.6-33.3 * At 1 November 2008 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 209 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH to other developed countries, although differences in study methodology need to be considered.1,3,20–25 This study includes the first six months of PCV7 immunisation in New Zealand, but predates the introduction of vaccines with increased serotype coverage and activity against other pathogens such as non-typeable Haemophilus influenzae. The gender distribution of the sample was comparable to the New Zealand population, but the ethnicity distribution significantly differed, with Maori being over-represented and Pacific children under-represented. However, calibration of the raw incidence to the New Zealand population by gender and ethnicity has minimal effect on the estimated incidence (i.e. 271/1000 children per annum). Applying Approximately half of the AOM episodes were treated with antibiotics. This is a significant decrease from a 96.6% prescription rate in Tilyard’s New Zealand study from 1993 to 1994, and the 95% rate self-reported by GPs in Ryan’s study in 2002. this rate to the New Zealand population (July 2009 projection) suggests there are 83 000 new episodes of AOM annually in children less than five years of age. The majority of diagnoses (82%) were well supported by at least two documented signs or symptoms, an accepted degree of diagnostic certainty across studies.1,17,26 There were only three cases of Level 3 diagnostic certainty, as tympanocentesis (penetration of the tympanic membrane to aspirate middle ear contents for diagnostic purposes) is not a routine procedure in primary care in New Zealand. The decline in incidence of new episodes of AOM with age is consistent with international studies.1,16,20,27 There was no significant difference in the incidence of AOM across ethnic groups 210 in this study. This was an unexpected finding as risk factors for otitis media (such as exposure to second-hand smoke and lower breastfeeding rates)28 are more prevalent in the Maori population, and a high prevalence of CSOM in Maori has been previously reported.12–14 Maori children have a greater ‘unmet need’ for GP services than non-Maori,28 which may have contributed to lower AOM rates than anticipated. For children of Pacific ethnicity, recent research showed 1.9% of two-year-old Pacific children investigated for chronic otitis media were incidentally diagnosed with AOM.29 The incidence of recurrent acute otitis media was lower than reported in other studies, but an inverse association between recurrent AOM and age was still demonstrated.1,2 One explanation for the lower incidence is that children in the study, although enrolled in the participating GP practices, may have presented to other general practices throughout the year. The occurrence of complicated acute otitis media may be under-represented in this study. ‘Mastoiditis’ and other complications were not search terms used to detect AOM; therefore, complications were only identified if they were simultaneously present with features of AOM during the study period. In addition, New Zealand children with complex disease (e.g. mastoiditis, absence of clinical improvement, out-of-hours deterioration) often present directly to non-GP services such as hospital emergency departments. The only complications identified were 62 visible perforations. In some instances these may have been chronic perforations with acute otorrhoea. Local data from Starship Hospital (the referral centre for acute otorhinolaryngology services for approximately one quarter of New Zealand’s paediatric population) indicate there were 59 admissions consistent with mastoiditis in children less than five years of age from 1 January 2005 to 31 December 2009.30 The incidence was disproportionately represented amongst Maori and Pacific populations (30.5% (18/59) New Zealand Maori, 27.1% (16/59) Pacific, 42.4% (25/59) ‘Other’) compared with population statistics 28.5% Maori, 11.1% Pacific, 60.4% ‘Other’.30 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Table 6. Number of cases of recurrent AOM by age Recurrent episodes Number of children with three AOM episodes in six months (N) Number of children with four AOM episodes in one year (N) Number of children with non-recurrent AOM (N) Proportion of children with AOM diagnosed with recurrent AOM (%) <1 38 38 1085 6.5 1161 1 27 18 962 4.5 1007 2 12 6 651 2.7 669 3 7 2 546 1.6 555 Age Total number of children with ≥1 episode of AOM (N) 4 1 3 489 0.8 493 All 85 67 3733 3.9 3885 Approximately half of the AOM episodes were treated with antibiotics. This is a significant decrease from a 96.6% prescription rate in Tilyard’s New Zealand study from 1993 to 1994, and the 95% rate self-reported by GPs in Ryan’s study in 2002.10,11 This decreasing trend complies with evidence for the management of AOM in developed countries with the exception of children less than six months of age for whom antibiotics are still recommended.26,31 It also compares favourably with international antibiotic prescription rates.20,23 Starship Children’s Hospital data over five years does not reflect an increasing rate of mastoiditis with the decrease in antibiotic therapy.30 All antibiotics prescribed were presumed to be oral antibiotics. The use of topical antibiotics as an adjunct or alternative to oral antibiotics or the presence of concurrent otitis externa was not reviewed. Limitations of this retrospective study include the lack of independent assessment of the diagnoses and the bias inherent in reviewing and interpreting medical records. Although over 90% of children nationally are able to see their GP when needed,28 some children may have presented to emergency departments, mobile ‘ear clinics’, or general practices not registered as their primary health care provider. This would lead to an underestimation of AOM incidence and related complications. Only two-thirds of the invited general practices agreed to participate in the study. This is likely because of the short recruitment period, as study participation rates are typically around 90%. Some practices do not routinely accept study invitations before they have been discussed at practice meetings, which may occur only quarterly. In spite of this, our study sample represented over 6% of the New Zealand population. This study demonstrates the importance of AOM in New Zealand children and creates baseline data to enable future analysis of the effect of changing management practices, including the effectiveness of current and future immunisations on the burden of AOM disease. References 1. Teele DW, Klein JO, Rosner B. Epidemiology of otitis media during the first seven years of life in children in Greater Boston: a prospective, cohort study. J Infect Dis. 1989;160(1):83–94. 2. Rovers MM, Schilder AGM, Zielhuis GA, Rosenfeld RM. Otitis media. The Lancet. 2004;363(9407):465–73. 3. Liese J CA, Cantarutti L, Silfverdal S-A, Fuat A, Vollmar J, Pircon J-Y, Rosenlund M. Incidence of acute otitis media in young children seen in European medical practices. 6th World Congress of the World Society for Pediatric Infectious Diseases. 18–22 November, 2009. 4. American Academy of Family Physicians, OtolaryngologyHead and Neck Surgery, American Academy of Pediatrics Subcommittee on Otitis Media With Effusion. Otitis media with effusion. Pediatrics. 2004;113(5):1412–29. 5. Morris PS, Leach AJ. Acute and chronic otitis media. Pediatr Clin North Am. 2009 Dec;56(6):1383–99. 6. Vergison A, Dagan R, Arguedas A, Bonhoeffer J, Cohen R, Dhooge I, Hoberman A, Liese J, Marchisio P, Palmu AA, Ray GT, Sanders EA, Simoes EA, Uhari M, van Eldere J, Pelton SI. Otitis media and its consequences: beyond the earache. Lancet Infect Dis. 2010 Mar;10(3):195–203. 7. O’Connor TE, Perry CF, Lannigan FJ. Complications of otitis media in indigenous and non-indigenous children. Med J Aust. 2009 Nov 2;191(9 Suppl):S60–4. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 211 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH FUNDING We wish to acknowledge GlaxoSmithKline (GSK) for providing the funding for this study. GSK manufactures the new pneumococcal conjugate vaccine which targets the two predominant bacterial pathogens of AOM (Streptococcus pneumoniae and nontypeable Haemophilus influenzae). COMPETING INTERESTS None declared. 212 8. Top 5 diagnoses at visits to office-based physicians and hospital outpatient departments by patient age and sex: United States, 2008. National Ambulatory Medical Care Survey and National Hospital Ambulatory Medical Care Survey. In: Centers for Disease Control and Prevention NCHS, editor. Atlanta 2008. 9. Owings MF, Kozak LJ. Ambulatory and inpatient procedures in the United States, 1996. Vital Health Stat. 1998 Nov;(139):1–119. 10. Tilyard MW, Dovey SM, Walker SA. Otitis media treatment in New Zealand general practice. N Z Med J. 1997 Apr 25;110(1042):143–5. 11. Ryan J, Giles M. Management of acute otitis media by New Zealand general practitioners. NZ Med J. 2002 Feb 22;115(1148):67–9. 12. Giles M, O’Brien P. Otitis media and hearing loss in the children of the Ruatoki valley: a continuing public health problem. NZ Med J. 1989 Apr 12;102(865):160–1. 13. Giles M, Asher I. Prevalence and natural history of otitis media with perforation in Maori school children. J Laryngol Otol. 1991 Apr;105(4):257–60. 14. Giles M, O’Brien P. The prevalence of hearing impairment amongst Maori schoolchildren. Clin Otolaryngol Allied Sci. 1991 Apr;16(2):174–8. 15. Milne RJ, Van der Hoorn S. Burden and cost of hospital admissions for vaccine-preventable paediatric pneumococcal disease and non-typable Haemophilus influenzae otitis media in New Zealand. Appl Health Econ Health Policy. 2010;8(5):281–300. 16. Fireman B, Black SB, Shinefield HR, Lee J, Lewis E, Ray P. Impact of the pneumococcal conjugate vaccine on otitis media. Pediatr Infect Dis J. 2003 Jan;22(1):10–6. 17. Eskola J, Kilpi T, Palmu A, Jokinen J, et al. Efficacy of a pneumococcal conjugate vaccine against acute otitis media. N Engl J Med. 2001;344(6):403–9. 18. Prymula R, Peeters P, Chrobok V, Kriz P, Novakova E, Kaliskova E, Kohl I, Lommel P, Poolman J, Prieels J-P, Schuerman L. Pneumococcal capsular polysaccharides conjugated to protein D for prevention of acute otitis media caused by both Streptococcus pneumoniae and non-typable Haemophilus influenzae: a randomised double-blind efficacy study. Lancet. 2006;367(9512):740–8. 19. HealthStat Primary Healthcare Intelligence. Available from: http://www.healthstat.co.nz/. 20. Plasschaert AI, Rovers MM, Schilder AG, Verheij TJ, Hak E. Trends in doctor consultations, antibiotic prescription, and specialist referrals for otitis media in children: 1995–2003. Pediatrics. 2006 Jun;117(6):1879–86. 21. Morris PS, Leach AJ, Silberberg P, Mellon G, Wilson C, Hamilton E, Beissbarth J. Otitis media in young Aboriginal children from remote communities in Northern and Central Australia: a cross-sectional survey. BMC Pediatr. 2005;5:27. 22. Wolleswinkel-van den Bosch JH, Stolk EA, Francois M, Gasparini R, Brosa M. The health care burden and societal impact of acute otitis media in seven European countries: results of an Internet survey. Vaccine. 2010 Nov 19;28 Suppl 6:G39–52. 23. Grossman Z, Silverman BG, Porter B, Miron D. Implementing the delayed antibiotic therapy approach significantly reduced antibiotics consumption in Israeli children with first documented acute otitis media. Pediatr Infect Dis J. 2010 Jul;29(7):595–9. 24. Thomas EM. Recent trends in upper respiratory infections, ear infections and asthma among young Canadian children. Health Rep. 2010 Dec;21(4):47–52. 25. Williamson I, Benge S, Mullee M, Little P. Consultations for middle ear disease, antibiotic prescribing and risk factors for reattendance: a case-linked cohort study. Br J Gen Pract. 2006 Mar;56(524):170–5. 26. Subcommittee on Management of Acute Otitis Media. Diagnosis and management of acute otitis media. Pediatrics. 2004;113(5):1451–65. 27. Shaikh N, Hoberman A. Update: acute otitis media. Pediatr Ann. 2010 Jan;39(1):28–33. 28. Ministry of Health. A focus on the health of Maori and Pacific children: key findings of the 2006/07 New Zealand Health Survey. In: Ministry of Health, editor. Wellington: Ministry of Health; 2009. 29. Paterson JE, Carter S, Wallace J, Ahmad Z, Garrett N, Silva PA. Pacific Islands families study: the prevalence of chronic middle ear disease in two-year-old Pacific children living in New Zealand. Int J Pediatr Otorhinolaryngol 2006 Oct;70(10):1771–8. 30. Data courtesy of Decision Support CRD, Starship Children’s Hospital, Auckland. 2010 Data obtained August, 2010. 31. Glasziou PP, Del Mar CB, Sanders SL, Hayem M. Antibiotics for acute otitis media in children. Cochrane Database Syst Rev 2004(1):CD000219. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH Does the order of presentation and number of online resources affect the frequency of access by learners? Steven Lillis MBChB, FRNZCGP, MGP, PhD;1 Samantha Murton MBChB, FRNZCGP2 1 ABSTRACT INTRODUCTION: Provision of web-based resources is a valuable addition to face-to-face teaching in a blended learning environment. AIM: To understand how both order of presentation and number of online resources impacts on the frequency of access by learners in postgraduate vocational training in general practice. Department of General Practice & Primary Health Care, Waikato Clinical School, University of Auckland, Hamilton, New Zealand 2 The Royal New Zealand College of General Practitioners, Wellington, New Zealand METHODS: Information was collected on how many times individual online resources were accessed. Data regarding access rates for 15 separate topics used in postgraduate general practice vocational training were aggregated. Analysis was on the basis of order of presentation where the mean of percentages of hits by order of presentation with standard deviations was calculated. RESULTS: The first four listed resources were accessed at a higher rate than the remainder of the resources. All resources after the first four were accessed at a relatively uniform low rate. DISCUSSION: It would appear that providing more than four resources per topic is associated with learner overload. The number of online resources to support face-to-face teaching should be limited to four. Resource material needs to be carefully considered in terms of how it adds educational value. The ability of resource material to present a different perspective on a topic and adherence to both curriculum and assessment objectives are important considerations. KEYWORDS: Education, medical; education, distance; internet; provision of resources; general practice Introduction Online learning has created diverse and innovative delivery mechanisms and an active education research stream.1 Outcomes research on internet-based learning suggest that enhanced learning opportunities may be generated, but there are no time efficiencies gained.2 One enhanced learning opportunity of e-learning is the ease of placing education material within reach of the learner and the self-paced nature of subsequent learning.3 The availability of online educational material does not, however, predict how this material will be used.4 Issues relevant to the use of resources within online resources include the level of literacy necessary to incorporate the available material, the complexity and quantity of material provided and the motivation of students to use the material.5 Many medical education courses now utilise both face-to-face small group learning and internet-based learning within an integrated and mutually supportive framework (blended learning). The creation, sharing and sourcing of this e-content is receiving increasing focus.6 Guidelines regarding content of e-learning within this blended environment are available, but tend to provide broad rather than specific information, such as not overloading the learner with content.7 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):213–216. CORRESPONDENCE TO: Steven Lillis Director of Assessment, Faculty of Medicine and Health Science, University of Auckland, PB 92019, Auckland, New Zealand slillis@wave.co.nz 213 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH There is little specific data on how learners access online resource content within a blended structure. Educators therefore lack important data that informs on the effect of volume or organisation of online resource content. This research reports on rates and patterns of access to online resources by learners in a blended learning environment. Methods All general practice registrars in The Royal New Zealand College of General Practitioners (RNZCGP) training scheme enter a year of intensive training that utilises a blended learning environment. Each clinical topic (adolescent medicine, for example) has both a set of online resources chosen by medical educators as well as several hours of structured interactive face-to-face learning with discussion, case studies, guest speakers etc. The order of presentation of the online resources within each clinical topic is unplanned. The online resources are available through a Moodle platform. Moodle is an open source learning management system that facilitates content access and communication for web-based teaching and learning. The number of online resources per clinical topic is variable. Data are collected on each topic, recording the number of times each of the online resources for that topic is accessed. There were 144 registrars in the programme with approximately 70% being female in the 2011 year in which the data were collected. Data were available for 15 topics and were aggregated and anonymised. Each access to, or download of, a resource by a registrar was considered as a ‘hit’. For each scaffold, the number of hits for individual resources was calculated as a percentage of all hits for that scaffold. Results There were between five and 32 resources available for each of 15 clinical topics, with a mean of 15.5 (SD of 6.9). The overall number of hits per clinical topic varied considerably (487 to 2136). The aggregated data on hits for the top 15 resources across all clinical topics is presented in Figure 1. The mean of the aggregated data over 15 topics is given with one standard deviation on Figure 1. Level of access to the 15 most accessed online resources across all clinical topics based on order of presentation 214 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH each side. The first listed resource accounts for just under 16% of all hits, with an SD of 7%. The first four resources accounted for 10% or more of all hits each. Thereafter, the percentage of hits that each resource attracted was lower. Discussion WHAT GAP THIS FILLS What we already know: Provision of web-based resources is a valuable addition to face-to-face teaching in a blended learning environment. What this study adds: The order in which resources are listed significantly dictates the frequency of access of individual resources. Providing more than four resources per topic is associated with decreased access rates for subsequently listed resources. It would be expected that learners may access resources that are at the beginning of a list more frequently than those at the bottom of a list and sheer curiosity may well be the major driver of exploring what the first-listed resources contain. What is surprising from this research is how marked this behaviour is. After the fourth resource, the frequency of access declines and remains low. Even within the first four listed resources, the order of appearance would appear to influence the frequency of accessing the material. that all participants will read them, others may be provided as a library of material that can be drawn upon if there is a need, but there is no expectation that all participants will access them. In this case, as is likely in the majority of blended educational courses, no distinction was drawn between resources considered essential and others. The data show that the third resource was more frequently accessed than the second resource. A The data also raise questions about the nature of learning from resources in this environment. It would be expected that learners may access resources that are at the beginning of a list more frequently than those at the bottom of a list and sheer curiosity may well be the major driver of exploring what the first-listed resources contain. What is surprising from this research is how marked this behaviour is. more detailed analysis of the resources was undertaken to seek an explanation for this. If appropriate for the topic of a scaffold, the Best Practice Advocacy Centre (BPAC) guidelines are included amongst other resources and guidelines. For three of the scaffolds, the BPAC guidelines were placed third in the order of presentation by coincidence. The hit rate for each of the BPAC guidelines was high in comparison to other resources and therefore skewed the results. These data suggest that providing more than four resources per clinical topic is associated with learner overload. However, there may be several reasons for making resources available; some resources may be provided with the intention Social learning theory promotes the notion that cognition is not an individual process. Rather, learning and knowing are shaped by both interaction with others and the context of these interactions.8 Resources within a social learning context become methods of exploring different ways of knowing as well as finding material that suits individual learning styles. As described by Cook and McDonald, ‘Ultimately, the rationale for using computers in education comes not because of any inherent instructional advantage but because computers facilitate the use of effective instructional methods’.3 Accordingly, resources should be structured to provide alternative ways of understanding or cater to differing learning styles using methods either VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 215 ORIGINAL SCIENTIFIC PAPERS QUANTITATIVE RESEARCH not available or inappropriate for face-to-face learning in small group meetings. As the availability of resources increases due to cooperation and sharing between institutions and countries, it will become even more important that careful thought is applied to what resources are included as part of the curriculum.9 An alternative view is that resources should be considered carefully for their ability to contribute to learning that is likely to be assessed. This position is based on the well-researched premise that what is assessed will drive what is learned.10,11 A plethora of resources for a curriculum topic, some of which contain conflicting and contradictory views or where the resource is of marginal interest, is clearly not examinable material irrespective of its ability to encourage deep learning. The response of assessment-driven students will be to avoid committing time to such learning. It is suggested that, in planning the use of online resources in a blended learning environment, the following guidelines may be helpful: References 1. Smothers V, Ellaway R, Greene P. The e-learning evolutionleveraging new technology approaches to advance healthcare education. Med Teach. 2008;30(2):117–8. 2. Cook DA, Levinson AJ, Garside S. Time and learning efficiency in Internet-based learning: a systematic review and meta-analysis. Adv Health Sci Educ Theory Pract. 2010;15(5):755–70. 3. Cook DA, McDonald FS. E-learning: is there anything special about the ‘E’? Perspect Biol Med. 2008;51(1):5–21. 4. D’Alessandro DM, Kreiter CD, Peterson MW. An evaluation of information-seeking behaviors of general pediatricians. Pediatrics. 2004;113(1 Pt 1):64–9. 5. Hannafin MJ, Hill JR. Resource-based Learning. In: Spector M, Merrill J, van Merrienboer, Driscoll MP, editors. Handbook of research on educational communications and technology. 3rd ed. London: Erlbaum; 2007. p 525–536. 6. Matyas ML. Online teaching resources: where in the WWW are they? Physiology (Bethesda). 2007;22:68–9. 7. Choules AP. The use of elearning in medical education: a review of the current situation. Postgrad Med J. 2007;83:212–216. 8. Hill JR, Song L, West RE. Social learning theory and webbased learning environments: a review of research and discussion of implications. Am J Distance Educ. 2009;23:2:88–103. 9. Soula G, Darmoni S, Le Beux P, Renard J Dahamna B, Fieschi M. An open repositories network development for medical teaching resources. Studies in health technology and informatics. Stud Health Technol Inform. 2010;160(Pt 1):610–4. 10. Newble DI, Jaeger K. The effect of assessments and examinations on the learning of medical students. Med Educ. 1983.17(3):165–71. 11. Fredericksen N. The real test bias. Influences of testing on teaching and learning. Am Psychol. 1984;39:193–202. • Limit the number of resources to four for each clinical topic unless there are pressing reasons to include more • Carefully consider the value of each resource for its ability to represent an alternative view of a topic or to cater to different learning styles • For each resource, consider if the material is considered essential or ‘if needed’ and clearly indicate this difference. Conclusions COMPETING INTERESTS None declared. 216 When lists of resources are provided in a blended learning environment, the order of presentation of the resources will substantially affect the rate at which the resources are accessed. The first four listed resources are likely to have much higher access rates than the remainder of the resources. Resources need to be considered carefully for how they add value to learning within a social learning theory, whether the resource is ‘essential’ or ‘if needed’ and the number of resources should be kept to four or below. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUALITATIVE RESEARCH Addressing patient alcohol use: a view from general practice Thomas Mules;1 Jennifer Taylor;1 Rachel Price;1 Logan Walker;1 Baneet Singh;1 Patrick Newsam;1 Thenmoli Palaniyappan;1 Toby Snook;1 Mahfuzah Ruselan;1 John Ryan;1 Jaishree Santhirasegaran;1 Phoebe Shearman;1 Petronella Watson;1 Richard Zino;1 Louise Signal;1 Geoff Fougere;1 Helen Moriarty;2 Gabrielle Jenkin1 1 Department of Public Health, University of Otago Wellington, New Zealand ABSTRACT INTRODUCTION: General practitioners (GPs) have the potential to promote alcohol harm minimisation via discussion of alcohol use with patients, but knowledge of GPs’ current practice and attitudes on this matter is limited. Our aim was to assess GPs’ current practice and attitudes towards discussing alcohol use with their patients. 2 Department of Primary Health Care and General Practice, University of Otago Wellington METHODS: This qualitative study involved semi-structured, face-to-face interviews with 19 GPs by a group of medical students in primary care practices in Wellington, New Zealand. FINDINGS: Despite agreement amongst GPs about the importance of their role in alcohol harm minimisation, alcohol was not often raised in patient consultations. GPs’ usual practice included referral to drug and alcohol services and advice. GPs were also aware of national drinking guidelines and alcohol screening tools, but in practice these were rarely utilised. Key barriers to discussing alcohol use included its societal ‘taboo’ nature, time constraints, and perceptions of patient dishonesty. CONCLUSION: In this study there is a fundamental mismatch between the health community’s expectations of GPs to discuss alcohol with patients and the reality. Potential solutions to the most commonly identified barriers include screening outside the GP consultation, incorporating screening tools into existing software used by GPs, exploring with GPs the social stigma associated with alcohol misuse, and framing alcohol misuse as a health issue. As it is unclear if these approaches will change GP practice, there remains scope for the development and pilot testing of potential solutions identified in this research, together with an assessment of their efficacy in reducing hazardous alcohol consumption. KEYWORDS: Primary health care; general practice; alcohol drinking; alcohol-related disorders, attitude of health personnel. Introduction Alcohol-related harm is a global problem. In developed countries, alcohol is responsible for 6.7% of disability-adjusted life years lost and 1.6% of deaths and is increasing.1 In New Zealand (NZ) more than 1000 deaths every year can be attributed to alcohol, resulting in 17 000 years of life lost annually.2 In a recent NZ study, the proportion of people drinking more alcohol over the past year had increased from 2% in 1998 to 16% in 2006.3 The worsening of alcohol-related health problems internationally was realised as early as 1980, when a World Health Organization (WHO) expert committee on alcohol stressed the need for increased efforts to prevent alcohol-related health harms.4 The committee called for the development of strategies that could be applied in primary health care settings with a minimum of time and resources. Primary care was identified as a key setting for the reduction of alcohol-related harm, with general practitioners (GPs) considered VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):217–222. CORRESPONDENCE TO: Louise Signal Associate Professor, Department of Public Health, University of Otago PO Box 7343, Mein St, Newtown, Wellington, New Zealand. louise.signal@otago.ac.nz 217 ORIGINAL SCIENTIFIC PAPERS QUALITATIVE RESEARCH to be in an ideal position to detect, prevent and manage patients’ alcohol problems. A number of advantages of a GP-based strategy have been identified.5 In addition to there being good evidence that brief interventions delivered by GPs have a positive impact on patients’ alcohol consumption, GPs are readily accessible to the general population and have role legitimacy in the delivery of advice about alcohol consumption.5 In most countries people with alcohol problems will first present to their GP6 rather than to specialist treatment services.7 It is perhaps because of this that there is a growing expectation that GPs will provide advice concerning lifestyle issues, including alcohol.8 Evidence in support of the efficacy of brief GP interventions in this area has also been accumulating, with longer duration of counselling having little additional effect.9 alcohol might be somewhat optimistic. Therefore, this study assessed GPs’ current practice in NZ (where GP consultation is based on a user-pays model with discounts for high needs and low income groups) and attitudes towards addressing alcohol use with their patients, together with barriers and supports to such initiatives identified by GPs. Methods A convenience sample of 19 GPs was interviewed by fourth-year medical students (the primary authors of the paper) while on a one-week general practice placement in Wellington, NZ. From 33 invited GPs, 14 declined (58% response rate), 11 because they were too busy, two were not interested and one was ill. GPs came from a variety of practices including privately owned and not-for- In addition to there being good evidence that brief interventions delivered by GPs have a positive impact on patients’ alcohol consumption, GPs are readily accessible to the general population and have role legitimacy in the delivery of advice about alcohol consumption. Despite the weight of evidence in favour of brief GP-based interventions to address hazardous alcohol consumption, uptake by GPs is limited.5 A number of challenges to introducing a discussion on alcohol in the GP consultation have been identified, including time pressure and sensitivity to the issue.5 Consequently, GPs do not always identify patients with hazardous levels of alcohol consumption.5,10 An Australian study5 reported that GPs were able to identify only 28% of patients classified by the Australian Medical Association criteria as ‘high risk’ drinkers, while another study found that between 65% and 82% of patients with alcohol-related problems (identified by consumption levels or screening tests) were not identified by GPs.11 Taken together, this evidence suggests that the expectation that GP-led brief interventions could facilitate significant reductions in harm from 218 profit, central city and suburban, and a range of socioeconomic areas. This sample was similar to GPs nationally. The average age of the sample was 47 (compared to 48 nationally), the proportion of males in the sample was 56% (compared to 59% nationally), almost half (n=9) had been practising for over 15 years and the same number worked in private practice compared to not-for-profit practice (no national data available for comparison). A semi-structured interview schedule included 15 open-ended questions covering GPs’ current practice and attitudes towards discussing alcohol use with their patients and barriers to, and support needed to facilitate, such discussions. Face-to-face interviews were approximately 20 minutes long, tape-recorded with consent, and transcribed. A thematic analysis12 was undertaken, with initial coding and analysis cross-checked by other research team members. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUALITATIVE RESEARCH Ethics approval was received from the Department of Public Health Ethics Committee, University of Otago Wellington. Findings Current practice Discussions with patients The GPs in this study reported rarely discussing alcohol in their consultations. Reasons for initiating discussions with patients around alcohol included: • suspicious clinical signs (raised by 15 GPs) • social issues, such as domestic violence, depression or frequent work absence (mentioned by six GPs), and • consultations with new patients (mentioned by six GPs). Other circumstances mentioned less frequently were: • as part of routine questioning • in the presence of other addictions • when the presenting complaint was related to mental health • when alcohol was a presenting complaint, or • when alcohol had been raised by a concerned family member. Screening Routine screening for alcohol misuse was not common practice in consultations conducted by these GPs. While six GPs screened all new patients for alcohol misuse, seven GPs reported rarely screening (all were from private practices and six had practised for over 15 years). Sixteen of the GPs who screened relied on a list of verbal questions (e.g. the CAGE tool—a mnemonic for attempts to reduce drinking, being annoyed when criticised about drinking, feeling guilty about drinking and using alcohol on waking),13 three used written screening tools (the Alcohol Use Disorders Identification Test—AUDIT)14 and, two used liver function blood tests. Of the six GPs who screened all new patients, four had practised less than 15 years and five were salaried and from not-for-profit practices (serving low income popu- WHAT GAP THIS FILLS What we already know: General practitioners (GPs) are thought to be well positioned to reduce alcohol-related harm via discussion of alcohol misuse with patients. GPs identified a number of barriers to discussing alcohol misuse as well as some potential solutions to these barriers. What this paper adds: Although GPs in this study thought it was important to discucss alcohol use with their patients, they rarely did so unless the patient’s drinking had signifant impacts on their health. lations). All the younger GPs who screened used a routine set of questions. Fifteen of the 19 GPs were aware of NZ guidelines for alcohol screening but 13 did not use them to guide their practice. The difficulty of raising the issue of alcohol with patients was noted by a number of GPs and some had developed tactical ways of approaching the topic, such as using the screening window (on the computer) as an excuse for asking: ‘So I’ll say what we’re supposed to be doing is… asking everybody how much alcohol they [consume] in a week’ or ‘[we just] need to update our details, are you allergic to anything, do you smoke or drink?’ Doctors said that patients expect smoking questions so some bundle alcohol into the same question. Interventions GPs identified interventions including GPdelivered advice, referral to drug and alcohol services, referral to Alcoholics Anonymous, medication, and family support. Only five GPs delivered alcohol intervention to patients such as, ‘counselling [during] the consultation and talking to the patient about what is a safe level of alcohol intake’. Many of these GPs found that it was sometimes very difficult to arrange followup visits as patients would not attend. Almost all of these GPs reiterated the importance of the patient’s willingness to talk for discussion to be successful. GP attitudes The role of the GP GPs in the study agreed that primary care has an important role to play in delivering primary prevention, including preventing harm from alcohol. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 219 ORIGINAL SCIENTIFIC PAPERS QUALITATIVE RESEARCH Eleven GPs felt their own advice about alcohol was useful to patients, six were undecided, and two thought it made no difference to patients’ drinking. GPs expressed concern about the number of primary prevention strategies they were expected to address. Perspectives on what their role should entail varied and included: raising patient (n=7) and public awareness (n=2), referral to specialist services (n=7), providing advice (n=5), and some were unsure (n=5). Use of screening and guidelines Overall, GPs in the study thought that opportunistic screening was more effective than routine screening. Those who used screening questions found them ‘too rigid’ for the flow of consultation and therefore modified them. Reasons given for not using guidelines included constant changes being made to the recommendations, guidelines not being widely applicable, and a lack of consistency between different guidelines. One of the two GPs who used guidelines talked about the difficulty in staying up to date with the changes. Barriers to addressing alcohol GPs identified a number of barriers to addressing alcohol use with patients. The most common was the ‘taboo’ nature of the subject (n=11). As one GP explained, ‘there are stigmas around [alcohol], so people don’t necessarily like talking about it’. This contributes to doctor and patient discomfort. One GP noted that, ‘if there is clinician discomfort about asking these [alcohol-related] questions, then they might not ask [the patient]’. Three GPs stated that they felt asking about alcohol use was ‘intruding into other people’s lives unnecessarily’ and they feared that raising the topic could damage the doctor–patient relationship, making future encounters difficult. Another GP noted that raising the subject often made patients ‘wince… and some became defensive’. Ten GPs identified lack of time as a significant barrier. One noted that ‘quality alcohol consultation should take… 15 minutes’. Perception of patient dishonesty was another commonly identified barrier. Four GPs believed that many patients were not honest about their alcohol use 220 and therefore doubted the value of questioning patients. One stated that ‘most people halve their alcohol [consumption]. We were always taught that you double what people say’. Three GPs identified that the presence of third parties, commonly family members, made it inappropriate to raise the issue. Other barriers included the patient not accepting that their drinking was a health issue, the GP feeling they did not have the expertise needed, and the GP being unable to define a safe level of alcohol consumption to the patient. When prompted on specific patient demographics that may act as barriers, a third of GPs in the study found it difficult to raise the topic of alcohol with people of differing ethnicity and gender, and a quarter found it hard to raise with people of a different age. As one GP stated: …it is always easier if you’re from a [patient’s] cultural group because they identify with you and it is difficult to tell someone from a very different cultural background about [alcohol]… they are less likely to listen to what you have to say. The age barrier was a particular issue with teenage patients who see themselves as ‘bulletproof’, often ‘don’t discuss [alcohol] with older people’, and often come with a parent or support person. The barriers of ethnicity, age and gender were more commonly identified by less experienced and younger GPs. Supports identified by GPs The majority of GPs (n=13) said that more support was needed to facilitate discussions of alcohol use with patients. Four said they did not need further support and two made no comment. Four GPs suggested this support should come from raising public awareness of the adverse health effects of alcohol. As one GP noted, ‘if it’s just doctors struggling with this problem it’s not enough’. Four GPs, all practising for less than 15 years, said longer consultations were needed. Four GPs suggested using standardised questionnaires, ideally administered by nurses prior to the consultation, would save time. One GP suggested that it VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS QUALITATIVE RESEARCH would be easier if a questionnaire was incorporated into a medical records computer programme. Discussion Routine screening of patients by GPs in this study was uncommon and alcohol use was rarely discussed with patients in their consultations. However, nearly a third of GPs screened all new patients. Slightly over a quarter of GPs delivered alcohol interventions to their patients, including GP-delivered interventions, referral and medication. This is despite evidence of the considerable potential for success.4,5,7,9,15–17 This finding supports earlier work.10 While some GPs in this study were confident in this arena, most did not address the issue unless the level of drinking had led to a disease state that could not be overlooked—a finding consistent with research in this area.18,19 Of the GPs who screened all new patients, the majority were younger and nearly all were salaried. This suggests that more recent GP training may encourage, or better equip, GPs to address alcohol issues. It may also suggest that the work context of salaried GPs is more supportive of alcohol screening than that of those who work for fee-for-service. It is possible that GPs in fee-for-service practices are under more pressure not to offend their patients in case they go elsewhere. While GPs agreed there is an important role for primary care in primary prevention, they were less clear on what this role was in relation to alcohol. GPs were generally of the view that screening for alcohol misuse should be opportunistic rather than routine. When raising alcohol use with patients, some GPs did so following use of screening tools and alcohol guidelines, and others relied on clinical or social indicators. The study suggests that GPs find evidence-based guidelines and structured questionnaires cumbersome and if they use them they alter the validated questions. Barriers to addressing alcohol use with patients include: • the taboo nature of the subject • concern about intruding into people’s lives • time shortages in consultations • GP perception of patient dishonesty • the presence of a third party, and • the challenge of raising alcohol with people of different ethnicity, gender and age. Age was a particular issue with teenagers. The majority of GPs thought more support was needed, including raising public awareness of the health effects of alcohol, longer consultations, screening prior to the consultation, and incorporation of screening tools into medical records programmes. These findings suggest the advantages GPs have in addressing alcohol use with patients that Paton-Simpson et al.5 identify are not widely experienced, at least by this study population. First, GPs may be accessible, but short con- While GPs agreed there is an important role for primary care in primary prevention, they were less clear on what this role was in relation to alcohol. GPs were generally of the view that screening for alcohol misuse should be opportunistic rather than routine. sultations appear to be a barrier to addressing alcohol use, although evidence suggests that five minutes may be all that is required for successful intervention.7 Further, at least in NZ, access to a GP is still limited for many. Second, many GPs in this study were not clear about the legitimacy of their role in alcohol health promotion. This may be in part because alcohol is seen as a social problem and its discussion therefore invokes social stigma rather than health issues. Third, while there is good evidence of effectiveness for GP intervention, many of the GPs in this study were unaware of, or did not believe, the evidence. The brief time required and the fact that brief interventions may be even more efficacious than specialist treatment9 are messages that GPs need to hear. This research also confirms and extends earlier research on barriers VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 221 ORIGINAL SCIENTIFIC PAPERS QUALITATIVE RESEARCH to progress in this area.10 GPs in the study provided advice about the support that would assist them. This is a small qualitative study; hence, the findings are not generalisable. However, the sample is similar to the NZ GP population and findings are consistent with previous NZ research10 and other research internationally.18,19 The response rate of 58% may have introduced some bias, as most of those who declined cited time constraints. It is possible that these are the GPs most likely to spend time with patients discussing issues such as alcohol use, and that this study therefore underestimates the extent that alcohol issues are addressed in general practice. Regardless, this research suggests that a substantial number of GPs in NZ may not routinely address alcohol issues with their patients. Conclusions ACKNOWLEDGEMENTS The authors would like to thank the GPs who participated in the study for generously giving of their time and experience. AUTHOR CONTRIBUTIONS The students (TM, JT, RP, LW, BS, PN, TP, PN, TS, MR, JR, JS, PS, PW RZ) participated in the research design, collected and analysed the data and wrote the initial report. LS, HM and GF initiated the research and supervised the researchers at all stages of the research. GJ wrote the first draft of this paper and GJ, LS, HM, GF and TM contributed to subsequent drafts and shared responsibility for editing the final version. All authors have contributed to, and approved, the final version. COMPETING INTERESTS None declared. 222 GPs in the current study rarely discussed alcohol use with their patients and most did not do so unless the level of drinking was significantly impacting on health. GPs thought that discussing alcohol use was important although they were less clear as to their role in this. These findings challenge the concept that GPs are well positioned to deliver community-based alcohol screening and brief intervention. There are unresolved societal, organisational and interpersonal barriers which deserve further exploration if GPs are to provide primary intervention to reduce alcohol-related harm. First is the need to legitimate the role of GPs in this arena, both in the minds of the public and with GPs. A social marketing campaign that promotes the health effects of alcohol misuse and the assistance GPs can provide may encourage patients to approach their doctor, and give GPs more confidence in raising the issue without concern that they were intruding in patients’ lives. Second, consideration could be given to how primary care is organised to promote alcohol health promotion, for example in terms of accessibility, funding models, time with patients, the place of screening, and screening prompts. Third, GP training should build the skills of individual clinicians including stressing the efficacy of brief intervention and building their confidence to undertake it. Addressing these issues will better equip GPs to help their patients reduce harm from alcohol misuse. Urgent action is needed if we are to stem the increasing harm from alcohol. References 1. World Health Organization. Global health risks: Mortality and burden of disease attributable to selected major risks. Geneva: World Health Organization; 2009. 2. Connor J, Broad J, Jackson R, Vander Hoorn S, Rehm J. The burden of death disease and disability due to alcohol in New Zealand. ALAC Occasional Publication 23. Wellington; 2005. 3. Wilkins C, Sweetsur P. Trends in population drug use in New Zealand: Findings from national household surveying of drug use in 1998, 2001, 2003, and 2006. N Z Med J. 2008;121(1274):61–71. 4. World Health Organization. Problems related to alcohol consumption: report of a WHO expert committee. Geneva, Switzerland: World Health Organization; 1980. 5. Paton-Simpson G, McCormick R, Powell A, Adams P, Bunbury D. Problem drinking profiles of patients presenting to general practitioners: Analysis of alcohol use disorders identification test (AUDIT) scores for the Auckland area. N Z Med J. 2000;113:74-77. 6. Rush B. The use of family medical practices by patients with drinking problems. Can Med Assoc J. 1989;140:35–39. 7. Anderson P. Alcohol and primary health care. WHO regional publications. In: WHO regional publications, editor. European series No. 64: World Health Organization; 1996. 8. Wallace P, Brennan P, Haines A. Are general practitioners doing enough to promote healthy lifestyle? BMJ. 1987;297:663–68. 9. Kaner EF, Dickinson HO, Beyer FR, Campbell F, Schlesinger C, Heather N, et al. Effectiveness of brief alcohol interventions in primary care populations (Review). Cochrane Database Syst Rev 2007(Issue 2):Art. No.: CD004148. DOI: 10.1002/14651858.CD004148.pub3. 10. Moriaty H, Stubbe M, Chen L, Tester R, Macdonald L, Dowell A, et al. Challenges to alcohol and other drug discussions in the general practice consultation Fam Pract. 2011. doi: 10.1093/fampra/cmr082. 11. Rydon P, Redman S, Sanson-Fisher R, Reid L. Detection of alcohol related problems in general practice. J Stud Alcohol. 1992;53:197–202. 12. Green J, Thorogood N. Qualitiative methods for health research. London: Sage; 2004. 13. Bush B. Screening for alcohol abuse using the CAGE questionnaire. Am J Med. 1987;82(2):231–35. 14. Senft R, Polen M, Freeborn D, Hollis J. Brief intervention in a primary care setting for hazardous drinkers. Am J Prev Med. 1997;13(6):464–70. 15. Bien T, Miller W, Tonigan J. Brief interventions for alcohol problems: A review. Addiction. 1993;88(3):315–36. 16. Chick J, Ritson B, Connaughton J, Stewart A. Advice versus extended treatment for alcoholism. Br J Addict. 1988;83:159–70. 17. Drummond D, Thom B, Brown C, Edwards G, Mullan M. Specialist versus general practitioner treatment of problem drinkers. Lancet. 1990;336:915–18. 18. Anderson P. Managing alcohol problems in general practice. BMJ. 1985;290:1873–5. 19. Powell A, Adams P, McCormick R. Preventive medicine in general practice with particular emphasis on early intervention for alcohol. N Z Fam Physician. 1996;23:44–7. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH Factors influencing diagnostic decision-making Kathleen SN Callaghan PhD ABSTRACT INTRODUCTION: Identifying influences on diagnostic decisions is important because diagnostic errors often have far-reaching consequences for an individual’s future within the workforce and their eligibility for Accident Compensation Corporation–funded treatment. Most investigations of factors biasing decision making have used quantitative techniques rather than qualitative methods. Human Factors Group, Department of Surgery, Faculty of Medical and Health Sciences, The University of Auckland, New Zealand AIM: To identify factors influencing GPs’ diagnostic decision-making and to develop a valid questionnaire to determine the desirability and importance of each factor’s influence. METHODS: Focus groups and the Delphi method were combined with Rasch analysis to identify factors influencing GPs’ diagnostic decision-making and then examine the strength and stability of ratings of the factors’ desirability and importance. RESULTS: Thirty-nine factors were identified. Factors demonstrating high stability but no consensus included the importance of evidence-based medicine, the potential ramifications of a diagnosis, and the desirability of medicolegal issues. Factors for which there was disagreement in the first Delphi round but consensus in the second round included the importance of patient advocacy/support groups and the desirability of examination findings. Rasch analysis indicated that the questionnaire was close to the model (88.6% and 86.2% of variance in the ratings of importance and desirability explained). DISCUSSION: Participants readily identified factors influencing GPs’ diagnostic decision-making. Their ratings did not appear to support a prescriptive model of medicine, yet two cornerstones of prescriptive medicine, clinical information and probability of disease, were rated as highly desirable and important. KEYWORDS: Decision-making; diagnosis; bias; Rasch analysis; general practitioners Introduction Making an accurate diagnosis and selecting an appropriate treatment can have profound consequences for an Accident Compensation Corporation (ACC) claimant, determining when or if they are able to return to work. Most investigations of factors influencing diagnostic or clinical decision-making employ quantitative methods rather than qualitative methods. Qualitative techniques can capture a range of factors which may be overlooked by quantitative methods.1 In one such qualitative study dermatologists were asked to identify non-clinical or diagnostic factors and to indicate how influential the factors were.1 Three types of factors were identified: patientcentred (e.g. patient choice, place of residence, ethnicity and age), clinician-related factors (e.g. time constraints, relationship with colleagues and staff, and pressure from the pharmaceutical industry), and practice-related factors (e.g. public or private treatment, cost of treatment to NHS). United States studies have found regional and practice variations in diagnostic practices (e.g. overuse of diagnostic tests) which were related to the intensity of hospital and physician services.2 Diagnostic practices were less likely to be related to patient characteristics. Hajjaj et al.3 argued that understanding how non-clinical factors affect evidence-based practice is important in clinical consultation, but usually overlooked. They argued that a combination of clinical and non-clinical factors may contribute to VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):223–230. CORRESPONDENCE TO: Kathleen Callaghan Human Factors Group, Department of Surgery, Faculty of Medical and Health Sciences, The University of Auckland, PB 92019 Auckland, New Zealand k.callaghan@ auckland.ac.nz 223 ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH a clinician’s decision without their awareness (e.g. patient adherence to advice and taking medication may influence the choice of treatment). They also suggested that a major challenge to practising medicine is integrating evidenced-based medicine (EBM) with important non-clinical factors while maintaining good standards of care. For example, a United Kingdom report on fitness for work advocated adopting the biopsychosocial approach to judging an employee’s capacity or fitness for work rather than focusing on their physical symptoms.4 Others have argued that while the biopsychosocial model is an inadequate scientific tool, it is nevertheless useful for clinical and teaching purposes5 or the foundation of a philosophy of medicine.6 The current study employed qualitative and quantitative techniques to examine factors influencing diagnostic decisions. Three methods were used to identify and evaluate factors affecting diagnostic decisions (especially within an ACC-related context): focus groups, the Delphi method, and Rasch analysis. Using mixed quantitative and qualitative methods combines the strengths of the two while reducing their weaknesses.7 The general practitioner’s (GP’s) evaluation of a patient’s injury or injuries is critical to the success of an ACC claim: ‘ACC legislation is a case in point where diagnosis routinely affects entitlement and cover decisions.’8 A misdiagnosis can result in temporary or long-term loss of employment, reduced employment opportunities, and poorer health.9,10 Analyses of ACC claims have revealed shortcomings in the diagnoses11–13 so a greater understanding of decision-making is vital. Given the predominance of the biomedical model of medicine and the prescriptive and normative approaches to decision-making,14,15 it was expected that the factors identified in the current study would reflect these models and approaches. believe such techniques are necessary to reveal hidden values and beliefs.18 The Delphi method The Delphi method was used to identify and prioritise factors that influence GPs’ decisionmaking. ‘Delphi may be characterised as a method for structuring a group communication process so that the process is effective in allowing a group of individuals, as a whole, to deal with a complex problem.’19 ‘The Delphi survey is a group facilitation technique, which is an iterative multistage process, designed to transform opinion into group consensus.’20 The Delphi method reduces the influence of more dominant participants and group pressure.21 Rasch analysis The Rasch model postulates that ‘…the probability of endorsing a statement is a logistic function of the difference between two independent quantities: the strength of the respondent’s attitude and the location of the statement on the attitude continuum.’22 Rasch analysis yields a measure on a unidimensional scale of the strength of a respondent’s attitude, and another measure of the location on the same scale of the opinion expressed by an item under study.23 The units of the scale are log-odds ratio or logits. In this study, the key respondent attitude is ‘a tendency to endorse’ the item and the key item measure is the importance or desirability of a diagnostic factor. The scale was calibrated such that the average of the items’ importance or desirability was arbitrarily set to zero. The University of Auckland Human Subjects Ethics Committee informed the investigator that ethical approval was not required. Method 224 Focus groups Participants Focus groups are ‘a group of individuals selected and assembled by researchers to discuss and comment on, from personal experience, the topic that is the subject of the research.’16 Focus groups may be biased if one individual dominates the discussion or if the participants are too readily influenced by the facilitator;17 however, some authors Expert groups responsible for setting GP standards in NZ were identified. There are a number of groups that influence GP standards; however, many are biased, either legislatively or otherwise, towards a particular viewpoint. Academic organisations were identified as being the least biased ‘standard setters’; therefore, participants were VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH selected from five academic organisations which provide GP training. These were the three Departments of General Practice at the University of Otago (Dunedin, Christchurch and Otago), the Departments of General Practice at The University of Auckland, and The Royal New Zealand College of General Practitioners (RNZCGP). Heads of department were asked to select a minimum of four participants from senior members of their department currently in clinical practice. A reasonable gender balance and a mix of rural/city practice familiarity was preferred. The characteristics of the 14 focus group participants were as follows: 10 were male, 4 were female; 13 were in current clinical practice (one had recently ceased practice to complete a full-time postgraduate degree); 11 were in urban practice, and 3 were in rural practice. Focus group meetings The investigator met with three of the five expert groups (time constraints prevented a meeting with the RNZCGP or the Christchurch Department of General Practice). Background information about the study was sent prior to the meetings. The investigator explained the purpose of these meetings, but did not participate in the subsequent discussion. The focus group from the Dunedin Department of General Practice comprised six participants while the other groups comprised four participants. The lists of factors obtained from each meeting was examined for overlapping concepts and repetitions, and a composite list embracing the intent of all the expert groups consulted was compiled. An electronic questionnaire for Round 1 of the Delphi process was constructed from the composite list. Copies of the questionnaire together with instructions and background information were sent by email to 23 experts who had not participated in the focus groups. The questionnaire contained 39 factors identified by the focus groups as influencing GP diagnostic decision-making. The experts were asked to rate the importance and desirability of each factor using a seven-point scale, where 1 represented ‘not at all important’ or ‘not at all desirable’ and 7 represented ‘very WHAT GAP THIS FILLS What we already know: Clinical and non-clinical factors may bias clinical decision-making. Bias is related to the GP’s perception of the costs and benefits of making one choice over another. What this study adds: Thirty-nine factors that potentially bias diagnostic decision-making by GPs and the subjective value placed on these factors were identified. Despite responses diverging from the prescriptive model, standard setters endorsed two cornerstones of the model—clinical information and probability of disease were rated as both highly important and desirable. important’, or ‘very desirable’. Experts were free to comment on each factor. Despite both email and telephone reminders, only 12 Round 1 responses were received—a response rate of 52%. Of the 12 respondents, 6 were male, 6 were female; 11 were in current clinical practice, 1 had recently ceased practice to complete a full time postgraduate degree; 5 were in urban practice, and 3 were in rural practice. Following standard Delphi methodology, the experts who responded in Round 1 were asked to re-rate each influencing factor and provide comments if they wished (Delphi Round 2). Several areas were clarified based on comments from the previous round, and one new factor added—a question designed to explore the influence of potential ramifications of the diagnosis. Clarifications and alterations were clearly identified by a different coloured font. Experts were provided with their original ratings from Round 1 together with the group mean rating for each item. Only 11 Round 2 responses were received. Given the effort and time required to elicit completed Round 2 responses, it was assumed that the cost of a further Delphi round was likely to outweigh any benefits obtained; hence, a third round was not undertaken. A Rasch analysis was performed using WINSTEPS® Version 3.55.24 The method adopted was the rating-scale version of the Rasch model25 where the ratings given to each item are not assumed to be equally spaced, but all items share the same structure. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 225 ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH Results Thirty-nine factors were identified by the focus groups as influencing GPs’ diagnostic decisionmaking (Tables 1 and 2). Standard Delphi methodology The summed mean ratings and stability or consensus of response for the group of GP standard setters are listed in Tables 1 and 2. Stability of response is an important, albeit often unreported, consideration in the analysis of Delphi responses. Whether or not further Delphi rounds may be productive can be established by measuring the stability of respondents’ opinion distribution curves over successive rounds. The assessment of stable non-consensual distributions should be of equal interest in assessing opinion as the assessment of stable consensual distributions. Group stability of opinion between the two Delphi rounds was assessed26 and is reported in Tables 1 and 2. The absolute differences in the histograms of responses were calculated (subtracted column-wise) for two successive rounds and then summed to show total units of change. Net person changes were then calculated by dividing the total units of change by two (any one participant’s change of opinion is reflected in the histogram differences by two units of change). Finally, the percentage change was calculated by dividing the net person changes by the number of participants. In this method, up to a 15% change level is accepted as representing a stage of equilibrium. The higher the percentage change, the less stable the group’s position. Responses from the participant who withdrew between Rounds 1 and 2 were eliminated from the analysis of stability. Similarly, factors which were not rated by one of the remaining participants could not be analysed accurately (no factor had more than one non-response). Stability was therefore represented as a range, with the lowest number representing the respondent as having maintained their previous rating and the highest number representing a change of rating. Of interest are factors demonstrating high group stability without consensus (indicating fixed disagreement). The importance of EBM, 226 potential ramifications of the diagnosis and the desirability of medicolegal issues are examples of these. In most Delphi analyses, consensus is assumed to have been reached when a specified percentage of responses falls within a prescribed range. Neither the proportion nor the range has been defined in the literature.20 For this study, consensus was defined as having been reached when all responses fell within a three-point range at Round 2. If all responses fell within a two-point range (or less) then this was defined as strong consensus. Responses falling outside of a three-point range constituted disagreement. This is consistent with other interpretations of agreement and disagreement.27 Some factors showing consensus in Round 2 were highly unstable indicating that the group had shifted from a state of disagreement to a state of agreement between Round 1 and Round 2. Examples of this are the importance of patient advocacy/support groups and the desirability of examination findings as influencing factors. It is often assumed that each item measured contributes in a meaningful way to the construct being investigated; that is, the questionnaire measures a single dimension. The dimensionality of the questionnaires used in this study was determined by principal components analysis. The Rasch measure explained 88.6% of the variance in ratings of importance (unexplained variance measured by the next component was 2.8%) and 86.2% of the variance in the ratings of desirability (unexplained variance measured by the next component was 4%). These results indicate that the questionnaire was unidimensional and close to the Rasch model’s intent of measuring ‘one attribute at a time’.23 While the investigator labelled the constructs ‘importance’ and ‘desirability’, it can be inferred from examination of the items that the one attribute measured is represented by the label. The assumption that ratings are equally spaced was tested using Masters and Wright’s method28 in which cumulative answers give the probability of responding greater or equal to each one of the available categories (e.g. ≥1, ≥2 etc). The threshold value is set where the cumulative probabilities equal 0.5. For example, in Table 3 the logit VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH Table 1. Summed mean ratings of importance, stability, degree of stability, consensus and strength of consensus Mean rating (importance) Stability Degree of stability Consensus Strength of consensus History 6.8 Yes 9.1 Yes † Examination findings 6.2 Yes 9.1 Yes * GP’s personal clinical experience 5.5 Yes 9.1 Yes * GP’s knowledge of local conditions 5.4 18.2 Yes † Yes * Influencing factor * Results of investigations 5.3 Possibly yes 13.6–18.2 Evidence-based medicine 5.2 Yes 9.1 Characteristics of the GP 4.8 22.7 Need to achieve an outcome 4.4 31.8–36.4 Patient expectations 4.2 36.4 Patient advocacy 4.0 22.7–27.3 Medicolegal issues 4.0 36.4 Characteristics of the patient 4.0 27.3 Reasonable patient pressure 4.0 Closeness of GP/patient relationship 4.0 27.3 Time available for the consultation 4.0 External feedback from a medical source 3.9 Potential ramifications of the diagnosis 3.9 The clinical setting 3.6 27.3 Need to justify a course of action 3.6 18.2 Personal circumstances of the patient 3.5 36.4 Expectations of external medical professionals 3.5 22.7 The Health and Disability Commissioner 3.1 18.2 Potential implications for the wider community 3.1 50–54.5 Context in which the diagnosis is made 3.1 18.2 Administrative requirements 3.0 27.3 Unreasonable patient pressure 3.0 36.4 Possibly yes 13.6–18.2 18.2 18.2 Yes 0.0 GP’s desire to please the patient 3.0 27.3 Diagnostic algorithms/categories/protocols 3.0 18.2 Fear of uncertainty 2.9 36.4 Whether or not treatment is available for the diagnosed condition 2.9 27.3 GP’s perception of what other external health professionals may think 2.8 36.4 GP’s perception of the state of the national health care system 2.6 27.3 Marketing/media 2.6 18.2 Legal requirements 2.6 27.3 Patient advocacy/support groups 2.5 36.4 Technological tools 2.3 27.3 Who is funding the consultation 1.7 Business considerations 1.6 External incentives 1.1 Yes Yes 9.1 Yes * 18.2 Yes † 9.1 Yes † Consensus † Strong consensus VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 227 ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH Table 2. Summed mean ratings of desirability, stability, degree of stability, consensus and strength of consensus. Mean rating (desirability) Stability History 6.6 Yes Examination findings 6.2 Evidence-based medicine 6.2 GP’s knowledge of local conditions 5.7 Results of investigations 5.5 Influencing factor * Degree of stability Consensus Strength of consensus 0.0 Yes † 27.3 Yes * Yes 9.1 Yes * Yes 9.1 Yes * Yes * Yes * 18.2 GP’s personal clinical experience 5.2 27.3 External feedback from a medical source 4.3 36.4 Technological tools 3.8 31.8–36.4 Potential implications for the wider community 3.6 Patient advocacy 3.6 Characteristics of the GP 3.5 Characteristics of the patient 3.5 45.5 Patient expectations 3.4 18.2 Need to achieve an outcome 3.3 22.7–27.3 27.3 Possibly yes 13.6–18.2 18.2 Closeness of GP/patient relationship 3.1 Medicolegal issues 2.8 40.9–45.5 Expectations of external medical professionals 2.8 Diagnostic algorithms/categories/protocols 2.8 27.3 Legal requirements 2.6 45.5 Need to justify a course of action 2.4 31.8–36.4 Yes 9.1 36.4 Reasonable patient pressure 2.4 36.4 Personal circumstances of the patient 2.3 22.7–27.3 The Health and Disability Commissioner 2.3 27.3 Potential ramifications of the diagnosis 2.3 18.2 GP’s perception of what other external health professionals may think 2.1 18.2 Administrative requirements 1.9 18.2 The clinical setting 1.9 18.2 Time available for the consultation 1.9 18.2 Patient advocacy/support groups 1.9 Marketing/media 1.6 36.4–40.9 Yes * 13.6–18.2 Yes † GP’s desire to please the patient 1.6 36.4 Yes * Fear of uncertainty 1.6 22.7–27.3 Yes † Whether or not treatment is available for the diagnosed condition 1.6 13.6–18.2 Yes * Possibly yes Possibly yes Unreasonable patient pressure 1.4 18.2 Yes † Who is funding the consultation 1.3 18.2 Yes † Context in which the diagnosis is made 1.3 4.5–9.1 Yes * GP’s perception of the state of the national health care system 1.2 27.3 Yes † Business considerations 1.2 Yes 9.1 Yes † External incentives 1.0 Yes 9.1 Yes † Yes Consensus † Strong consensus 228 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH Table 3. Fifty percent cumulative probabilities for ratings of importance and desirability Likert category (desirability) 1 2 3 4 5 6 7 50% cumulative probability (logits) … -1.97 -0.96 -0.27 0.38 1.02 1.81 Likert category (importance) 1 2 3 4 5 6 7 -2.77 -1.70 -1.00 0.15 1.62 3.70 50% cumulative probability (logits) value of -2.77 is the threshold at which a Likert rating of 1 is equally probable as a rating of 2 or above. The logit value of 3.70 is the threshold at which a rating of below 6 is as equally probable as a rating of 7. The thresholds are not spread equidistantly and it is easier to move between categories when rating desirability than when rating the importance of influencing factors. When the distribution of participants and factors was plotted, the most important influencing factor was history, followed by examination findings, and the GP’s personal clinical experience. However, the difference between history and examination findings is 2.22 logits versus 1.43 logits between examination findings and GP’s knowledge of local conditions; hence, history would appear to be a much more important influencing factor than would be expected from knowledge of its ordinal position alone. Similarly, external incentives are regarded as being much less important than business considerations, with a distance of 1.89 logits separating these factors. While history is viewed as the most desirable influencing factor, it is not as desirable as it is important, according to the difference in ratings (-5.95 logits for importance versus 3.56 logits for desirability). This indicates that the experts preferred to de-emphasise history as an influencing factor while maintaining its ordinal position. EBM was the sixth most important influencing factor but the second most desirable factor, despite representing only a difference of 0.91 logits. At the other end of the importance and desirability ordinal scale, factors such as external incentives, business considerations, and who is funding the consultation show very little logit mismatch. Both the importance and desirability questionnaires are marked by a very broad span of 10.3 and 7.97 logits respectively, indicating a quite marked ‘black and white’ view of the factors at either extreme. The item reliability index was used to test the replicability of item placements (influencing factors) assuming these same factors were to be given to another group of comparable standard setters. For ratings of importance the item reliability index was 0.94 (an estimated 94% of the observed variance in the results obtained can be regarded as true variance, with the remaining 6% attributable to measurement error). In other words, the results obtained could be expected to correlate 0.94 with the results from a parallel measure of the same underlying construct. For ratings of desirability, the item reliability index was 0.93. ACC-specific ramifications were considered less important than generic ramifications (0.01 versus 0.18 logits). The ramifications of a diagnosis (whether generic or ACC-specific) were not considered to be a desirable influence on diagnostic decision-making (0.34 and 0.39 logits). Discussion The relatively low number of participants in this study and the low response rate of 52% raises some concerns over the reliability29,30 of the results obtained; however, the item reliability index values ameliorate these concerns. The current study used a qualitative approach to identify factors influencing GPs’ decision-making together with quantitative techniques. The stability of responses was also assessed. An advantage of Rasch scaling is that it enables meaningful comparisons between different groups undertaking the same questionnaire. It is noteworthy that the standard setters readily identified 39 factors that influence GPs’ diagnostic decision-making. The factors identified are not unexpected because they reflect everyday influences such as what others think of us, what we think of others, legality, need to achieve an outcome VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 229 ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH and so forth. That these factors can be identified supports the controversial notion that it is not possible to divorce the practice of medicine from the society in which it is practised.31 Interestingly, while a prescriptive view of diagnostic decisionmaking (in which decision-making is improved through the use of clinical guidelines etc.32) appears invalidated by the responses obtained in this study, the factors on which a prescriptive view of medicine relies (clinical information and probability of disease) are rated as both highly important and highly desirable. This apparent dissonance might be explained by Elstein:33 ‘All models and theories are simplifications of reality. Models are not reality however, and theory is not practice.’ Further research is required to determine the prevalence and consequences of diagnostic errors and to obtain further insight from an in-depth ‘systems’ approach (including nature of the error and underlying causal/contributory factors). A diagnosis relating to an ACC claim can be critical to whether a patient returns to work. Remaining in or returning to work is recognised as providing a significant contribution to health and wellbeing.34–36 A better understanding of diagnostic decision-making within an ACC-related context, therefore, has important financial and social consequences not only for the patient but also for their community. References ACKNOWLEDGEMENTS We thank John Irwin and Pat Bullen for their invaluable comments on this manuscript. FUNDING The source of funding for the study was a University of Auckland Senior Health Research Scholarship. COMPETING INTERESTS None declared. 230 1. Hajjaj F, Salek M, Basra M, Finlay A. Nonclinical influences, beyond diagnosis and severity, on clinical decision making in dermatology: understanding the gap between guidelines and practice. Br J Dermatol. 2010;163(4):789–99. 2. Song Y, Skinner J, Bynum J, Sutherland J, Wennberg JE, Fisher ES. Regional variations in diagnostic practices. N Engl J Med. 2010;363(1):45–53. 3. Hajjaj F, Salek M, Basra M, Finlay A. Non-clinical influences on clinical decision-making: a major challenge to evidence-based practice. J R Soc Med. 2010;103(5):178–87. 4. Bevan S, Passmore E, Mahdon M. Fit for work? Musculoskeletal disorders and labour market participation. London; 2007. 5. Kendler KS. The rise and fall of the biopsychosocial model: reconciling art and science in psychiatry. Am J Psychiatry. 2010;167(8):999–1000. 6. Lewis B. The biopsychosocial model and philosophic pragmatism: is George Engel a pragmatist? Philos Psychiatr Psychol. 2008;14(4):299–310. 7. Johnson R, Onwuegbuzie A. Mixed methods research: a research paradigm whose time has come. Educ Res. 2004;33(7):14–26. 8. Callaghan KSN. Factors that influence general practitioner diagnostic decision-making and a comparison with other stakeholders. University of Auckland; 2006. 9. The Royal Australasian College of Physicians. Compensable injuries and health outcomes. Sydney: The Royal Australasian College of Physicians; 2001. 10. The Royal Australasian College of Physicians AFOEM. Helping people return to work: using evidence for better outcomes. Sydney; 2010. 11. Gorman DF. Audit of the ACC Workwise Clinic—December 1997. Auckland: The University of Auckland; 1997. 12. Gorman DF. Audit of the ACC Workwise Clinic—September 1997. Auckland: The University of Auckland; 1997. 13. Gorman DF, Jarvie P, Robinson P. Occupational health practice in New Zealand. NZ Med J. 1999;112:79–82. 14. Shaban RZ. Theories of clinical judgment and decisionmaking: a review of the theoretical literature. J Emerg Primary Health Care. 2005;3(1–2). 15. Wilson HJ. The myth of objectivity: is medicine moving towards a social constructivist medical paradigm? Fam Pract. 2000;17(2):203–09. 16. Powell RA, Single HM. Focus groups. Int J Qual Health Care. 1996;8(5):499–504. 17. Cox SJ, Cheyne AJT. Assessing safety culture in offshore environments. Saf Sci. 2000;34(1–3):111–29. 18. DeJoy DM. Behavior change versus culture change: divergent approaches to managing workplace safety. Saf Sci. 2005;43(2):105–29. 19. Linstone H, Turoff M. Introduction. In: Linstone H, Turoff M, editors. The Delphi method. Techniques and applications. Massachusetts: Addison-Wesley Publishing Company; 1975. 20. Hasson F, Keeney S, McKenna H. Research guidelines for the Delphi survey technique. J Adv Nurs. 2000;32(4):1008–15. 21. Normand S-LT, McNeil BJ, Peterson LE, Palmer RH. Eliciting expert opinion using the Delphi technique: identifying performance indicators for cardiovascular disease. Int J Qual Health Care. 1998;10(3):247–60. 22. Irwin RJ, Irwin KC. Attitude towards mathematics of 11-yearolds: analysis of a questionnaire and its relation to mathematical achievement. In: Waugh RF, editor. Frontiers in educational psychology. New York: Nova Science Publishers; 2004:107–16. 23. Bond TG, Fox CM. Applying the Rasch model: fundamental measurement in the human sciences. Mahwah, New Jersey: Lawrence Erlbaum Assoicates; 2007. 24. WINSTEPS: Multiple-choice, rating scale, and partial credit Rashch analysis [program]. Chicago: MESA Press; 2000. 25. Andrich D. A rating formulation for ordered response categories. Psychometrika. 1978;43:561–74. 26. Scheibe M, Skutsch M, Schofer J. Experiments in Delphi methodology. In: Linstone H, Turoff M, editors. The Delphi method. Techniques and applications. Reading, Massachusetts: Addison-Wesley Publishing Company; 1975:263–83. 27. Brook RH, Chassin MR, Fink A, Solomon DH, Kosecoff J, Park RE. A method for the detailed assessment of the appropriateness of medical technologies. Int J Tech Assessment Health Care. 1986;2 (Part 1):53–63. 28. Masters GN, Wright BD. The partial credit model. In: van der Linden WJ, Hambleton RK, editors. Handbook of modern item-response theory. New York: Springer; 1997:101–21. 29. Frisbie DA. Reliability of scores from teacher-made tests. Educ Meas: Issues and pract. 1988;7(1):25–35. 30. Traub RE, Rowley GL. Understanding reliability. Educ Meas: Issues and pract. 1991;10(1):37–45. 31. Gorman DF, Scott J. The social distortion of medical practice. Med Today. 2003;4(11):75–77. 32. Shaban RZ. Theories of clinical judgment and decisionmaking: a review of the theoretical literature. J Emerg Primary Health Care. 2005;3(1–2):23–33. 33. Elstein AS. Clinical problem solving and decision psychology: comment on ‘The epistemology of clinical reasoning’. Acad Med. 2000;75(10):S134–S36. 34. Bevan S, Quadrello T, McGee R, Mahdon M, Vavrovsky A, Barham L. Fit for work? Musculoskeletal disorders in the European workforce. London; 2009. 35. Black DC. Working for a healthier tomorrow. Norwich; 2008. 36. The Royal Australasian College of Physicians AFOEM. Realising the health benefits of work. Sydney; 2010. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH Maori cultural adaptation of a brief mental health intervention in primary care Fiona Mathieson MA (applied) Clinical and Community Psychology, Dip Tchg;1 Kara Mihaere PGdip Clinical Psychology, PhD;2 Sunny Collings MBChB, PhD;3 Anthony Dowell MBChB;4 James Stanley PhD4 1 Department of Psychological Medicine, University of Otago Wellington, New Zealand ABSTRACT INTRODUCTION: There are no brief psychological mental health interventions designed specifically for Maori in a primary care setting. AIM: To adapt an existing cognitive behavioural therapy–based, guided self-management intervention for near-threshold mental health syndromes in primary care, for Maori, and to examine its acceptability and effectiveness. METHODS: Semi-structured interviews with primary care clinicians and Maori patients were conducted to inform adaptations to the intervention. Clinicians were then trained in intervention delivery. Patients were recruited if they self-identified as Maori, were aged 18–65 years, were experiencing stress or distress and scored ≤35 on the Kessler-10 (K10) measure of global psychological distress. Patient and clinician satisfaction was measured through a questionnaire and semi-structured interviews. Post-intervention, patients’ mental health status was measured at two weeks, six weeks and three months. 2 Te Whare Marie, Specialist Maori Mental Health Service, Capital Coast DHB, Wellington 3 Director of Social Psychiatry and Population Mental Health Research Unit, University of Otago Wellington 4 Primary Health Care and General Practice, University of Otago Wellington RESULTS: Maori adaptations included increased emphasis on forming a relationship; spirituality; increased use of Maori language and changes to imagery in the self-management booklets. Nine of the 16 patients recruited into the study completed the intervention. Patients and clinicians rated the intervention favourably and provided positive feedback. Improvement was seen in patients’ K10 scores using intention-to-treat rated global psychological distress following intervention. DISCUSSION: This study found that it was not difficult to adapt an existing approach and resources, and they were well received by both providers and Maori patients. Further research is required with a larger sample utilising a randomised controlled trial, to establish whether this approach is effective. KEYWORDS: Primary health care; Maori mental health; patient satisfaction; brief intervention Introduction Disseminating evidence-based mental health treatments into real-world settings poses a number of challenges for clinicians and policy makers. The field of cross-cultural mental health is a fledgling one.1 This is certainly the case in New Zealand, with little research on talking therapies for Maori,2 none of which are in primary care mental health. There are disparities between Maori and nonMaori regarding access to mental health services and, when contact occurs, it is often with primary care services.3 Near-threshold mental health symptom complexes such as anxiety, depression, harmful use of recreational or prescribed substances, or a combination of these (not at a diagnostic level), are common (15.3%) in Maori.3 ‘Near-threshold’ anxiety, depression and substance use are common presentations in primary care and often coexist. While not meeting the threshold for disorder in standard diagnostic systems such as DSM-IV,4 these symptom complexes carry a significant disability burden along with the risk of developing a diagnosable disorder.5,6 These symptom complexes often arise in the context of social problems and cause distress and VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):231–238. CORRESPONDENCE TO: Fiona Mathieson Department of Psychological Medicine, University of Otago Wellington, PO Box 7343, Wellington 6242, New Zealand fiona.mathieson@ otago.ac.nz 231 ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH significantly impaired work and role functioning.7 GPs perceive and manage them as clinically significant.8,9 There is evidence that mental health services are negatively perceived by Maori,10,11 and it has been recommended that mental health services for them should be more closely aligned with primary health care, with services geared to their cultural expectations, if they wish. Increasing primary care clinician skills in assessment and management of Maori mental health will maximise early intervention and minimise referral out.12 primary care setting. The sessions comprise one hour in total and are supported by a treatment manual and self-help booklets on the topics of relationships, stress and breaking habits. The clinician role is that of coach and change facilitator rather than ‘therapist’. Patients work to develop a specific plan for change, which is printed out as a prescription. Prescriptions are integrated with the computerised patient management system for ease of access. There is one follow-up phone call/ email to ask how this plan is working. The literature on talking therapy adaptation for Maori is scant.2 One study adapted cognitive behaviour therapy for Maori based on input from a range of Maori advisors, with promising outcomes.13 However, there are no studies in this area in primary care. Clinicians were trained in a two-hour training session. Clinician and patient satisfaction ratings were very positive and patient outcome scores indicated that this approach may be beneficial, although the sample was small in the acceptability study.19 Expert opinion on cultural adaptation suggests that the following aspects of Maori culture need to be incorporated into therapy in order to be effective for Maori: an emphasis on whakawhanaungatanga (the process of forming relationships); an emphasis on the whanau (family) and iwi (tribe);14 use of te reo (Maori language); an emphasis on spirituality and fostering of strong cultural identity.15,16 At the same time, it is important to maintain awareness of cultural diversity among Maori and generalisations regarding cultural or therapeutic mores should be avoided.15,16 While it is important to be aware of common Maori worldviews and how they can be accommodated in practice, it is equally important to be guided by the individual tangata whaiora (person seeking wellness) and their whanau (extended family). International indigenous literature reflects similar concepts and values to those that have been identified in the New Zealand literature.2 An innovative, ultra-brief intervention (UBI) was designed for people aged 18–65 with near-threshold mental health syndromes. The approach was developed through a collaborative process involving patients, clinicians and researchers.19 UBI consists of three brief sessions of guided, cognitive behaviour therapy–based18 self-management, with a focus on problem solving and behaviour change, delivered by the primary care clinician in the 232 This paper reports on the process and outcomes of a feasibility study into adapting the UBI approach to better fit the needs of Maori patients. The aim was to develop a credible and acceptable UBI for subthreshold mental health syndromes among Maori, to be delivered in primary care settings by non-specialist clinicians. The questions were: 1. To what extent and in what way does UBI need to be adapted for Maori? 2. How acceptable is the adapted intervention to clinicians and Maori patients? Methods Adaptation process The adaptation process comprised a review of the literature, along with both individual face-to-face and group interviews with primary care clinicians and individual face-to-face interviews with Maori potential users of the intervention. The rationale for this approach was that a collaborative approach to therapy adaptation was more likely to result in an approach that was feasible and acceptable in this context. Early in the research process, a partnership was formed with a Maori health researcher (KM, co-investigator), who worked clinically in a primary care setting. She established relationships with local providers, conducted the interviews with patients and clinicians, and had face-to-face contact with patients to collect intake and outcome data. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH Interviews WHAT GAP THIS FILLS Semi-structured interviews were conducted with nine general practitioners (GPs) and nine primary care nurses. Of these, four identified as Maori (one GP, three nurses), six as New Zealand European and eight as ‘other’ ethnicity. Clinicians were based in urban group practice primary care settings, none of which were specifically focused on Maori health needs, but all of which had some Maori patients attending their practices. What we already know: Near-threshold mental health syndromes are common among Maori. Little is known about brief interventions in primary care for these syndromes and, similarly, the literature on adapting talking therapies for Maori is scant. Clinicians were shown the existing UBI booklets and the existing clinical approach was outlined to them. They were asked to respond to the existing UBI approach and to suggest how it would need to be adapted to better meet the needs of Maori patients. Clinicians reported that the existing approach and structure appealed. They felt the same basic format could be retained, but they thought it had a ‘middle-class European flavour’. They suggested that we have more and different imagery and more Maori language, easier language, fewer words, more relevant scenarios in the vignettes, a karakia (prayer) and a whakawhanaungatanga process (forming connections). Clinicians were keen to get continuing education credits for the training and preferred to have the training during regular peer meetings. Semi-structured, individual face-to-face interviews were also conducted with six potential patient users of the intervention, who all identified as Maori. Recruitment was through an article in the local community newspaper. Feedback from the potential patients was similar to that of the clinicians. There was a range of responses to the idea of offering karakia, with some more comfortable with the idea than others. What this study adds: This paper describes the process and outcomes of an adaptation of an innovative, ultra-brief intervention, creating a version with promising clinical effectiveness and acceptability for Maori. option of offering a karakia in English or Maori, if they were comfortable. A whakawhanaungatanga process was added at the start of the first session. In this process, clinicians were encouraged to self-disclose a little bit about themselves, such as family background or work history, and ask their patient about these areas. The imagery in the self-help booklets was changed by increasing the number of images and using more Maori designs. A Maori graphic designer was employed to make these changes. The covers of the booklets and treatment manual and some of the imagery contained in the booklets are shown in Figure 1. The scenarios in the vignettes were changed to be more culturally relevant, such as playing Figure 1. The covers of the self-help booklets and treatment manual and some of the imagery contained in the booklets Adaptations made In response to the feedback, the following adaptations were made to UBI: Patients were invited to begin sessions with an optional karakia or whakatauki (proverb). Whakatauki were provided, such as ‘He manga wai koia kia kore e whikitia; It is a big river indeed that cannot be crossed’. Alternatively, clinicians and patients were given the VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 233 ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH Figure 2. Recruitment of patient participants and data collection process Consenting patients make appointment for first session and make a problem list Patient presents to GP Routine clinical assessment CHAT/ clinical judgment Distress? Research assistant contacts patient and conducts baseline assessment Invite patient to participate touch rugby, rather than soccer. The sessions were changed to include a stronger emphasis on how whanau (extended family) could be involved in supporting the person, and patients were also asked about whether wairua (spirituality) was an important part of their wellbeing and, if so, how they nourished it. There was a slight increase in the use of Maori language in the booklets, largely through the use of whakatauki, and terms such as ‘wairua’ and ‘whanau’. Both patient and clinician feedback had suggested that full translation would not be helpful. Session 1 Clinician training Session 2 Identify as Maori; English speaking and aged 18–65? Session 3 + follow-up phone call And not suicidal Administer K10 K10 scores of 35 or less? Post-intervention K10 assessment of patients at 2 weeks YES Patient satisfaction questionnaire administered 6 weeks NO K10 scores of more than 35 Clinician satisfaction questionnaire administered The clinician training session was two hours long and included a presentation and role-play practice. In addition, the adaptations made for Maori were outlined and a video role-play of a clinician working with a Maori patient was added. This demonstrated skills such as offering the whakatauki and whakawhanaungatanga process. Recruitment As shown in Figure 2, patient eligibility was determined by clinician assessment of the presence of psychological distress. This was assessed by clinical judgment and/or the Case-finding and Help Assessment Tool (CHAT) for lifestyle and mental health assessment of adult patients in primary care.20 After applying the exclusion criteria (non-English speaking, under age 18 or over age 65, no suicidal ideation in the past two weeks), the remaining patients were screened by the clinician for the presence of psychological distress using the Kessler-10 (K10) measure of global psychological distress (a standard 10-item selfreport questionnaire based on questions about the level of anxiety and depressive symptoms in the preceding four-week period). Higher scores on the K10 indicate greater distress.21 Patient ineligible Offer medication or specialist referral 234 A score of 35 or less on the K10 determined final eligibility. Patients with K10 scores over 35 received their GP’s usual care as appropriate—generally medication or referral to a mental health professional. Patients meeting criteria for inclusion were invited to participate in the study. The VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH right-hand column of boxes in Figure 2 shows the study data collection process before, during, and after the UBI-Maori (UBI-M) sessions. Clinician and patient satisfaction Clinician and patient satisfaction questionnaires were administered following completion of the intervention (patient questionnaires at two weeks post-intervention and clinician surveys at the end of the study). Possible responses to each statement were based on a five-point Likert scale to indicate levels of agreement or disagreement. The statements were designed by the investigators and informed by feedback obtained from clinician and patient interviews conducted during the design phase of the intervention. Patient K10 scores were reassessed at two weeks, six weeks and three months post-intervention. Patients and clinicians were also asked a semi-structured series of questions about what they liked and did not like about the intervention and any changes they would like to see. Data collation and analysis All semi-structured interviews were audio recorded and transcribed verbatim. The transcript data was collated thematically under the headings of each question by the data gatherer (KM). Questionnaire data were entered into Microsoft Access. Data analysis was performed using R 2.11. The 95% confidence intervals reported were based on the t-distribution. Formal comparisons between K10 scores at intake with K10 scores post-intervention used paired-sample t-tests across participants. As only some participants who completed the intervention completed all three follow-up outcome measures (n=7), the descriptive statistics and hypothesis testing for the K10 data were calculated on an intentionto-treat basis—that is, missing data points were replaced with the most recent K10 score available for that participant. This produces a slightly more conservative estimate of the magnitude of improvement compared to excluding these individuals from analysis. Descriptive statistics (range, inter-quartile range, and median) were calculated for the Likert-response items on the patient satisfaction scales. The feasibility study was approved by the Central Regional Ethics Committee. (CEN/09/11/085). Approval was also gained from the Ngai Tahu Research Consultation Committee which is a partnership between Te Runanga o Ngai Tahu and the University of Otago. Findings Sample Sixteen patients were recruited into the study. The majority of these patients were female (13/16, 81%) and the median age was 38.9 years, with a range from 20 to 65 years. They came from a low socioeconomic, urban group, with 56% not in paid employment and 71% having an annual household income of less than $20,000. Five people did not complete the study for work or personal reasons, and two did not complete due to clinical reasons (including one patient referred to secondary mental health services). This left nine patient participants who completed the intervention. Seven of these completed all the follow-up measures. Of the 22 clinicians trained in the intervention, five clinicians used the intervention, each seeing between one and seven patients. Levels of patient satisfaction The results of the patient satisfaction questionnaire at two weeks post-intervention are presented in Figure 3. Patients who completed all three UBI-M sessions reported high levels of satisfaction with most aspects of the intervention. Some would have preferred more sessions and/or longer sessions. Some did not find the self-help booklets particularly relevant. Patient comments from the semi-structured interview are shown in Figure 4. Levels of clinician satisfaction The clinician satisfaction questionnaire ratings are not shown graphically due to the small number of clinicians who delivered the complete intervention (n=4). The interview data was thematically VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 235 ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH Figure 3. Boxplots (median, interquartile range, range) showing patient satisfaction ratings (n=9) collated under the headings of each question by the data gatherer (KM). Interview data and ratings indicated a high degree of clinician satisfaction with many aspects of UBI-M. These included the intervention, timing between sessions, the logical structure, the manual was easy to use, the prescriptive approach was acceptable, and it seemed to meet patients’ needs. Clinicians reported that they largely kept to the framework and it was different from their usual practice. They were happy to use the approach again and would recommend it to colleagues. However, some felt that the length of sessions was inadequate. Some suggestions were made for improvements, such as a booster training session soon after the initial training and assistance with pronunciation of Maori words. Clinician comments are shown in Figure 5. Figure 4. Patient satisfaction comments ‘It felt as though he was helping me that step further’ ‘It helped me as a person to have faith in myself again that I can actually pull through these tough times’ Mental health status of patients at follow-up Changes in the mental health status of the patient participants are illustrated in Figure 6, as measured by mean K10 scores and 95% confidence intervals (n=16 intention-to-treat analysis) at intake and post-intervention (at two weeks, six weeks and three months). Results in the lefthand part of the figure are from the original UBI study,19 which had a larger sample size (n=18). The results in the right-hand side of the figure are from the UBI-M study. For both studies, analysis was based on an intention-to-treat analysis. Improvements in score from the intake period were calculated for the UBI-M study. Mean improvement at Week 2 was 5.2 points (95% CI 1.5–8.8; one-sample t-test t(15)=2.99, p=0.009). Mean improvement at Week 6 was similar at 5.1 points (95% CI is 0.9–9.3; one-sample t-test t(15)=2.60, p=0.020). By month 3 post-treatment, the mean improvement from intake was no longer statistically significant at 4.2 points (95% CI -0.5–8.96; one-sample t-test t(15)=1.90, p=0.077). ‘It gave me the power back’ ‘I just felt safe in that environment’ Discussion ‘He was lifting me out of my hole but not mollycoddling me. I was doing it but he was there too’ A particular strength of this study is the demonstration that a culturally appropriate collaborative approach to intervention adaptation can result in 236 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH a talking therapy which has both clinical promise and high credibility and acceptability to patients and clinicians. Improvement was seen in patients’ rated global psychological distress following intervention. While the improvement was not statistically significant, the confidence intervals indicate that it is highly likely that the true mean improvement will be greater than zero (i.e. an actual improvement). However, this study is limited by the absence of a control group; hence, it remains unclear whether this is any greater improvement than might have occurred with no intervention. Other limitations include the relatively short follow-up (three months) and the small sample. There was relatively low uptake by clinicians. Reasons given for not using the intervention included not finding suitable patients (some practices had a low percentage of Maori patients), finding change difficult and being busy with competing demands.22 Unlike secondary care, primary care has a broad scope and patients may not present there with mental health issues as the main problem. The Figure 5. Clinician satisfaction comments ‘Good to have another tool’ ‘Familiar face, familiar environment were helpful’ ‘It normalises it. That fear is not so much there’ ‘I needed help with pronunciation’ ‘Like the getting to know process’ ‘Different from my usual practice. I do not usually self-disclose any personal information’ ‘The karakia put people at ease; patients seemed surprised but appreciative at being offered it’ ‘Had to chase people for sessions 2 and 3 (text reminders helped)’ ‘An extra training session would help, to help people get started’ ‘Bang on, really’ ‘It empowered clients to go away and do their own mahi (work) and then come back’ ‘I liked how the client made their own realisations’ ‘I liked having a framework so you don’t fall apart in the middle of it’ ‘Nobody in her life had ever paid her that kind of attention… it was quite humbling’ Figure 6. Mean K10 scores (95% confidence intervals) for UBI and UBI-Maori adaptation at intake, two weeks, six weeks and three months. Analysis includes intention-to-treat carry-over of initial K10 scores for non-completers. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 237 ORIGINAL SCIENTIFIC PAPERS MIXED METHOD RESEARCH competing demands on clinicians in terms of time and multiple presenting problems make it challenging for clinicians to incorporate mental health interventions into routine clinical practice, particularly if a suitable patient is not identified soon after UBI training occurs. This uptake issue could be addressed by providing top-up training approximately three weeks after the initial training occurs. Confidence with pronouncing Maori words could be assisted by the provision of a CD recording, practising common words and whakatauki. The training could be incorporated into ongoing peer review processes. The total amount of contact time for this intervention is one hour and although this is brief in terms of traditional therapy approaches, for many clinicians in the primary care context this is a long time. However, it can be argued that this is a relatively small amount of time compared to an area like diabetes care, which is long term and time consuming due to the emphasis on teaching self-management. Clinicians may nevertheless need both protected time and resources for this approach to be sustainable. ACKNOWLEDGEMENTS Thanks to Hori Mataki from Ariki Creative for the design work and Professor Felicity Goodyear-Smith, Associate Professor Simon Hatcher and Dr Simon Bennett for their contribution to the original UBI design. FUNDING Oakley Mental Health Research Foundation and University of Otago Research Grants. COMPETING INTERESTS None declared. 238 Our experience suggests that there is potential for clinical improvement to occur through brief talking interventions in primary care with both Maori and non-Maori versions of UBI. It seems likely that there are similar change processes that operate across cultural groups, at least for this near-threshold group of patients. The strong tradition of Maori oratory in healing aligns well with the overall principles of the brief intervention and particularly with the Maori adaptation. Despite the limitations, the findings of this study suggest that cultural adaptation of an intervention was feasible and was well received by both patients and clinicians. All but one of the research team in the study were non-Maori, and concerned about ‘getting it right’ culturally. The experience of this adaptation was rewarding for the research team and offers a model for other areas of clinical practice. Further research is required by way of a randomised controlled trial to establish whether this approach is effective. References 1. Hayes PA, Iwamasa GY, editors. Culturally responsive cognitive behavioral therapy: assessment, practice and supervision. American Psychological Association, Washington DC; 2006. 2. Te Pou. He rongoa kei te korero. Talking therapies for Maori: Wise practice guide for mental health and addiction services. Auckland: Te Pou o Te whakaaro Nui; 2010. 3. Oakley Browne MA, Wells JE, Scott KM, editors. Te Rau Hinengaro: the New Zealand Mental Health Survey. Wellington: Ministry of Health; 2006. 4. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994. 5. Judd L, Schettler P, Akiskal H. The prevalence, clinical relevance and public health significance of subthreshold depressions. Psychiatr Clin North Am. 2002;25:685–698. 6. Kessler R, Zhao S, Blazer D, Swartz M. Prevalence, correlates, and course of minor depression in the national comorbidity survey. J Affect Disord. 1997;45(1–2):19–30. 7. Collings S and The MaGPIe Research Group. Disability and the detection of mental disorder in primary care. Soc Psychiatry Psychiatr Epidemiol. 2005;40(12),994–1002. 8. Dowell A, Garrett S, Collings S, et al. Primary Mental Health Initiatives: Interim Report. Wellington: School of Medicine and Health Sciences, University of Otago, Wellington; 2007:242. 9. Backenstrass M, Joest K, Rosemann T, et al. The care of patients with subthreshold depression in primary care: is it all that bad? A qualitative study on the views of general practitioners and patients. BMC Health Serv Res. 2007;7(1):190. 10. Durie M. Whaiora: Maori Health Development. Auckland: Oxford University Press; 1994. 11. Jansen P, Bacal K, Crengle S. He Ritenga Whakaaro: Maori Experiences of health services. Auckland: Mauri Ora Associates; 2008. 12. Durie M. Mental health and Maori development. Aust N Z J Psychiatry. 1999;33(1):5–12. 13. Bennett S. Te Huanga o te Ao Maori: Cognitive behavioural therapy for Maori clients with depression—development and evaluation of a culturally adapted treatment programme. A dissertation presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Psychology, Massey University, Wellington, New Zealand; 2009. 14. Durie MH, Hermansson G. Counselling Maori people in New Zealand [Aotearoa]. Int J Adv Couns. 1990;13(2),107–118. 15. Bennett S, Flett R. Te Hua o te Ao Maori. He Pukenga Korero: A Journal of Maori Studies. 2001;6(2), 29–34. 16. Hirini P. Counselling Maori clients—He Whakawhiti Nga Whakaaro i te Tangata Whaiora Maori. NZ J Psychol. 1997;26(2):13–18. 17. Mathieson F, Collings S, Dowell A, Hatcher S, GoodyearSmith F, Stanley J. From technology transfer to collaboration: an example of real world therapy adaptation (under review, Cognitive and Behavioral Practice). 18. Beck A, Rush A, Shaw B, Emery G. Cognitive therapy of depression. The Guilford Press; 1979. 19. Collings S, Mathieson F, Dowell A, Stanley J, Jenkin G, Goodyear-Smith F, Hatcher S. Acceptability of a self-help mental health intervention in general practice. Fam Pract. 2011;0:1–7. 20. Goodyear-Smith F, Coupe N, Arroll B, Elley C, Sullivan S, McGill A. Case finding of lifestyle and mental health disorders in primary care: validation of the ‘CHAT’ tool. Br J Gen Pract. 2008 January; 26–31. 21. Andrews G, Slade T. Interpreting scores on the Kessler Psychological Distress Scale (K10). Aust N Z J Public Health. 2001;25(6):494–97. 22. Davis L, Collings S. A bit set in my ways: clinicians’ willingness to implement a primary mental health intervention. Social Psychiatry and Population Mental Health Research Unit, University of Otago Wellington; 2011. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS SHORT REPORT The anatomical placement of body organs by Australian and New Zealand patients and health professionals in general practice Marjan Kljakovic MBChB, FRNZCGP, PhD ABSTRACT INTRODUCTION: Understanding patients’ awareness of the anatomical placement of their body organs is important for doctor–patient communication. AIM: To measure the correct anatomical placement of body organs by people from Australian and New Zealand general practices Academic Unit of General Practice and Community Health, School of General Practice, Rural and Indigenous Health, Australian National University Medical School, Woden, ACT, Australia. METHOD: A questionnaire survey containing drawings of 11 organs placed in different locations within each drawing. RESULTS: Among 1156 participants, there was no difference in the proportion of correct placement of 11 organs between Australian (51.7%) and New Zealand (49.6%) general practices. There was a positive correlation between the proportion of correctly placed organs and the age participants left school (p=0.012) and a negative correlation with the number of GP visits in the previous year (p=0.040). Participants from rural Australia were more likely to correctly place organs than urban participants (p=0.018). The mean proportion of organs correctly placed for doctors was 80.5%, nurses 66.5%, allied health 61.5%, health administrators 50.6% and the remaining consulting patients 51.3%. DISCUSSION: Patients from Australian and New Zealand general practice were poorly aware of the correct placement of organs. Health professionals were moderately better than patients at correct placement. KEYWORDS: Health knowledge; attitudes; practice; anatomy; general practice Introduction Methods The general public has been poorly aware of the anatomical placement of body organs for many years. Three UK studies found that only 50% of people correctly identified the placement of organs.1–3 Understanding how patients place their organs is important for doctor–patient communication.1 A convenience sample of three general practices from urban Australia, three from rural Australia, and two from urban New Zealand was approached. All participants over 18 years of age consented to completing the questionnaire while sitting in waiting rooms during normal practice hours over a three-day period. Health professionals from the practices, staff from the ACT Division of General Practice, and staff from the Southern General Practice Network also completed the questionnaire. A literature search found only UK studies in this area.1–4 The aims of this study were to test the null hypothesis that the rate of correct placement of organs was 50% for patients attending Australian and New Zealand general practices and to compare the rate of correct placement of organs by patients and health professionals. Weinman’s cross-sectional questionnaire-based design was replicated.1 Anatomical placement was the dependent variable. The first part of VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):239–241. 239 ORIGINAL SCIENTIFIC PAPERS SHORT REPORT Table 1. Comparison of the percentage of correct anatomical placement of 11 organs for 1156 participants from Australia and New Zealand who reported having the following chronic diseases: asthma, chronic obstructive pulmonary disease, diabetes, heart failure, or hypertension Total N (%) Body organ Asthma N=142 COPD N=26 Depression N=151 Diabetes N=75 Heart failure N=26 Hypertension N=257 94.0% Bladder 1046 (93.5) 94.2% 96.2% 90.7% 83.8%† 93.3% Intestine 1031 (91.3) 93.6% 88.5% 91.4% 90.7% 80.6% ‡ 89.8% Liver 636 (57.5) 61.9% 48% 57.4% 50.7% 58.6% 54.2% Stomach 557 (49.6) 51.8% 38.5% 46.3% 45.2% 30% ‡ 45.2% Gallbladder 552 (50.5) 49.6% 65.2% 40.7% 47.9% 48.1% 54.1% Pancreas 498 (45.6) 40.6% 48% 40.4% 46.5% 46.2% 51.0% Heart 480 (42.9) 46.4% 38.5% 43.2% 45.2% 36.7% 36.4%§ Kidney 474 (42.2) 43.6% 46.2% 42.3% 38.4% 30.0% 44.2% Ovaries 381 (34.1) 30.4% 52% 35.1% 38.7% 48.3% 32.4% Lungs 351 (31.5) 28.1% 24% 31.1% 34.7% 35.5% 31.1% Thyroid 270 (24.4) 17.9% 4.3%* 20.1% 17.8% 20.7% 25.8% * † ‡ § p=0.024 p=0.001 p=0.033, p=0.02 Table 2. Comparison of the percentage of correct anatomical placement of body organs by 1088 participants from occupations in Australia and New Zealand Occupation N Mean % of correct placement SD Health professionals* 141 63.8% 18.9% Teaching 56 51.8% 15.5% Students 34 53.4% 17.3% Employed 490 49.1% 16.1% Not employed† 367 49.2% 16.3% Total 1088 51.3% 17.2% * Includes doctors, nurses, allied health, health administrators † Includes unemployed, retired, at home, beneficiaries ment of the practice; and number of times they had visited a general practitioner (GP) and/or a practice nurse over the previous 12 months. The data were entered into PASW® Statistics 18 for appropriate statistical analysis. The Australian National University Human Ethics Committee and the New Zealand Health and Disability Ethics Committee, University of Otago Medical School, approved the study. Questionnaire responses were anonymous. Results the questionnaire consisted of drawings testing the placement of the heart, lungs, stomach, intestines, bladder, thyroid, liver, kidneys, pancreas, gallbladder, and ovaries.1 Each drawing depicted four body outlines each with the organ placed in different locations within an outline. Participants selected one out of four drawings they felt correctly placed the organ within the body outline. The independent variables were gender; age; occupation; age on leaving full-time education; which, if any, of the following chronic diseases they had—asthma, chronic obstructive pulmonary disease (COPD), bronchitis/emphysema, depression, diabetes, heart failure, and hypertension/high blood pressure; urban or rural place- 240 Out of 1184 questionnaires returned, 28 were removed because participants were too young. Australians compared to New Zealanders were no different in their mean (SD) age (51.18 (16.23) years versus 49.26 (18.04) years, t-test=1.621, p=0.105). They left school at a slightly older age (19.20 (6.63) years versus 18.2 (6.25) years, t-test=2.115, p=0.035). They visited their GP more often in the previous year (4.32 (2.94) visits versus 3.59 (2.75) visits, t-test=3.299, p<0.001). And visited the nurse less often in the previous year (1.05 (2.14) visits versus 1.97 (2.32) visits, t-test=-5.734, p<0.001). The mean (SD) proportion of organs correctly placed was 55.1% (19.2%) with no difference between Australia (51.7% (7.1%)) and New Zealand (49.6% (17.4%)). VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS SHORT REPORT There was no gender difference in the overall mean proportion of correct placement of organs (males 50.0% versus females 52.0%, p=0.061). There was a positive correlation between correct placement and the age participants left school (r=0.076, p=0.012). There was a negative correlation between correct placement and the number of GP visits in the previous year (r=-0.061, p=0.040). Participants from rural Australia were more likely to correctly place organs than from urban Australia (mean (SD) correct placement 5.68 (1.94) versus 5.41 (1.90), t=2.36, p=0.018). There was no correlation found for age of patient, number of nurse visits in the previous year and the number of reported chronic diseases. Table 1 shows no difference in the proportion of correctly placed organs for participants reporting they had asthma or depression. Participants reporting COPD were less likely to correctly place the thyroid (Fisher’s exact test, p=0.034). Participants reporting diabetes were less likely to correctly place the bladder (χ2=11.484, df=1, p=0.001). Participants reporting heart failure were less likely to correctly place the intestine (χ2=4.549, df=1, p=0.033) and the stomach (χ2=4.528, df=1, p=0.033). Participants reporting hypertension were less likely to correctly place the heart (χ2=5.396, df=1, p=0.02). Table 2 lists the occupations of 1088 participants including 27 (2.5%) doctors, 40 (3.7%) nurses, 38 (3.5%) health administrators and 37 (3.4%) allied health workers. The mean (SD) proportion of organs correctly placed for doctors was 80.5% (11.5%), nurses 66.5% (13.4%), allied health 61.5% (18.1%), health administrators 50.6% (19.7%) and the remaining consulting patients 51.3% (17.2%). Health professionals (63.8%) were modestly better than the remaining occupations (51.3%) at correctly placing organs (one-way ANOVA, p<0.001). Discussion This study found people from Australia and New Zealand were poorly aware of the correct anatomical placement of 11 organs—a finding no different from the UK.1 The biological characteristics of age, gender (as in the UK1), and chronic disease did not appear to influence the correct WHAT THIS GAP FILLS What we already know: Patients in the United Kingdom have a poor understanding of anatomy. What this study adds: Patients from Australian and New Zealand general practice also are poorly aware of the correct anatomical placement of organs. Health professionals were moderately better than patients at correct placement. placement of organs. Social characteristics such as education (as in the UK,1 the longer patients were educated, the more correct they were at placing organs), living in a rural location in Australia, or access to general practice services did appear to have an influence. A limitation of this study is that it did not distinguish between patients’ personal experience from what they had been taught when deciding on the correct placement of organs. Understanding how patients place their organs is important for doctor–patient communication. For example, only 28% of patients with asthma could place their lungs correctly. This poor understanding of lung anatomy might influence the poor peak flow meter technique found among patients with asthma.5 Furthermore, the negative correlation between the correct placement of organs and the number of GP visits in the previous year, and the lack of correlation with visits to the nurse, suggests contact with health professionals did not add to patients’ knowledge of anatomy. Perhaps also, communication difficulties arose because health professionals were only modestly better than their patients at correctly placing organs. References 1. Weinman J, Yusuf G, Berks R, Rayner S, Petrie KJ. How accurate is patients’ anatomical knowledge: a cross-sectional, questionnaire study of six patient groups and a general public sample. BMC Fam Pract. 2009;10:43. 2. Rashid A, Jagger C. Patients’ knowledge of anatomical location of major organs within the human body: a comparison between Asians and non-Asians. Family Pract 1996;13(5):450. 3. Boyle C. Difference between patients’ and doctors’ interpretation of some common medical terms. BMJ. 1970;2(5704):286–89. 4. Leventhal H, Weinman J, Leventhal EA, Phillips LA. Health psychology: the search for pathways between behavior and health. Annu Rev Psychol. 2008;59:477–505. 5. RNZCGP Wellington Faculty General Practice Network, Kljakovic M. Utilisation of the peak flow meter for acute asthma in general practice. NZ Fam Physician 2000;27(3):58-64. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ACKNOWLEDGEMENT The author would like to thank the people who contributed to this study. COMPETING INTERESTS None declared. 241 ORIGINAL SCIENTIFIC PAPERS IMPROVING PERFORMANCE Evaluation of Tu Meke PHO’s Wairua Tangata Programme: a primary mental health initiative for underserved communities Sally Abel PhD;1 Bob Marshall PhD;1 Donny Riki BASS (psychotherapy);2 Tania Luscombe BAgEcon, DipTchg2 1 Eastern Institute of Technology, Hawke’s Bay ABSTRACT 2 Wairua Tangata Kaiwhakaora Hinengaro/Lead Practitioner, Innov8 Group Ltd BACKGROUND AND CONTEXT: New Zealand’s primary mental health initiatives (PMHIs) have successfully filled a health service gap and shown good outcomes for many presenting with mild to moderate anxiety/depression in primary health care settings. Maori have higher rates of mental health disorders and complexity of social and mental health needs not matched by access to PMHIs. ASSESSMENT OF PROBLEM: The Wairua Tangata Programme (WTP), a Hawkes Bay PMHI, aimed to provide an integrated, flexible, holistic, tikanga Maori–based therapeutic service targeting underserved Maori, Pacific and Quintile 5 populations. External evaluation of the programme provided formative and outcome feedback. RESULTS: The WTP reported high engagement of Maori (particularly women), low non-attendance rates, good improvements in mental health assessment exit scores, strong stakeholder support and service user gratitude. GPs reported willingness to explore mental health issues in this high needs population. Challenges included engaging Pacific peoples and males and recruiting from scarce Maori, Pacific and male therapist workforces. STRATEGIES FOR IMPROVEMENT: Effectively meeting the target population’s complex social and therapeutic needs required considerable programme flexibility, referral back into the programme and assistance with transitioning to other therapeutic or social support services. Referral criteria required adaptation to accommodate some sectors, especially youth. A group programme was developed specifically for males. LESSONS: A holistic PMHI programme delivered with considerable flexibility and a skilled, culturally fluent team working closely with primary care providers can successfully engage and benefit underserved Maori communities with complex social and mental health needs. Successful targeted programmes are integral to reducing mental health disparities. KEYWORDS: Primary health care; mental health; Maori; medically underserved areas; evaluation. J PRIM HEALTH CARE 2012;4(3):242–248. Background CORRESPONDENCE TO: Sally Abel Senior Research Fellow, Eastern Institute of Technology, PB 1201, Hawkes Bay Mail Centre, Napier 4142, New Zealand sabel@eit.ac.nz In New Zealand, over a third of general practice patients are reported to have a diagnosable mental disorder, most commonly conditions relating to anxiety, depression and substance use.1 Major time constraints can militate against GPs counselling effectively within the context of the general practice setting,2 yet private therapy is costly and community-based 242 secondary care services, although free, are very stretched and have long waiting lists.3 The Ministry of Health (MoH) funded primary mental health initiatives (PMHIs), which commenced in 2004, allowed primary health care providers around the country to establish programmes suited to their patient population that could address mild/moderate anxiety or depression and prevent progression into more complex (and costly) conditions. An evaluation VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS IMPROVING PERFORMANCE of the first round of PMHI programmes found that a number of different models of care were developed, including some Kaupapa Maori programmes; programmes were viewed very favourably by primary health care providers and their patients; and some important service-user outcomes were demonstrated.3 The PMHIs (both mainstream and Kaupapa Maori) were intended to be responsive to Maori and an aim of their evaluation was to assess whether they were reducing mental health disparities and improving health outcomes.3 Maori have higher rates of mental health disorders that are not matched by access to mental health interventions.4 In addition, NZ GPs are less likely to diagnose depression in Maori patients than in non-Maori patients.5 The PMHI evaluation found that Maori comprised 17.5% of service users in the programmes evaluated, slightly higher than the 15% in the general population, and Kaupapa Maori and some mainstream PMHI programmes were considered responsive to Maori. However, many providers experienced difficulty providing Maori-focused programmes and recruiting a Maori workforce, and the evaluators noted that there remained a relative underutilisation of PMHI services by Maori, given their higher prevalence of mental health conditions. Evaluation feedback from Maori PMHI service users also highlighted the multiple everyday life stresses they faced and the sheer complexity of their mental health issues.3 This paper reports on an evaluation of a primary mental health initiative in Hawkes Bay that was developed a little later than the above-mentioned programmes. The Wairua Tangata Programme (WTP) specifically targets Maori, Pacific and NZDep Quintile 5 communities using a holistic tikanga Maori–based approach, and has shown some success at engaging and improving outcomes for this historically underserved population. The aims of the evaluation were to provide formative feedback to the implementation team and, through the collection and analysis of process and outcome data, to determine the success of programme outcomes. The paper reports on results from the programme’s first year with further follow-up results and comments. WHAT GAP THIS FILLS: What we already know: Primary mental health initiatives (PMHIs) have shown successful outcomes for many presenting with mild to moderate anxiety/depression in the primary health care setting, but appear to be underutilised by Maori who have higher rates of these conditions. What this study adds: A holistic PMHI programme delivered with flexibility and a skilled, culturally fluent team who work closely with primary care health practitioners can successfully engage and benefit underserved Maori communities with complex social and mental health needs. Context The WTP was established in 2008 by Tu Meke First Choice PHO (TMFCPHO), which at that time serviced a high needs population of 13 712 people (including 47% Maori and 11% Pacific) resident in Hastings and Flaxmere, Hawkes Bay. TMFCPHO was one of the three Hawkes Bay PHOs that have since been amalgamated into one, Health Hawke’s Bay: Te Oranga Hawke’s Bay. The WTP was funded through the second round of MoH PMHI contracts. Its target population was 18–60-year-old Maori, Pacific residing in an NZDep Quintile 5 area. Access was by GP referral and, as was required by the MoH contract, a mental health assessment tool was used prior to entry and on exit from the programme. The Kessler-10 (K10) assessment tool was chosen from a range of options, although none had been validated for Maori. The K10 criterion for admission was a score of 20–29 (mild–moderate depression and/or anxiety) out of a possible 50. Those referred to the programme received an initial therapeutic programme of four to six (more or less) sessions of counselling or psychotherapy and the services of a social worker if required. The service was free of charge. Initiation of the programme was driven by the PHO management and general practitioners. The programme design was developed by a predominantly Maori team who were mindful of early evaluation feedback about the successes and weakness of other initiatives around the country.6 A popular modality used in other PMHIs was cognitive behavioural therapy (CBT). Although one early commentator argued that this cognitive VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 243 ORIGINAL SCIENTIFIC PAPERS IMPROVING PERFORMANCE approach may not work well for Maori,8 recent research on how CBT might be adapted to be responsive to Maori looks promising.7 However, the WTP team decided on a more holistic approach with spiritual and whanau dimensions. The programme was embedded in a Maori framework and ethos that embraced all comers. Practice guidelines embraced a holistic view of the person founded on Te Whare Tapa Wha model9 and on a parallel Pacific Fonofale model (developed by Traci Tuamaseve, Trail Media, Flaxmere). Key features considered important and integral to the programme were: • an integrated model at both structural and therapeutic levels • considerable flexibility in a range of domains (e.g. venue, time, style of engagement and personnel) to enable appropriate responsiveness • a polymodal, systemic, strengthsbased therapeutic approach • tikanga Maori–based practice, including whakawhanaungatanga (acknowledging relationships), whakawatea/whakanoa (creating a sense of safety), whakapapa/pepeha (reclaiming ancestral origins), waiata/haka composition (expression of new beginnings) and decolonisation techniques (dealing with the sense of historical injustice) • inclusion of whanau, when appropriate • prompt, face-to-face engagement with the programme and therapist. Tikanga Maori–based practice is delivered from a contemporary understanding of a traditional Maori worldview that provides a context to remember, reclaim and reaffirm the empowering elements of Maori cultural heritage. The practice provides spaciousness that aims to counteract the negative effects of colonisation and proactively provide the potential to expand, transform and heal. The programme was led by an experienced Maori service development manager and a Maori therapist, the lead practitioner, with a programme advisory team and support from the PHO management and board. The therapeutic team comprised the full-time lead practitioner, a part-time Pacific social worker and a number of therapists or counsellors who were contracted part-time to 244 the programme. The selection of therapists, most of whom were Pakeha women, was based on both cultural competence and appropriate professional scope of practice. Cultural supervision was provided by the lead practitioner. The social worker assisted service users with a range of practical problems, including liaison and advocacy around access to benefits and services, assistance with budgeting and food, and transport to therapy and other services. The team liaised closely with PHO clinical staff, providing programme information and training, including cultural competency, where necessary. Some data not required by the contract but considered necessary to monitor for meaningful and positive outcomes (such as attendance rates, onward pathways and a satisfaction survey) were collected. Assessment Evaluation method The Eastern Institute of Technology Hawke’s Bay was commissioned by the PHO to undertake an evaluation of the WTP in early 2008. The evaluation objectives were to describe the service-user population and their use of the service, report on effects of the therapeutic treatment in terms of changes to K10 scores, and assist programme improvement by identifying implementation process issues. As no service users were interviewed personally, the Central Regional Ethics Committee confirmed that the evaluation did not need ethics approval or expedited review.10 The evaluation covered the first 15 months of the programme: 1 April 2008 to 30 June 2009. A mixed methods approach was used to provide a picture of formative issues and programme processes and outcomes.11 Qualitative data were obtained from in-depth interviews or focus groups with key informants, the formative evaluator’s notes on team meetings, and document review. Open-ended interview questions focused on participants’ views of programme progress, what they considered key programme strengths and difficulties, and what recommendations they had for improvement. In the early phase, 13 key people were interviewed as part of the formative evaluation. Face-to-face key informant interviews or small focus groups were then undertaken midway (n=15) and at the end of VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS IMPROVING PERFORMANCE the first 15 months (n=21), with many people interviewed on both occasions. Those interviewed at least once included therapists (6), social worker (1), community worker (1), general practitioners (GPs) (10), nurse (1); admin/management staff (6). The 10 GPs interviewed had made 72% of all GP referrals to the programme. Quantitative data were obtained from the programme database. This comprised service user demographics, patterns of service use and K10 pre- and post-treatment scores. In addition, service-user satisfaction surveys, which comprised both quantitative and qualitative components, were analysed. For this paper, some descriptive quantitative data for the programme’s second MoH contract (July 2009 to June 2010) have also been included and compared to the previous period. Outcome evaluation results Quantitative data The quantitative data (Table 1) indicated that the programme successfully reached most of its target population, with particularly high engagement of Maori women, high appointment attendance rates and improvement in average exit K10 scores, the latter being slightly higher than the 80% average shown in the national evaluation.3 The low nonattendance (DNA) rates recorded are considered very good for therapy in the primary health setting,12,13 and are likely to be particularly so for this demographic group. All the above measures were demonstrated to improve over the two-year period. The engagement of Pacific peoples and men, however, was low and remained so despite efforts at improvement. Entry K10 scores tended to be considerably higher than the admission criteria. In the first 15-month period, the average score was 33.4, with 73% of users having a score of 30 or more. Although exit K10 scores averaged 25.3, a highly significant (p<0.001) improvement, this was still in the mid-range of scores denoting a mild/moderate depression or anxiety. In the first 15-month period, the majority (79%) of service users were diagnosed with some form of depression (depression, moderate depression, reactive depression, postnatal depression). Other diag- Table 1. WTP quantitative outcomes 2008–2009 % (n) 2009–2010 % (n) Maori 65% (138) 71.5% (236) European Referrals 23% (49) 25.8% (85) Pacific 3% (6) 1.5% (5) Other 9% (19) 1.2% (4) Females 75% 74% Males 24% 26% No. of packages of care 257 330 No. of therapeutic sessions 1228 1860 Average sessions per package of care 4.8 5.6 No. of social worker interventions 204 262 84.5% Attendance rates 86.4% Did not attend (DNA) rates 10.2% 5.3% Late cancellation 3.4% 10.2% Kessler 10 score—reduced 87% 83% Kessler 10 score—increased 5% 8% Kessler 10 score—no change 8% 9% noses included anxiety, post-traumatic stress disorder, stress, grieving and relationship problems. Anecdotal reports of high levels of distress and comorbidities were confirmed by the K10 scores, stakeholder interview data and the programme database, with 71% of service users having physical and/or mental comorbidities recorded. Those who required more than the standard package of care were either referred back into the programme (4%) or on to other services (20%). Service user feedback Service users completed a short satisfaction survey approximately six weeks after exiting the programme. Results must be viewed knowing that this was undertaken over the phone by a therapist and thus responses may be more positive than if collected by an independent person. Within the first 15-month period, 113 of the 170 people who had exited the programme at least six weeks before had provided feedback via the survey. Table 2 details responses to the closed response questions. Some did not answer all questions and percentages are of those who responded. They show that a significant majority of those providing feedback responded positively to all questions. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 245 ORIGINAL SCIENTIFIC PAPERS IMPROVING PERFORMANCE Table 2. WTP satisfaction survey responses six weeks post-exit Yes n (%) No n (%) More content and happier in yourself? 88 (84%) 17 Healthier from a spiritual point of view? 73 (73%) 27 Stronger in yourself as a Maori or other cultural group you identify with? 67 (72%) 26 More able to set goals for yourself? 86 (81%) 20 More able to think and feel positive? 82 (77%) 24 More able to manage unwelcome thoughts and feelings? 72 (67%) 35 More committed to having good physical health? 73 (72%) 28 More able to communicate with your whanau? 95 (88%) 13 More confident in your relationships with other people? 87 (81%) 20 105 (98%) 2 As a result of the counselling/therapy programme do you generally feel: As a result of the counselling/therapy programme are you: Were you happy with our service? Responses to the small number of open-ended questions gauging satisfaction with the service were very positive, with strong appreciation expressed for the kindness and support of the therapists and social worker, the availability of Maori therapists, the fact the service was no cost, the flexibility around venue and the ability to involve whanau members. Stakeholder feedback Stakeholder interviews showed strong confidence in the WTP team and good relationships between the team and primary care practices. General practitioners felt many of their patients did not seek help unless their situation was serious and, whilst previously they might have avoided probing a problem in the 15-minute consultation, they reported an increased willingness to pursue a patient’s depression or anxiety because of their confidence in the programme now available to them. Some changes in GP prescribing practices occurred, with several reporting they used medication less frequently or delayed use to see if therapy was useful. General practitioners also reported seeing important and significant changes in many of their referred patients, such as visible reductions in anxiety and distress and, in some cases, major lifestyle changes. 246 Both GPs and therapists expressed ambivalence about the cultural and social appropriateness of the K10 assessment tool for this high-needs Maori community, but it was considered useful to the extent that entry and exit scores could be compared. Participating GPs, who were very experienced in this community, reported, however, that they preferred their own clinical judgment to the K10 score when assessing the appropriateness of the referral. Therapists trusted their judgment and liaised closely with them. The WTP commitment to accepting those with K10 scores higher than the referral criteria meant there was no temptation for GPs to adjust scores downward in order for the person to be admitted to the programme, as anecdotal comments suggest may have occurred elsewhere. Strategies for quality improvement/change This section describes quality improvement strategies adopted by the team to address issues identified through formative evaluation feedback and internal review. To effectively meet the complex needs of many referred, more than the standard package of care was often required. This was achieved through referral back into the programme for another package of care or transitioning on to other therapeutic or social support services. The latter required the programme manager to seek alternative funding options and liaise with a range of agencies (e.g. Work and Income NZ, Ministry of Social Development and Accident Compensation Corporation). The team found the requirement of GP referral to the programme limited entry for those who did not use GP services, so direct access through whanau or self-referral was initiated in the second year. Rangatahi/youth under 18 years were initially not eligible for admission to the programme, but the holistic therapeutic model placed importance on the integrity of the whanau so strategies such as including rangatahi in whanau sessions and liaising with existing youth services were explored as necessary. In the second year, those aged 12 years and over became eligible for referral to the programme and were seen by either the programme’s therapeutic team or the team of a contracted dedicated youth service. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE ORIGINAL SCIENTIFIC PAPERS IMPROVING PERFORMANCE Key informants considered reasons for relatively low Pacific referral rates were complex, but included Pacific community sensitivity around confidentiality, unfamiliarity with seeking help from outsiders, low usage of GP services through which referral occurred, GPs not recognising need because of cultural or language barriers, and the lack of a Pacific therapist. To optimise effective engagement with those referred, the team involved the Pacific social worker in first sessions if appropriate. Reasons for low male referral rates were seen to be a combination of men being reluctant to access therapy and the shortage of male therapists. A male therapist was later recruited and a men’s group programme, Tane Toa, funded from an alternative source, was started in the second year. The scarcity of available Maori, Pacific and male therapists was an ongoing issue. To increase the pool of male and Maori therapists for the programme, a mentoring system was developed with senior psychotherapy students from the regional Eastern Institute of Technology. Ensuring availability of Maori and male therapists remained a priority and, at the time of writing (September 2011), the therapeutic team comprises five therapists; three are Maori, one of whom is male. Discussion Programme data showed that the WTP successfully reached and maintained engagement with many of the historically underserved target population. In addition, it showed successful outcomes for many of those finishing the programme, with ongoing referral if necessary. Key features which differed from many mainstream PMHIs included the use of tikanga Maori to build rapport, enhance feelings of safety and strengthen cultural identity; the emphasis on culturally competent personnel; the availability of a social worker; flexibility around venue and time; and the inclusion of family/whanau if wanted. Issues identified through the formative evaluation, for which quality improvement strategies were developed, were generally specific to high needs, underserved populations. Many of the WTP’s reported successes (improvements in K10 scores, stakeholder support and service-user gratitude) and difficulties (low recruitment of Pacific peoples and men and lack of available Maori, Pacific and male therapist workforce) were a feature of other PMHIs around the country.3 However, high Maori engagement, low non-attendance rates and high percentage improvements in K10 exit scores for this population are distinctive WTP features. Two possible explanations might account for the high entry K10 scores and, although significantly reduced, relatively high exit scores. Firstly, that those referred were the more distressed of a population in which high levels of distress and anxiety are perhaps the norm. Secondly, as key informants in both this and the national PMHI evaluation6 have suggested, that the assessment tool might not be as meaningful or culturally appropriate for this population. Despite ambivalence about the tool, both GPs and therapists were happy with the referral process and service users who required more or ongoing therapy or support were assisted on to other services. An evaluation recommendation, however, was that the K10 entry criteria be reassessed. Perhaps more relevant markers of success for this population were the low DNA rates, GPs’ confidence in the programme and reported observed improvements in their patients. Of note was GPs’ greater willingness to pursue mental health issues in this high needs population and reported changes in their prescribing practices. Lessons and messages A key message to take from the WTP programme is that a holistic, tikanga Maori–based programme delivered with considerable flexibility and skilled, culturally fluent programme management and therapeutic teams who work closely with primary care providers can successfully engage and benefit underserved communities with complex social and mental health needs. The complexity of need and high prevalence of comorbidities amongst Maori service users are consistent with national data14 and, given Maori have higher unmet need for general practitioner services than the general population,15 alternative referral pathways to VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 247 ORIGINAL SCIENTIFIC PAPERS IMPROVING PERFORMANCE PMHIs and a dedicated specialist team, such as the WTP, appear necessary to ensure services to this population. The PMHIs have clearly increased services to, and shown benefits for, the general population, including Maori, and new developments, such as screening tools16,17 and brief interventions in the primary health care setting,18 continue to be assessed. While the general view is that Maori probably benefit best from having both Kaupapa Maori and mainstream PMHI models available,3 the challenge is to ensure that new developments do not disproportionately benefit some groups over others and increase mental health disparities. A recent NZMA position statement, supported by mounting research evidence, has prioritised a focus on reducing health disparities in New Zealand.19 This signals to PMHI funders and providers the importance of ensuring that new and existing services are at least as effective for Maori and others in underserved communities as the general population and that successfully targeted programmes, like the WTP, which are developed around the specific needs in these communities, are sufficiently resourced. ACKNOWLEDGEMENTS We wish to thank the erstwhile Tu Meke First Choice PHO Operations Manager and Board for providing access to WTP data and supporting the writing of this paper and the busy professionals interviewed for the evaluation. We would also like to acknowledge Dr David Tipene-Leach and two anonymous reviewers for their helpful peer review comments. FUNDING We are grateful to Tu Meke First Choice PHO for funding the evaluation and EIT Hawke’s Bay for funding the writing of this paper. COMPETING INTERESTS Donny Riki is the WTP lead practitioner and Tania Luscombe is the WTP project manager. 248 8. Hirini, P. Counselling Maori clients: He whakawhiti nga whakaaro i te tangata whaiora Maori. NZ J Psych. 1997; 26(2):13–18. 9. Durie M. Whaiora: Maori health development. Auckland: Oxford University Press; 1994. 10. Abel S, Marshall B. Evaluation of the Wairua Tangata Project, Tu Meke First Choice PHO’s primary mental health initiative, Hawke’s Bay. Final report, April 2008–June 2009. Hawke’s Bay: Eastern Institute of Technology; 2009. 11. Patton MQ. Qualitative research and evaluation methods. 3rd ed. Thousand Oaks: Sage Publications; 2002. 12. Gilbert N, Barkham M, Richards A, Cameron I. The effectiveness of a primary care mental health service delivering brief psychological interventions: a benchmarking study using the CORE system. Prim Care Ment Health. 2005;3(4):241–251. 13. Hamilton R, Gordon P, Naji S. Service innovation: the first year of a brief psychiatric screening clinic in primary care. Psychiatric Bulletin. 2002;26:218–21. 14. Baxter J. Maori mental health needs profile—summary: a review of the evidence. Wellington: Ministry of Health; 2008. 15. Ministry of Health. A portrait of health: key results of the 2006/07 New Zealand Health Survey. Wellington: Ministry of Health; 2008. 16. Calveley J, Verhoeven A, Hopcroft D. A patient-centred referral pathway for mild to moderate lifestyle and mental health problems: does this model work in practice? J Prim Health Care. 2009;1(1):50–56. 17. Goodyear-Smith F, Arroll B, Coupe N, Buetow S. Ethnic differences in mental health and lifestyle issues: results from multi-item general practice screening. NZ Med J. 2005; 18(1212):U1374. 18. Mathieson F, Collings S, Dowell A. Sub-threshold mental health syndromes: finding an alternative to the medication of unhappiness. J Prim Health Care. 2009;1(1):74–7. 19. New Zealand Medical Association. Health equity position statement. NZ Med J. 2011;124(1330):99. References 1. The MaGPIe Research Group. The nature and prevalence of psychological problems in New Zealand healthcare: a report on mental health and general practice investigation. NZ Med J. 2003;116(1171):1–15. 2. Hutton C, Gunn J. Do longer consultations improve the management of psychological problems in general prac tice? A systematic literature review. BMC Health Serv Res. 2007 May 17;7:71. PMID:17506904. 3. Dowell AC, Garrett S, Collings S, McBain L, McKinlay E, Stanley J. Evaluation of the Primary Mental Health Initiatives: summary report 2008. Wellington: University of Otago and Ministry of Health; 2009. 4. Oakley Browne MA, Wells JE, Scott KM, editors. Te Rau Hinengaro—the New Zealand Mental Health Survey. Wellington: Ministry of Health; 2006. 5. Thomas DR, Arlidge B, Arroll B, Elder H. General practitioners’ views about diagnosing and treating depression in Maori and non-Maori patients. J Prim Health Care. 2010;2(3):208–16. 6. Department of Primary Healthcare and General Practice. Interim report: primary mental health initiatives evaluation. Wellington: School of Medicine and Health Sciences, University of Otago; 2007. 7. Bennett ST, Flett RA, Babbage, DR. The adaptation of cognitive behavioural therapy for adult Maori clients with depression: a pilot study. In: Levy M, Nikora LW, Masters-Awatere B, Rua M, Waitoki W, editors. Claiming spaces: proceedings of the 2007 national Maori and Pacific psychologies symposium; 23–24 November 2007, Hamilton, New Zealand. Hamilton: Maori and Psychology Research Unit, University of Waikato; 2008. p. 83–91. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE CONTINUING PROFESSIONAL DEVELOPMENT COCHRANE CORNER PEARLS Topical antibiotics are probably better than placebo for acute conjunctivitis but most get better anyway Bruce Arroll MBChB, PhD, FRNZCGP; Professor of General Practice and Primary Health Care, The University of Auckland, PB 92019, Auckland, New Zealand; Email: b.arroll@auckland.ac.nz THE PROBLEM: Acute bacterial conjunctivitis is an infective condition in which the eyes become red and inflamed. The condition is not normally serious and usually recedes spontaneously within about a week. People with acute conjunctivitis are often given antibiotics, usually as eye drops or ointment, to speed recovery. If there is any loss of vision, the problem is more serious and needs an immediate eye consult. The benefits of antibiotics to the sufferer of conjunctivitis have been questioned. I am always impressed by the ‘mass hysteria’ of ‘pink eye’ in day care centres where the staff act like it was the plague as opposed to a self-limiting viral infection. However, there are cases of bacterial infection and I am more inclined to give antibiotics if the eye is really stuck down in the morning. CLINICAL BOTTOM LINE: The review of trials found that the signs of conjunctivitis went away more quickly in people taking antibiotics, but the benefits are marginal as in most cases the infection is self-limiting. However, most cases resolve spontaneously with clinical remission being achieved in 65% (95% CI 59–70) by days 2–5 in those receiving placebo. Treatments for acute bacterial conjunctivitis with topical antibiotics Topical antibiotics vs placebo Success Evidence Harms Effective; NNT = 6 (range 6–15) Cochrane review1 No major harms String of PEARLS Practical Evidence About Real Life Situations Musculoskeletal conditions PEARLS are succinct summaries of Cochrane Systematic Reviews for primary care practitioners—developed by Prof. Brian McAvoy for the Cochrane Primary Care Field (www.cochraneprimarycare.org), New Zealand Branch of the Australasian Cochrane Centre at the Department of General Practice and Primary Health Care, University of Auckland (www.auckland.ac.nz/uoa), funded by the Ministry of Health (www. health.govt.nz), and published in NZ Doctor (www.nzdoctor.co.nz.). Pronation may be more effective than supination in correcting pulled elbow Local corticosteroid injections improve symptoms of carpal tunnel Corticosteroid injections are effective for trigger finger Oral steroids are effective short-term treatment for adhesive capsulitis Topical glyceryl trinitrate may be effective in rotator cuff disease Surgery for rotator cuff disease is no more effective than active non-surgical treatment Insufficient evidence of benefit of topical rubefacients for acute and chronic painful musculoskeletal conditions in adults DISCLAIMER: PEARLS are for educational use only and are not meant to guide clinical activity, nor are they a clinical guideline. NNT = numbers needed to treat. An NNT of 6 means that for every six people given the treatment, one additional person will get better due to the effective treatment Reference 1. Sheikh A, Hurwitz B. Antibiotics versus placebo for acute bacterial conjunctivitis. Cochrane Database of Systematic Reviews. 2006, Issue 2. Art. No.: CD001211. DOI: All people residing in New Zealand have access to the Cochrane Library via the Ministry website www.health.govt.nz/cochrane-library VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 249 CONTINUING PROFESSIONAL DEVELOPMENT VAIKOLOA Hearing loss among Pacific peoples Ofa Dewes PhD N VAIKOLOA Pacific Primary Health Care Treasures Vai (water) is a symbol of ‘life-source’ and koloa (treasures) to share oa’ia e mauri. Pacific peoples in Aotearoa New Zealand represent a distinct population group. With a population of approximately 266 000, we are undergoing significant demographic changes and growing rapidly into a sizeable number and social significance.1,2 Low health indicators, particularly in relation to non-communicable diseases and capacity and capability gaps that impact service delivery, highlight for us the urgent need for more effective and efficient services, including models of care for Pacific peoples and the development of Pacific capacity and capabilities.3,4,5,6 A health indicator in the health and disability sector which falls into the non-communicable disease category and calls for our attention is hearing loss. Hearing loss carries a financial burden on individuals, their families, and on governments. The effects of hearing impairment can impact our children’s education, social development and quality of life.7,8,9 Among adults, hearing loss can result in poor communication, social isolation, reduced employment opportunities and poor quality of life.7,8,9 Untreated ear disease can have serious clinical consequences such as middle ear erosion and chronic infections as well as brain abscesses and meningitis. According to the World Health Organization, hearing loss and ear disease is considered to be highly prevalent in developing countries with estimates of up to 21% of the population affected.10,11 J PRIM HEALTH CARE 2012;4(3):250–251. CORRESPONDENCE TO: Ofa Dewes School of Population Health, The University of Auckland, PB 92019, Auckland, New Zealand o.dewes@auckland.ac.nz 250 There is some anecdotal evidence on population groups in the Pacific region which supports high prevalence estimates. However, services are fragmented and variable. Services are generally ad hoc or focused around specific non-governmental organisations with voluntary help provided by nurses, hearing therapists, audiologists and ENT surgeons. Our knowledge and understanding of hearing loss prevalence among Pacific peoples is also limited. Credible information and evidence to help leverage appropriate resources to meet the needs of Pacific peoples and inform national policy development and services are required. Motivations to reduce the burgeoning cost of hearing loss through preventative efforts and decisive action that will lead to service and clinical quality improvement and innovation in service delivery for Pacific peoples are warranted. In New Zealand the government is committed to leading significant change in the health and disability sector through its increased focus on evidence-based decisionmaking processes and by prioritising resources for frontline services.12,13,14 Here at the School of Population Health we are committed to overcoming this knowledge gap and developing a strong relationship with other universities in the region, especially around research and education initiatives. Growing interest is emerging to utilise Pacific researchers and the expertise in our Audiology section to look into the nature of hearing loss and assist with the development of sustainable services for Pacific children, young people and adults. This interest has led us to conduct an epidemiological pilot study on hearing loss among the Fijian (Indigenous Fijian, Indo-Fijian, Rotuman) population in Auckland to determine the prevalence of hearing loss and ear disease in this group. Study findings have been disseminated to the Fijian communities in Auckland and in Fiji, highlighting the need to strengthen the global discourse on hearing loss among Pacific peoples, to identify their needs and regional distribution, inform service development and explore opportunities to develop a critical mass of Pacific frontline workers including audiologists and hearing therapists to provide and maintain hearing services for Pacific peoples. We are committed to strengthening and maintaining collaborative relationships with key stakeholders and Pacific communities to develop sustainable hearing services for Pacific peoples. We are also committed to generating new knowledge and VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE CONTINUING PROFESSIONAL DEVELOPMENT POTION OR POISON? articulating the issues that Pacific peoples face, as well as encouraging public debate to improve the hearing health for Pacific peoples. In this connection, some of the implications for future work on hearing loss among Pacific peoples may include the need to: Colloidal silver • overcome data limitations by collecting credible evidence, investigating the hearing health of Pacific peoples by ethnicity, and describing the prevalence of health conditions of each group • know what’s important and works for Pacific peoples and their lived realities in order to improve care for them and maintain relationships • understand the issues within the institutional framework in which they currently exist • work across other sectors like education • support local ownership of services that are adaptable, innovative and responsive to the needs of Pacific peoples, and • develop local capacity and capability to support hearing services for and with Pacific peoples. Silver is a metallic element. Through the ages there has been the belief that silverware used for water storage had disinfecting properties. People are exposed to silver, usually in tiny amounts, through environment and in certain work activities or hobbies. Colloidal silver is a suspension of very small metallic silver particles in a liquid base. It has been marketed in various forms for many years as a topical antimicrobial preparation and as a dietary supplement to stimulate the immune system. References 1. Statistics New Zealand. Pacific Peoples: 2006 Census of Population and Dwellings. Wellington: Statistics NZ; 2007. 2. Statistics New Zealand and Ministry of Pacific Island Affairs. Demographics of New Zealand’s Pacific population. Wellington: Statistics NZ; 2010. ISBN 978-0-478-35344-0 (online). 3. Ministry of Health. A portrait of health: key results of the 2006/07 NZ Health Survey. Wellington: Ministry of Health; 2008. 4. Ministry of Health. Improving quality of care for Pacific peoples. Wellington: Ministry of Health; 2008. 5. Ministry of Health. The health of Pacific peoples. Wellington: Ministry of Health; 2005. 6. Gribben B. Improving access to primary health care: an evaluation of 35 reducing inequalities projects. Overview. Wellington: Ministry of Health; 2007. 7. Greville A. Hearing impaired and deaf people in New Zealand: an update on population numbers and characteristics. 2005. [cited 2011 Nov]. Available from: http://www.grevilleconsulting.co.nz/HearingimpaireddeafpeopleNZMar05a.pdf. 8. Ministry of Health. Pacific peoples’ experience of disability: a paper for the Pacific Health and Disability Action Plan Review. Wellington: Ministry of Health; 2008. 9. Ministry of Health. Pacific child health: a paper for the Pacific Health and Disability Action Plan Review. Wellington: Ministry of Health; 2008. 10. World Health Organization. Deafness and hearing impairment. 2010. Available from: http://www.who.int/mediacentre/factsheets/fs300/en/. 11. World Health Organization, 2008. The global burden of disease: 2004 update. Available from: http://www.who.int/ healthinfo/global_burden_disease/2004_report_update/en/. 12. Ministry of Health. Tupu Ola Moui: Pacific Health Chart Book. Wellington: Ministry of Health; 2004. 13. Ministry of Health. Making a Pacific difference: strategic initiatives for the health of Pacific peoples in New Zealand. Wellington: Ministry of Health; 1997. 14. Minister of Health and Minister of Pacific Island Affairs. ‘Ala Moui: Pathways to Pacific Health and Wellbeing 2010–2014. Wellington: Ministry of Health; 2010. David J Woods BSc (Hons), MPharm, MPSNZ, MRPharmS, FNZHPA PREPARATIONS: Available products range from liquids for oral adminis- tration to topical gels, ointments and sprays, usually supported by websites that describe the manufacturing process and their silver content. ACTIVE CONSTITUENTS: Elemental silver, usually its ionic form, is pro- posed as the active ingredient as an antimicrobial and immunostimulant. MAIN USES AND CLAIMS: Silver in various forms and salts has had medicinal uses since ancient times. Most recently, silver nitrate has been used to treat corns and warts and silver sulfadiazine to treat burns. These uses should not be confused with the claims associated with colloidal silver. Colloidal silver products are often marketed as dietary supplements or topical products with various health-related claims, including that it kills disease-causing agents such as bacteria, viruses, and fungi (alternative to prescription antibiotics); treats diseases such as cancer, HIV/AIDS, diabetes, tuberculosis, syphilis, scarlet fever, shingles, herpes, pneumonia and prostatitis; promotes bladder and lung health; and stimulates the immune system. EVIDENCE FOR EFFICACY: Official drug compendia such as the United States Pharmacopeia and National Formulary have not listed colloidal silver products since 1975 due to lack of evidence of effectiveness and growing concerns about possible toxicity. Colloidal silver products are still advertised and available in New Zealand as an alternative therapy. The scientific literature comprises numerous studies showing that silver in various forms does possess in vitro antimicrobial activity, but there is no clinical evidence of effectiveness. Whilst the intake or application of small amounts of colloidal silver is unlikely to be harmful, chronic use or use of Herbal medicines are a popular health care choice, but few have been tested to contemporary standards. POTION OR POISON? summarises the evidence for the potential benefits and possible harms of well-known herbal medicines. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 251 CONTINUING PROFESSIONAL DEVELOPMENT POTION OR POISON? preparations with unstated or high silver content may lead to significant adverse effects.1 A recent review by the National Center for Complementary and Alternative Medicine in the United States2 concluded that: • silver has no recognised function in the body and is not an essential mineral supplement • silver should not be promoted as a cure-all or to treat a ‘deficiency’ of silver in the body • claims made about the effectiveness of colloidal silver products for numerous diseases are unsupported scientifically • colloidal silver products can have serious adverse effects • the amounts of silver in supplements can vary greatly, which can pose risks to the consumer. Summary Message Colloidal silver is widely marketed for a variety of health benefits. A range of potentially useful in vitro effects have been shown, but no scientific evidence of actual patient benefit. Adverse effects are possible and are more likely with products with high silver content or long-term use. References 1. Fung MC, Bowen DL. Silver products for medical indications: risk-benefit assessment. J Clin Toxicol. 1996;34(1):119–126. 2. Colloidal silver products. A review by the National Center for Complimentary and Alternative Medicine (NCCAM); August 2010. [cited 2012 Jan]. Available from: http://nccam.nih.gov/ health/silver. ADVERSE EFFECTS: Adverse effects are relatively unlikely if a product with a stated low amount of colloidal silver is used for a short time. Longterm use can lead to argyria (silver salts deposit in the skin, gums, nails, eyes, and internal organs causes a permanent bluish-grey discolouration).1 Other possible adverse effects include neurological problems, renal damage, gastrointestinal upset, headaches, fatigue, and skin irritation. DRUG INTERACTIONS: Although there is no supporting clinical evidence, it is possible that colloidal silver taken by mouth may interfere with the absorption of some drugs, e.g. penicillamine, quinolones, tetracyclines, and thyroxine. J PRIM HEALTH CARE 2012;4(3):251–252. CORRESPONDENCE TO: David J Woods School of Pharmacy, University of Otago, PO Box 56, Dunedin 9054, New Zealand david.woods@otago.ac.nz 252 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE CONTINUING PROFESSIONAL DEVELOPMENT NUGGETS OF KNOWLEDGE Statins and memory loss Linda Bryant MClinPharm, PGDipHospPharmAdmin, PhD, FNZHPA, FNZCP, FPSNZ, MCAPA ‘I read in the newspaper that statins may cause dementia… so I want to stop my statin’ In February 2012 the FDA Consumer Information website1 warned that statins may cause subtle cognitive impairment such as mild memory loss or forgetfulness and ‘fuzzy’ or unfocused thinking. This uncommon adverse effect is reversible on stopping the statin. The mild cognitive impairment is not dementia. Statins do not prevent or treat dementia per se,2,3 but may help reduce risk of multi-infarct or vascular dementia. It is important for the patient to know that subtle changes in memory or cognitive function does not mean statin-induced dementia. Onset of cognitive impairment may be within weeks or after several years of statin therapy. The median time of onset is about six months.1,2 In a survey of 171 people with self-reported memory or cognition impairment, improvement was reported a median of two to three weeks after stopping the statin. A number of patients were rechallenged, with rapid recurrence of cognitive impairment.4 The mechanism for statin-induced memory loss is unclear, although it is likely to relate to the medicine rather than the extent of cholesterollowering. There is a view that impairment may be related to demyelination of CNS nerve fibres. Although it is difficult to distinguish between any of the statins, it may be that pravastatin, with less tendency to cross the blood-brain barrier or being less potent, has less impact on memory. Practical approach to the person who claims to have statin-induced memory impairment tion for statin-induced cognitive impairment. If a person thinks that their statin is causing cognitive impairment, a possible method to test this is: • Do a baseline TSH, lipid profile and cardiovascular risk assessment. • Administer the Montreal Cognitive Assessment (MoCA, http://www.mocatest.org/ )—takes 10–15 minutes. • Stop statin for three months. • Repeat MoCA and lipid profile. • If cognitive function improves, rechallenge with a statin, perhaps pravastatin. Rechallenge is important in case the decline in cognitive function was coincidental to stress or depression. Statins may cause mild cognitive impairment: • Assess cognitive function • Stop statins for three months • If function improves, rechallenge with statins And another potential adverse effect—fatigue A recent study found that statins may cause low energy and fatigue on exertion in up to 20% of people, particularly women.5 Benefits of statins are well established, and they certainly have a mortality benefit. Application of this study is to use statins for people with a calculated risk of having a cardiac event the next five years >15%. For people with a lower risk, the potential impact of fatigue on quality of life and activity may outweigh the benefits. References 1 2. 3. 4. Because cognitive impairment is mild there is a need to identify subtle changes in cognition, usually at a high functioning level. It is difficult to find a suitable evidence-based test of cogni- KEY POINTS 5. FDA expands advice on statin risks (for Consumers) 29 Feb 2012. http://www.fda.gov/ForConsumers/ConsumerUpdates/ucm293330.htm. [cited 2012 Jun 15]. McGuinness B, O’Hare J, Craig D et al. Statins for the treatment of dementia. Cochrane Database Syst Rev. 2010 (8):CD000714. McGuinness B, Carig D, Bullock R, Passmore P. Statins for the prevention of dementia. Cochrane Database Syst Rev. 2009 (2):CD003160. Evans M, Golomb B. Statin-associated adverse cognitive effects: survey results from 171 patients. Pharmacotherapy. 2009;29:800–11. Golomb B, et al. Effects of statins on energy and fatigue with exertion: Results from a randomized controlled trial. Arch Intern Med 2012; doi:10.1001/archinternmed.2012.2171. NUGGETS of KNOWLEDGE provides succinct summaries of pharmaceutical evidence about treatment of common conditions presenting in primary care and possible adverse drug reactions. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):253. CORRESPONDENCE TO: Linda Bryant Clinical Manager, Clinical Advisory Pharmacist, East Health Trust PHO PO Box 38248, Howick Auckland, New Zealand l.bryant@auckland.ac.nz 253 VIEWPOINT From ladder to platform: a new concept for pain management Lawrence Leung MBBChir(Cantab), MFM(Clin), CCFP, FRACGP, FRCGP ABSTRACT BACKGROUND: Pain remains one of the top five reasons for consultations in general practice, presenting either alone or as comorbidity. The World Health Organization (WHO) analgesic ladder proposed in 1986 has been the cornerstone of pain management, but is often inadequate in daily practice, especially when dealing with the diverse nature and etiology of various pain conditions. There is a need for a better concept which is universally applicable that acknowledges the value of, and need for, other domains of treatment for pain. OBJECTIVE: This article reviews the original ideas of the WHO analgesic ladder and proposes its extension to a platform model in the context of pain management. DISCUSSION: Pain affects both the physical and psychological wellbeing of patients and should not be treated with pharmacotherapy alone. The model of WHO analgesic ladder provides guidelines for choosing the analgesic agents, but has its limitations. Incorporating the latest paradigm of neuromatrix theory, both acute and chronic pain should be best managed with a broader perspective incorporating multimodal non-pharmacological and supportive treatments, illustrated by the concept of interacting domains on a broad platform as presented in this article. Different levels of pain severity and chronicity necessitate different analgesic platforms of management, and the clinician should move up or down the appropriate platform to explore the various treatment options as per the status and needs of the patient. KEYWORDS: Analgesic ladder; pain management; analgesic platform The original WHO analgesic ladder J PRIM HEALTH CARE 2012;4(3):254–258. CORRESPONDENCE TO: Lawrence Leung Department of Family Medicine, Queen’s University, PO Bag 8888, Kingston ON K7L 5E9, Canada leungl@queensu.ca 254 In 1986, the World Health Organization (WHO) published a set of guidelines regarding the use of analgesics in treating cancer pain.1 It described a three-step approach of sequential use of pharmacological agents commensurate with the pain level as reported by the patient. This stepwise concept with practical recommendations eponymously became the analgesic ladder which was later translated into 22 languages and became one of the most adopted standards for general pain therapy in the next three decades (see Figure 1).2 Moving up from no treatment, the original ladder model starts with non-opioids (e.g. aspirin, paracetamol or nonsteroidal anti-inflammatory drugs, NSAIDs) for mild pain, then increasing to weak opioids like codeine and its derivatives as the second step for intermediate level of pain, and finally escalating to strong opioids like morphine, methadone and even fentanyl as the third step for the highest level of pain (Figure 1). Apart from the choice of pharmacological agents, the original analgesic ladder states that:3 • assessment of the patient is necessary prior to initiation and at regular intervals of therapy • oral form of analgesics is preferred wherever possible • analgesic should be given at regular intervals rather than on demand • there is no standardised dosage and therapy should be individualised according to the level of pain as perceived • the central aim is to relieve as much pain as possible • adjuvant treatment should be added where necessary. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE VIEWPOINT Advantages of the analgesic ladder The original ladder was put forward to address the prevalent obstacles of effective cancer pain relief at that time, which included inadequate training amongst carers in cancer pain management, overt fear of addiction, poor drug availability and lack of public awareness. It both legitimises and rationalises the use of opioids in a cost-effective way for cancer pain. Two years after launching, the analgesic ladder was already validated to be useful in 80–90% of cases.4 The guidelines were further revised in 1997 and their efficacies have stood the test of time. Because of its step-by-step concept, the ladder approach is extended to management of acute and chronic pain, and modified in various specialties. Controversies and limitations of the WHO ladder Right from the start We may take for granted that NSAIDs are the bedside norm for mild acute pain, the evidence for standalone use in mild chronic and cancer pain is still inconclusive. Studies, however, have shown additional benefits of NSAIDs when used in combination with opioids for moderate to severe pain.5–7 North American physicians are less enthusiastic in general than their European colleagues in prescribing long-term NSAIDs for fear of gastrointestinal and renal complications.8 Thus said, in managing chronic pain, it would not be advisable to proceed to fentanyl patches too soon without properly titrating the actual opioid requirement. In fact, for cancer-type pain with breakthrough phenomena, fentanyl patches may not be a suitable choice due to their 72-hour longacting nature and relative lack of dose gradation. Taboo of switching agents or using more than two agents in the same rung The original ladder advises against using two or more agents of the same rung simultaneously.3 Nowadays, this advice is seldom adhered to. In managing mild acute pain, it is not unusual to prescribe both paracetamol and NSAIDs in step 1 of the ladder. For moderate to severe chronic and cancer pain, it is now standard practice to give a long-acting opioid for basal control of pain plus a short-acting opioid for breakthrough pain.14 In managing rheumatological pain, it is Figure 1. The original WHO analgesic ladder The value of the second step The role of weak opioids (e.g. codeine, dihydrocodeine, dextroproxyphene and tramadol) as the second step is also intensely debated. Two studies have shown better pain control with these step 2 agents as compared to morphine.9,10 One randomised controlled trial favoured using strong (step 3) opioids as initial agents for cancer pain and question the need of weak opioids;11 however, the validity of this was confounded by the use of a number of opioids at varying doses and 50% controls do not have matching levels of pain. Two other studies using transdermal fentanyl patches (step 3 opioid) demonstrated that 25µg/hr patches produce good pain relief in opioid-naive patients with chronic pain, omitting the need for step 2 of the WHO ladder.12,13 This shows the stepwise recommendation of pharmacotherapy from non-opioids (e.g. acetaminophen, aspirin, NSAIDs), weak opioids (e.g. codeine, hydrocodone, tramadol) to strong opioids (morphine, hydromorphone, fentanyl) VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 255 VIEWPOINT analgesics are poised in an interactive way. These domains include: Figure 2. The concept of the neuromatrix theory for pain Itself visualised as an entity (like an incessant spinning sphere) comprising the somatosensory (S), cognitive (C) and affective (A) domains, it receives inputs from areas of the brain governing sensation, emotions and cognitions and, in return, churns out a neurosignature (output) which activates various programmes for pain recognition, motor response, emotional and stress reactions. (Adapted from Melzack, Evolution of the neuromatrix theory of pain. The Prithvi Raj Lecture: presented at the third World Congress of World Institute of Pain, Barcelona 2004. Pain Pract. 2005 Jun;5(2):85–94.) often necessary to try two or three NSAIDs and even use them in combination, as failure of one agent does not predict inefficacy of another.15 Similarly, switching between different types and deliverable forms of potent opioids (step 3 of the analgesic ladder) in cancer pain is now commonly practised to achieve better outcomes and lesser toxicity.16 From rungs to platforms: time to think broad Pain in human beings, whether acute or chronic, has both physical and psychological components. Theories of pain have also moved from the original linear concept of gate-control17 to the latest three-dimensional neuromatrix system,18 incorporating the emotional and cognitive domains with the processing of pain signals in the central nervous system (see Figure 2). In the same line of thought, management of pain should no longer be a linear direction going up or down the narrow rungs of a ladder. Rather, it should be conceptualised as various levels of analgesic platforms extending horizontally from the rungs of the analgesic ladder, on which other domains to alleviate pain in addition to the recommended 256 • adjuvant pharmacological agents like muscle relaxants (e.g. cyclobenzaprine, baclofen and dantrolene), anticonvulsants (e.g. gabapentin, pregabalin and lamotrigine), antidepressants (e.g. tricyclics, SSRI, SNRI), injectable agents (steroids, local anaesthetics), compounds that act synergistically with opioids like cannabinoids (nabilone) • physiotherapy and physical therapy • surgical and neurosurgical procedures (e.g. spinal cord stimulation, deep brain stimulation, spinal delivery of opioids, ganglion ablation by phenol or electrofrequency, sympathectomy) • cognitive behavioural therapy and psychological counselling • interpersonal reinforcement (e.g. support group) • mind–body integration (e.g. yoga, meditation and religious support) • hypnosis and relaxation therapy • acupuncture and chiropractic • other complementary and alternative medicine (CAM) options. Each of these domains can be present in every platform and should be considered as an adjunct where appropriate. Depending on the chronicity and severity of the pain and physical/psychosocial construct of the patient, the clinician moves up or down the platforms in time enlisting different domains (Figure 3), similar to the traditional WHO analgesic ladder which stipulates different choice of pharmacological agents. However, unlike the traditional ladder which needs frequent modifications, either with extra rungs2,19 for cancer or severe types of pain, or fast-tracking2 in cases of acute pain, this novel concept of analgesic platform model is universally applicable to all pain scenarios, capable of describing the dynamics of pain management in a broader and holistic manner. Relevance in general practice This platform modification of the original WHO analgesic ladder extends the concept of pain management from an up–down linear pharmacological adjustment to a broader three-dimensional VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE VIEWPOINT Figure 3. Change of concept from the analgesic ladder to the analgesic platform A—Physiotherapy and physical therapy | B—Mind–body integration (e.g. yoga, meditation and religious support) | C—Hypnosis and relaxation therapy | D—Acupuncture | E—Chiropractic | F—External rub/lotions | G—Other CAM options (Tai chi, Tui Na) | H—Muscle relaxants (e.g. cyclobenzaprine, baclofen and dantrolene) | I—Injectable agents (steroids, local anaesthetics) | J—Interpersonal reinforcement (e.g. support group) | K—Anticonvulsants (e.g. gabapentin, pregabalin and lamotrigine) | L—Antidepressants (e.g. tricyclics, SSRI, SNRI) | M—Compounds that act synergistically with opioids like cannabinoids (nabilone) | N—Cognitive behaviour therapy and psychological counselling | O—Surgical and neurosurgical procedures (e.g. spinal cord stimulation, deep brain stimulation, spinal delivery of opioids, ganglion ablation by phenol or electrofrequency, sympathectomy) perspective that encompasses other domains and disciplines of therapies. Adopting this novel analgesic platform concept will not only remind practising clinicians to think and act broadly in seeking solutions, it will also align the treatment goals with the modern neuromatrix theory in dealing with acute and chronic pain. In primary care, the analgesic platform is of particular relevance as patients with pain often present to their primary care providers with psychosocial overlays that are best approached and managed with a broad and multimodal horizon. Case study 1 A 56-year-old retired manual labourer with four years’ history of chronic back pain as a result of traumatic herniation of L5/S1 disc. He underwent discectomy which gave him little improvement, and he experienced constant sharp pain in his lower back which would often radiate down to both of his thighs. Patient responded initially to non-steroidal anti-inflammatory drugs, but soon proceeded to codeine and hydrocodone. He tried three courses of physiotherapy and noticed very slight benefits. He was given amitriptyline but developed dizziness. As you reviewed his file, a pain specialist recommended that the patient should be moved up the analgesic ladder for stronger choices like morphine or hydromorphone. On examination, patient seemed distressed and tired. Further questioning revealed recent stress in the patient’s family with ongoing financial difficulties. Patient also admitted having poor quality of sleep and depressed mood, albeit lack of suicidal ideation. You realised that there were coexisting psychosocial issues in the patient’s life that could perpetuate and, in fact, exacerbate the chronic pain conditions. Hence, you decide to think broad and adopt the analgesic platform, adding in duloxetine 30mg which would help both of his mood and neuropathic pain and a muscle relaxant to relieve VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 257 VIEWPOINT the muscle spasm. You also suggested a meditation class and structured stretching exercises for his back. Last but not least, you enrolled him for a support group for chronic pain counselling. Patient made steady improvement in the next six months, gradually stepping down to non-opioid platform (NSAIDs) with the addition of a lowdose pregabalin. Case study 2 A 35-year-old professional footballer came to consult about pain in his right knee. A previous injury of that knee had involved partially torn cruciate ligaments. After successful repair, he still felt dull pain and stiffness in the knee at times. He consciously stayed away from any opioid containing analgesia for fear of affecting his professional career. He used up to 3 g of paracetamol plus 1.8 g of ibuprofen and reported the pain as 5/10. You reviewed his musculoskeletal pain and realised that moving up to opioid analgesia is not an option; hence, you went horizontally along the analgesic platform and suggested a combination of physiotherapy, muscle relaxants, massage therapy and acupuncture. In three months, patient went back to his football games with a manageable pain level of 2/10. Case study 3 ACKNOWLEDGEMENTS Diagrams by Arthur Leung. COMPETING INTERESTS None declared. 258 A 65-year-old gentleman suffered from intractable pain due to disseminated carcinomatosis and nerve root compression from metastatic vertebral collapse. He took continuous-release hydromorphone 60 mg twice a day and short acting hydromorphone 8 mg tablets up to six times a day for breakthrough pain. He consulted you for better pain control as part of his palliative care. In view of the high demand of narcotics and the need for constant titration, you discarded the option of fentanyl patch even though it is indicated at the highest rung of the traditional WHO analgesic ladder. Instead you adopted the analgesic platform concept and enlisted the help of neurosurgeon and palliative care physicians, who arranged for the patient to have continuous intrathecal morphine pump. Patient experienced remarkable pain relief and significant improvement in his mood in the following three months until he died. References 1. Ventafridda V, Saita L, Ripamonti C, De Conno F. WHO guidelines for the use of analgesics in cancer pain. Int J Tissue React. 1985;7(1):93–6. 2. Vargas-Schaffer G. Is the WHO analgesic ladder still valid? Twenty-four years of experience. Can Fam Physician. 2010;56(6):514–7, e202–5. 3. Ventafridda V, Tamburini M, Caraceni A, De Conno F, Naldi F. A validation study of the WHO method for cancer pain relief. Cancer. 1987;59(4):850–6. 4. Stjernsward J. WHO cancer pain relief programme. Cancer Surv. 1988;7(1):195–208. 5. Mercadante S, Fulfaro F, Casuccio A. A randomised controlled study on the use of anti-inflammatory drugs in patients with cancer pain on morphine therapy: effects on doseescalation and a pharmacoeconomic analysis. Eur J Cancer. 2002;38(10):1358–63. 6. Dhillon S. Tramadol/paracetamol fixed-dose combination: a review of its use in the management of moderate to severe pain. Clin Drug Investig. 2010;30(10):711–38. 7. Gatti A, Sabato E, Di Paolo AR, Mammucari M, Sabato AF. Oxycodone/paracetamol: a low-dose synergic combination useful in different types of pain. Clin Drug Investig. 2010;30 Suppl 2:3–14. 8. Mercadante S. Management of cancer pain. Intern Emerg Med. 2010;5 Suppl 1:S31–5. 9. Grond S, Radbruch L, Meuser T, Loick G, Sabatowski R, Lehmann KA. High-dose tramadol in comparison to low-dose morphine for cancer pain relief. J Pain Symptom Manage. 1999;18(3):174–9. 10. Mercadante S, Salvaggio L, Dardanoni G, Agnello A, Garofalo S. Dextropropoxyphene versus morphine in opioidnaive cancer patients with pain. J Pain Symptom Manage. 1998;15(2):76–81. 11. Marinangeli F, Ciccozzi A, Leonardis M, Aloisio L, Mazzei A, Paladini A, et al. Use of strong opioids in advanced cancer pain: a randomized trial. J Pain Symptom Manage. 2004;27(5):409–16. 12. Mystakidou K, Tsilika E, Parpa E, Kouloulias V, Kouvaris I, Georgaki S, et al. Long-term cancer pain management in morphine pre-treated and opioid naive patients with transdermal fentanyl. Int J Cancer. 2003;107(3):486–92. 13. Vielvoye-Kerkmeer AP, Mattern C, Uitendaal MP. Transdermal fentanyl in opioid-naive cancer pain patients: an open trial using transdermal fentanyl for the treatment of chronic cancer pain in opioid-naive patients and a group using codeine. J Pain Symptom Manage. 2000;19(3):185–92. 14. Davis MP, Walsh D, Lagman R, LeGrand SB. Controversies in pharmacotherapy of pain management. Lancet Oncol. 2005;6(9):696–704. 15. Bannwarth B. Irrelevance of the WHO analgesic ladder for managing rheumatic pain. Joint Bone Spine. 2010;77(1):1–3. 16. Vadalouca A, Moka E, Argyra E, Sikioti P, Siafaka I. Opioid rotation in patients with cancer: a review of the current literature. J Opioid Manag. 2008;4(4):213–50. 17. Melzack R, Wall PD. Pain mechanisms: a new theory. Science. 1965;150(699):971–9. 18. Melzack R. Evolution of the neuromatrix theory of pain. The Prithvi Raj Lecture: presented at the third World Congress of World Institute of Pain, Barcelona 2004. Pain Pract. 2005;5(2):85–94. 19. Tay W, Ho KY. The role of interventional therapies in cancer pain management. Ann Acad Med Singapore. 2009;38(11):989–97. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE VIEWPOINT Frequently asked questions on measurement of bone mineral densitometry Joseph C Lee MBBS, FRACP;1,2 Nelson K Loh BMedSci, MBBS, FRACP3 Introduction What is the accuracy? Osteoporosis currently afflicts over half of women and almost 30% of men aged over 60 years; this figure will grow as the population ages. Fractures—especially of the hip—result in considerable social and economic costs to the individual and to society, as well as leading to an appreciable mortality. The precision of DXA is approximately 1.0–1.5% for spine and total hip BMD.1 DXA scanners have excellent long-term precision because their calibration is very stable and there are effective instrument quality control procedures (provided by the manufacturers) to detect any long-term drifts. The test–retest variation for spine and hip DXA is 2–3%, resulting in T-score accuracy errors of ± 0.5.2 Bone mineral density measured with dual-energy x-ray absorptiometry is the ‘gatekeeper’ to osteoporosis treatments in New Zealand and the best way to estimate future fracture risk. Hence it is important to be familiar with some of the technical aspects of densitometry and its application in the community. Clarification of some of these areas may facilitate more effective utilisation. What are the indications for bone mineral density (BMD) assessment? The indications for BMD testing, as recommended on the Osteoporosis New Zealand website (http://www.bones.org.nz), are listed in Table 1. As pointed out on the site, BMD should only be measured if it will impact management. 1 Department of Nuclear Medicine, The Prince Charles Hospital, Queensland, Australia 2 School of Medicine, University of Queensland, Queensland, Australia 3 Department of Nuclear Medicine, Sir Charles Gairdner Hospital, Western Australia, Australia What is the radiation dose? The effective radiation dose is 1–10 µSv.3 Context is provided in Table 2. Are spine or hip measurements better? The best way of predicting fracture risk at any given site is to measure BMD at that site—for example, DXA of the proximal femur predicts femoral fracture better than measurements at other sites.4 Similarly, for vertebral or wrist fractures. The spine is the best site for follow-up measurements because treatment changes are usually the Table 1. Indications for bone densitometry Women over 60 years and men over 70 years with risk factors such as: Glucocorticoid therapy Parental history of a hip fracture Low body weight (< 58 kg) or body mass index (< 20 kg/m²) History of smoking or heavy alcohol intake Premature menopause in women or hypogonadism in males Rheumatoid arthritis Malabsorption, chronic liver or renal disease Any woman over 65 years or man over 75 years considering specific measures to prevent osteoporosis Any individual prescribed long-term glucocorticoids or other medications associated with osteoporosis, e.g. anticonvulsants, aromatase inhibitors Women with a history of premature menopause Postmenopausal women or older men with a history of minimal trauma fracture (from Osteoporosis New Zealand—www.bones.org.nz) VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE J PRIM HEALTH CARE 2012;4(3):259–261. CORRESPONDENCE TO: Joseph C Lee Staff Specialist Consultant Physician, Department of Nuclear Medicine, The Prince Charles Hospital, Queensland, Australia Joseph_Lee@health. qld.gov.au 259 VIEWPOINT Table 2. Effective radiation dose of DXA compared to urban background and other medical procedures3 Table 3. Clinical risk factors included in FRAX algorithm— www.shef.ac.uk/frax DXA 1-10 µSv Country of geographic origin Chest x-ray 30 µSv Ethnic origin Mammogram 130 µSv Age Annual background (urban environment) 2000 µSv Sex CT pelvis 5000 µSv Weight Height largest at this site and the precision error is lower than at other sites.5,6 Parental history of hip fracture Current smoking status Corticosteroid therapy (current or past) How do the fracture risk calculators compare? Rheumatoid arthritis Secondary osteoporosis The WHO Fracture Risk Assessment Tool (FRAX) algorithm (available at http://www.shef. ac.uk/frax) accounts for 14 risk factors (as listed in Table 3) in calculating 10-year probabilities of hip and major osteoporotic (humerus, forearm, hip or clinical vertebral) fractures.7 It incorporates data from 12 independent fracture studies, over 60 000 men and women (from four continents) Table 4. Pharmac Special Authority for subsidy criteria for alendronate10 One significant osteoporotic fracture demonstrated radiologically and T-score ≤ -2.5 One significant osteoporotic fracture demonstrated radiologically, and either the patient is elderly, or densitometry scanning cannot be performed because of major logistical, technical or pathophysiological reasons Two significant osteoporotic fractures demonstrated radiologically T-score ≤ -3.0 Ten-year risk of hip fracture (according to FRAX or Garvan calculator) ≥ 3% Patient has had a Special Authority approval for zoledronic acid (underlying cause— osteoporosis) or raloxifene The patient is receiving systemic glucocorticosteriod therapy (≥ 5 mg per day prednisone equivalents) and has already received or is expected to receive therapy for at least three months; and • T-score ≤ -1.5; or • one significant osteoporotic fracture demonstrated radiologically; or • Special Authority approval for zoledronic acid (underlying cause— glucocorticosteroid therapy) or raloxifene Table 5. Recommended intervals for serial DXA for detecting development of osteoporosis14 Normal BMD or ‘mild osteopenia’ (T-score not less than -1.5) 15 years ‘Moderate osteopenia’ (T-score between -1.5 and -2.0) 5 years ‘Advanced osteopenia’ (T- score between -2.0 and -2.5) 1 year 260 Previous history of fracture (after age 50) Alcohol intake (≥3 units daily) Hip BMD and 250 000 person-years of follow-up.8 The Garvan Institute fracture risk calculator (http:// garvan.org.au/promotions/bone-fracture-risk/calculator) requires fewer data and emphasises past fractures and falls. A recent study found that the Garvan Institute calculator overestimated hip fracture rates and FRAX (with BMD measurements) underestimated osteoporotic and hip fracture rates.9 However, BMD should continue to be used to calculate fracture risk which informs treatment decisions. How can BMD results be used to plan treatment? BMD and/or fracture risk scores are used as the arbiters for osteoporosis treatment as treatment is considered cost-effective at certain thresholds of fracture risk. For example, Special Authority for Subsidy is granted under the Pharmac schedule10 for alendronate according to the criteria in Table 4. By comparison, the threshold for treatment is set at a hip fracture risk in the next 10 years of 4% in the United Kingdom and Sweden.11 Why are there discrepancies between hip and spine T-scores? A number of factors can spuriously affect the T- (and Z-) score at the hip and spine, including VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE VIEWPOINT degenerative bone/joint disease and aortic calcification which leads to spuriously high BMD measurements.12 However, some conditions do tend to preferentially affect BMD at the spine more than the hip (e.g. steroid use) or the hip over the spine (e.g. hyperparathyroidism). How often should DXA be repeated? The International Society of Clinical Densitometry (ISCD) currently recommends no more frequent than biennial re-testing,13 formulated from the known precision of DXA equipment and that the normal rate of postmenopausal bone loss is 1–2% per year. Prospective clinical study of postmenopausal women concluded that development of osteoporosis is adequately detected by DXA frequency stratified by previous DXA findings for that individual14 (summarised in Table 5). In most instances, this is much less frequent than the minimum two-year interval recommended by ISCD. However, it is also suggested that patients with advanced osteopenia should have repeat testing in one year, which directly contradicts the ISCD’s recommendation. Are DXA results from different facilities interchangable? No. Different facilities use different equipment made by different manufacturers which, in turn, utilise different reference databases. It is best, therefore, to use the same software for repeat scans wherever possible. SUMMARY • DXA is a fast, reliable, low-radiation method for assessing BMD, which is the arbiter for osteoporosis treatment eligibility and the best determinant of future fracture risk. • Web-based fracture risk calculators can utilise BMD data to provide further information for medical practitioners and patients. • The interval for serial BMD assessment for detecting development of osteoporosis should be planned according to the previous DXA findings. 7. Kanis JA, Oden A, Johnell O, et al. The use of clinical risk factors enhances the performance of BMD in the prediction of osteoporotic fractures in men and women. Osteoporos Int. 2007;18:1033–46. 8. Blake GM, Fogelman I. An update on dual-energy x-ray absorptiometry. Semin Nucl Med 2010;40:62–73. 9. Bolland MJ, Siu AT, Mason BH, et al. Evaluation of the FRAX and Garvan fracture risk calculators in older women. J Bone Miner Res. 2011;26:420–7. 10. Wilson K, Bloor R, Jennings D, editors. TEX Schedule February 2012 Volume 19. New Zealand: Pharmaceutical Management Agency; 2012. 11. Chun KJ. Bone densitometry. Semin Nucl Med. 2011;41:220–8. 12. Reid IR, Evans MC, Ames R, et al. The influence of osteophytes and aortic calcification on spinal mineral density in postmenopausal women. J Clin Endocrinol Metab. 1991;72:1372–4. 13. Baim S, Binkley N, Bilezikian JP, et al. Official positions of the International Society for Clinical Densitometry and executive summary of the 2007 ISCD Position Development Conference. J Clin Densitom. 2008;11:75–91. 14. Gourlay ML, Fine JP, Preisser JS, et al. Bone-density testing interval and transition to osteoporosis in older women. New Engl J Med. 2012;366:225–33. References 1. Patel R, Blake GM, Rymer J, et al. Long-term precision of DXA scanning assessed over seven years in forty post-menopausal women. Osteoporos Int. 200;11:68–75. 2. Blake GM, Fogelman I. How important are BMD accuracy errors for the clinical interpretation of DXA scans? J Bone Miner Res. 2008;23:457–62. 3. Fogelman I, Blake GM. Different approaches to bone densitometry. J Nucl Med. 2000;41:2015–25. 4. Marshall D, Johnell O, Wedel H. Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ. 1996;312:1254–9. 5. Blake GM, Herd RJ, Fogelman I. A longitudinal study of supine lateral DXA of the lumbar spine: a comparison with posteroanterior spine, hip and total-body DXA. Osteoporos Int. 1996;6:462–70. 6. Faulkner KG. Bone densitometry: choosing the proper site to measure. J Clin Densitom. 1998;1:279–85. VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE COMPETING INTERESTS None declared. 261 LETTERS TO THE EDITOR Malaysian GAD-7 less sensitive than reported I n their paper on the validation of the Malay version of the GAD-7, Sidik et al. reported acceptable sensitivity (76.3%) and excellent specificity (94.4%) for detecting anxiety disorder in consecutive women attending a primary care clinic.1 However, these results are biased because the authors failed to account for the stratified sampling procedure. When a stratified sampling procedure has been applied, the data must be ‘weighted back’ to mirror the source population in order to produce unbiased estimates of the operating characteristics for that population. Sidik et al. included 895 women who completed a PHQ-9 and GAD-7. Then they divided the group into women with normal scores (PHQ-9 <10 and GAD-7 <5) and women with high scores (PHQ-9 ≥10 and GAD-7 ≥5). From the first group, one in 10 women and from the second group one in two women were selected for a psychiatric diagnostic interview. Fifty participants were excluded, leaving 845 overall participants of whom 146 underwent a diagnostic interview. Sidik et al.’s Table 1 informs us that 38 of the latter 146 women had an anxiety disorder diagnosis and 35 of the 146 women had a positive GAD-7 test (≥8). In addition, Sidik et al. provided the information that 66 of the original 845 women (i.e. 7.8%) had a positive GAD-7 test. We can use this information to try to calculate the expected numbers for Table 1 when all 845 women (representing the source population of women attending the clinic) would have been examined instead of the stratified sample of 146: From a total of 845 women, 66 had a positive and 779 a negative GAD-7. Thus, the 35 women with a positive GAD-7 (from the 146 women examined) represented a sample of the original 66 women with a positive GAD-7. To weight back these women to the source population, they are assigned a weight factor of 66/35=1.9. The 29 women with a positive GAD-7 and an anxiety disorder diagnosis therefore probably represented 29 x 1.9 = 55 women in the source population. Similarly, the six women with a positive GAD-7 and no anxiety disorder diagnosis probably represented 6 x 1.9 = 11 women in the source population. Analogously, the weight factor for women with a negative GAD-7 is 779/111 = 7.0. The nine women with a negative GAD-7 and an anxiety disorder diagnosis probably represented 9 x 7.0 = 63 women in the source population (!). A complete redraw of Table 1 looks like this: Table 1. Numbers weighted back to mirror the source population Anxiety diagnosis No anxiety diagnosis Total GAD-7 positive 55 11 66 GAD-7 negative 63 716 779 Total 118 727 845 Sensitivity of the GAD-7 in the source population was 55/118 = 46.6% (instead of 76.3% in the stratified sample). Specificity was 98.5%, the LR of a positive test was 30.81 and the LR of a negative test was 0.54%. The revised conclusion should be that the GAD-7 did not perform very well in detecting anxiety disorder in the Malaysian women attending a primary care clinic: the GAD-7 missed more than half of all anxiety disorders. Berend Terluin, MD, PhD Dpt. of General Practice and Elderly Care Medicine, EMGO Institute for Health and Care Research, VU University Medical Centre, Amsterdam, Netherlands Reference 1. Sidik S, Arroll B, Goodyear-Smith F. Validation of the GAD-7 (Malay version) among women attending a primary care clinic in Malaysia. J Prim Health Care. 2012;4(1):5–11. Authors’ response In response to the ‘Letter to the Editor’ on the validity results of the GAD-7 (Malay version), the authors would like to maintain that their original calculations are correct. The validity of the GAD-7 (Malay version) was calculated based on the 146 participants interviewed with the CIDI, and not the whole study population (n=845). The objective of the paper was to validate the GAD-7 against the CIDI as the reference standard, which was stated in the abstract and method sections. Therefore, the authors are justified in calculating the sensitivity and specificity of the GAD-7 only among the 146 participants who were interviewed with the CIDI. The remaining 699 respondents were not interviewed with the CIDI, and therefore the validity findings were not generalised to the whole study population of 845. The sensitivity and specificity of a test are independent of the prevalence of a disorder. We accept that the likelihood ratios are based on small numbers but they are also independent of prevalence. Dr Terluin’s extrapolation of results from a subset onto the total population, including those who had not had the reference standard risks, must be interpreted with caution due to the potential magnification of error, where numbers in some groups are actually small. Letters may respond to published papers, briefly report original research or case reports, or raise matters of interest relevant to primary health care. The best letters are succinct and stimulating. Letters of no more than 400 words may be emailed to: editor@rnzcgp.org.nz. All letters are subject to editing and may be shortened. 262 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE FILM REVIEW A good death: a film about end-of-life care and advance care planning Produced by Prof. D Robin Taylor and Paul Trotman Reviewed by Prof. Rod MacLeod PhD, FRCGP, FAChPM; Senior Staff Specialist in Palliative Care, Hammond Care, Sydney, and Conjoint Professor in Palliative Care, University of Sydney, Australia ‘A wise physician will consider whether a disease be incurable: if he find it to be such, let him resort to palliation: and alleviate the symptom.’ Bacon, Sylva, 1626 T he modern palliative care movement has been around for over 40 years and yet many health practitioners still find it hard to provide appropriate management near the end of life. This film is made for people with COPD or serious lung conditions, their families and the doctors and nurses who provide care for them. One of the aims is to introduce the concept of advance care planning and to prompt discussion about patient choices, but the real power of the film is in telling the story of the final illness of Martin Cavanagh as he is accompanied by his wife Tui and their family towards the end of his life. Robin Taylor writes on the website that they ‘wanted to make this film because we realised that the care we were providing to patients with respiratory disease at the end of life was missing the mark.’ On the evidence provided in this film they have little to be worried about. The care captured on film is exemplary. Professor Taylor is shown as empathetic, forthright and honest in his approach to Martin, the patient, Martin’s wife Tui and to the audience of the film. The most striking person on film, however, is Martin himself, who displays dignity and control throughout. One of the sadnesses pointed out by Professor Taylor is that doctors generally ‘do what they are trained to do.’ Sadly, we know that the training provided to young doctors and nurses in end-oflife care is inadequate in many ways. By making this film, the producers have shown that there is power in the telling of stories; there is power in excellent camera work (here by Scott Mouat), and there is power in watching and listening in order to learn. It is well worth listening and watching more than once as there are so many snippets of wisdom contained in the spoken words. There are weak points. Some of the dialogue is hard to hear (especially through Martin’s oxygen mask), but overall the quality of the film-making is very high. The website provides useful web addresses for further information. This is an important piece of work that should be shown to every New Zealand medical student and as many medical specialists and primary physicians as possible. It is an honest way of indicating the importance of advance care planning and a way of showing that it is possible to have a good death from a chronic disabling disease. I recommend it highly. PRN Films, Dunedin, 2012 www.agooddeath.co.nz VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE 263 ABOUT THE JOURNAL OF PRIMARY HEALTH CARE T he Journal of Primary Health Care (JPHC) began publishing in 2009, superseding the previous RNZCGP journal the New Zealand Family Physician. It is an interdisciplinary publication aimed at moving research into primary health care practice and practice into research. This includes the fields of family practice, primary health care nursing and community pharmacy as well as areas such as health care delivery, health promotion, epidemiology, public health and medical sociology of interest to a primary health care provider audience. It is positioned as relevant to countries within the Pacific rim. JPHC publishes peer-reviewed quantitative and qualitative original research, systematic reviews, papers on improving performance and short reports that are relevant to its primary health care practitioners. For the aim, scope, instructions to authors and templates for publications see www.rnzcgp.org.nz/ journal-of-primary-health-care/. JPHC includes pithy digests of the latest evidence including a String of PEARLS (Practical Evidence About Real Life Situations), Potion or Poison? (evidence for the potential benefits and possible harms of well-known herbal medicines), Cochrane Corner (the summary of a Cochrane review), Nuggets of Knowledge (succinct synopses of pharmaceutical evidence for primary care), Pounamu and Vaikoloa, (Maori and Pacific primary health care treasures respectively) and Gems, 100-word outlines of NZ primary care research published in other journals. JPHC publishes viewpoints, commentaries and reflections that explore areas of uncertainty on aspects of care for which there is no one right answer. Debate is stimulated in Back to Back, where two professionals present their opposing views on a topic. There is a regular Ethics column. Letters to the Editor are welcomed. INDEXING The Journal is indexed in MEDLINE, Excerpta Medica (EMBASE), Cumulative Index to Nursing and Allied Health Literature (CINAHL), Scopus and Index New Zealand (INNZ). Complete text of the Journal is available online at http://www.rnzcgp.org. nz/journal-of-primary-health-care/ and through various aggregators including PubMed Central and EBSCO. EDITOR Prof. Felicity Goodyear-Smith: Professor and Goodfellow Postgraduate Chair, Department of General Practice and Primary Health Care, The University of Auckland, Auckland, New Zealand; editor@rnzcgp.org.nz Dr Ofa Dewes: Research Fellow, Department of Pacific Health, School of Population Health, University of Auckland Prof. Tony Dowell: Professor and Head of the Department of Primary Health Care and General Practice, Wellington School of Medicine, University of Otago, NZ Ms Eileen McKinlay: Senior Lecturer in Primary Health Care, Department of Primary Health Care and General Practice, University of Otago Wellington, NZ Prof. Pauline Norris: Professor and Chair in Social Pharmacy, School of Pharmacy, University of Otago, Dunedin, NZ EDITORIAL BOARD The Editorial Board comprises renowned and active primary care clinicians, clinical and scientific academics and health policy experts with both New Zealand and international representation. Dr Barry Parsonson: Psychologist for NZ Ministry of Education and International Consultant, UNICEF (Georgia) Training Project for Institutional Staff working with disabled children, Napier, NZ Prof. Bruce Arroll: Professor and Head of the Department of General Practice and Primary Health Care, The University of Auckland, Auckland, NZ Dr Shane Reti (QSM): International Program Director Clinical Informatics and CEO of Clinical Informatics Industrial Research, Harvard Medical School, USA Prof. Jenny Carryer: Professor of Nursing, School of Health and Social Services, Massey University, Palmerston North, NZ Dr Jo Scott-Jones: General practitioner, Opotiki, Eastern Bay of Plenty Prof. Peter Crampton: Pro-Vice-Chancellor, Division of Health Sciences, University of Otago, Dunedin NZ Prof. Kurt Stange: Professor of Family Medicine, Case Western Reserve University, Cleveland, OH, USA and Editor, Annals of Family Medicine SUBMISSIONS Full instructions for authors can be found at: http://www.rnzcgp.org.nz/information-for-authors Please send all submissions to the Editor: editor@rnzcgp.org.nz JPHC ADMINISTRATION For subscription and advertising queries, or to sign up for email alerts, please contact the Publications Coordinator, RNZCGP, PO Box 10440, Wellington 6143, New Zealand; jphcnz@rnzcgp.org.nz JPHC is printed on uncoated, acid-free paper which meets the archival requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper) and is Forest Stewardship Council (FSC)–certified which meets the highest environmentally responsible standards. The Journal of Primary Health Care is the official journal of the RNZCGP. However, views expressed are not necessarily those of the College, the Editor, or the Editorial Board. ©The Royal New Zealand College of General Practitioners 2012. All Rights Reserved. 264 VOLUME 4 • NUMBER 3 • SEPTEMBER 2012 J OURNAL OF PRIMARY HEALTH CARE