A 4-year evaluation of blood pressure management in Trinidad and Tobago

Transcription

1 Journal of Human Hypertension (1999) 13, Stockton Press. All rights reserved /99 $ ORIGINAL ARTICLE A 4-year evaluation of blood pressure management in Trinidad and Tobago D Mahabir 1 and MC Gulliford 2 1 Nutrition and Metabolism Division, Ministry of Health, Trinidad and Tobago; 2 Department of Public Health Sciences, King s College, London SE1 3QD, UK This paper reports a 4-year evaluation of government primary care services in Trinidad and Tobago. The sample included 16 primary care health centres in Trinidad and Tobago with cross-sectional surveys of adult attenders in 1994 and Data were obtained from clinic records including details of processes of care, drug utilisation and blood pressure (BP) control. Intervention following the initial survey was by means of reports to the Ministry of Health and annual training workshops for medical officers. Data were analysed for 1759 adult attenders at study clinics (981 in 1994 and 778 in 1998). Most attenders had blood pressure (96% in 1994 and 98% in 1998). Among 1176 subjects with hypertension (BP 160/95 mm Hg or treated with drugs) the proportion with dietary advice ever increased from 139/662 (21%) in 1994 to 185/514 (36%) in 1998, and recording of exercise advice increased from 36 (5%) in 1994 to 99 (19%) in Among subjects treated with drugs, use of Brinerdin decreased from 191/625 (31%) in 1994 to 29/486 (6%) in 1998, while use of diuretics, beta-blockers, calcium antagonists and ACE inhibitors increased. The proportion of patients with blood pressure adequately controlled ( 160/95 mm Hg) was 338 (51%) in 1994 and 297 (58%) in 1998 (odds ratio 1.39, 95% confidence interval 0.96 to 2.00). An audit study used to inform health care policy, staff training and clinical practice may contribute to improving processes of hypertension management in a middleincome country like Trinidad and Tobago. Keywords: medical audit; health care delivery; quality of health care; West Indies Introduction Hypertension is judged to be a serious health problem in the Caribbean region whether measured in terms of the high prevalence of the condition, 1,2 or the mortality 3,4 and health service use 5 which result from its complications. Evidence from randomised controlled trials shows that treatment of hypertension can reduce the risk of stroke by 40% and coronary heart disease by approximately 15% 6 but poor quality of care may often diminish the effectiveness of blood pressure management. 7 Our aim in the present study was to evaluate the quality of blood pressure management in government primary care health centres in Trinidad and Tobago. This twin-island republic has a predominantly bi-ethnic population, with some 40% being of African descent and 40% of Indian descent, with most of the remainder being of mixed ethnicity. The per capita GDP of Trinidad and Tobago was US$3600 in with a median monthly income in men of about 150 per month. 9 Government health expenditure was estimated to be 2.8% of GDP in 1990 and this accounted for 63% of total health expenditure. 10 Government primary care services in Trinidad Correspondence: MC Gulliford, Department of Public Health Sciences, King s College London, Capital House, 42 Weston St, London SE1 3QD, UK Received 12 February 1999; revised 25 March 1999; accepted 26 March 1999 and Tobago are provided through a network of approximately 100 health centres which are distributed throughout the country. Health centres do not maintain a register of the local population, instead subjects attend by self-referral. Most health centres run clinic sessions for unselected adult attenders but, because of the importance of diabetes and hypertension in this community, many health centres run specially organised chronic disease clinics for clients with these conditions. In a recently reported survey of medical practitioners views on hypertension management, limited availability of drugs was one of the main concerns of doctors working in the public sector. 11 In 1994 we evaluated blood pressure management in public and private primary care settings in Barbados, Jamaica and Trinidad and Tobago and reported our findings in abstract form In government primary care health centres in Trinidad and Tobago, most adult attenders had their blood pressure, many patients with hypertension were treated, but few hypertensive patients had their blood pressure satisfactorily controlled. 13 There was widespread use of pharmaceutical preparations which did not conform with recommendations, especially Brinerdin (a fixed dose combination of clopamide, reserpine and dihydroergocristine) while advice on non-pharmacological measures was neglected. We intervened by reporting our findings directly to the Ministry of Health in Trinidad and Tobago

2 456 and we made presentations at national and regional meetings of scientific and professional bodies. The survey findings were also reported to medical officers in government primary care services in the setting of annual training workshops at which draft guidelines were discussed. This process led to the drafting of new management guidelines at a workshop held in Port of Spain in 1996 and these were formally published early in Our aim in the present study was to determine whether intervention by means of audit, linked with training, resulted in changes in the management of high blood pressure in government primary care health centres in Trinidad and Tobago between 1994 and Subjects and methods Study design and selection of subjects The design was a single group before and after study. Studies with follow-up may be carried out using either cross-sectional or cohort sampling. Cross-sectional sampling was preferred in the present study, both on grounds of administrative convenience and because this approach was more likely to provide a representative sample of clinic attenders at each measurement occasion. 16,17 We report data obtained from 16 health centres which were surveyed on both measurement occasions. The health centres were nominated by County Medical Officers of Health and were located in each of the administrative counties of Trinidad and Tobago. Within each health centre we carried out a census of attenders over a defined time period at the clinic sessions designated for adult attenders. We listed all adult attenders who were aged 30 years and who were neither diabetic nor pregnant. In 1994 the census period was for 4 weeks while in 1998 it was for 3 weeks. Data collection We abstracted data from the subjects clinical records excluding data from the index visit. The data abstracted included: the date of the first clinic visit; records of weight, height, urinalysis, and records of blood glucose, urea and creatinine. We also noted the first and the most recent blood pressure recordings, details of current drug treatment, and whether advice on diet or exercise had ever been. In 1994 data were abstracted by MG and a research nurse, while in 1998 the data were abstracted by staff from the Nutrition Division of the Ministry of Health. Analysis We analysed data using two definitions of hypertension. The WHO definition of hypertension 18 is of blood pressures 160/95 mm Hg, while the US Joint National Committee 19 gives a definition of 140/90 mm Hg. Subjects were also considered to be hypertensive if they were taking antihypertensive drugs including thiazide diuretics, beta-blockers, calcium channel antagonists, angiotensin-converting enzyme inhibitors, methyldopa, reserpine, vasodilators, alpha-blockers, adrenergic neuron inhibiting drugs, Brinerdin or other antihypertensive drugs. In order to test for the effect of the survey year while allowing for correlation of responses within health centres, we used the method of Generalised Estimating Equations (GEEs) implemented in the package Stata. 20 The GEE method provides a computationally intensive approach to the estimation of regression models for clustered data. In the present data, the GEE method was used because individual patient values were clustered within health centres. This approach corrects for the correlation of responses within health centres and leads to more conservative estimates than would be obtained with standard individual level analyses. Where appropriate analyses were adjusted for age (by 10-year age group), sex, and duration of follow-up (by quartile). Results Data were analysed for 981 cases in 1994 and 778 in The retrieval rate for eligible case notes was approximately 92% in 1994 and 91% in Twothirds of the subjects were women with a mean age of approximately 60 years. The duration of followup at the health centre was longer in 1998 than in 1994 but there were no differences in mean systolic or diastolic blood pressure between the 2 years (Table 1). Most patients attending the health centres had blood pressures at some time: 945 (96%) in 1994 and 763 (98%) in The blood pressure Table 1 Case mix and process of care variables. Figures are frequencies (per cent of column total) except where stated 1994 (981) 1998 (778) P value Case-mix Female 677 (69) 533 (69) Age (years) a 61 (14) 60 (14) Duration of follow-up 6.5 (4.1) 8.0 (5.2) (years) a Systolic blood pressure 139 (24) 137 (22) (mm Hg) a Diastolic blood pressure 87 (15) 86 (13) (mm Hg) a Process of care BP ever 945 (96) 763 (98) BP in last (86) 695 (89) Weight in last 336 (34) 400 (51) Height ever 11 (1) 76 (10) Urine glucose in last (23) 229 (29) Urine protein in last (22) 153 (20) Smoking habit ever 82 (8) 59 (8) Alcohol use ever 76 (8) 46 (6) Diet advice ever 150 (15) 223 (29) Exercise advice ever 37 (4) 117 (15) a Figures are mean (s.d.).

3 was within the last 12 for 848 (86%) in 1994 and 695 (89%) in 1998 with no significant difference between the 2 years (Table 1). However, the recording of body weight increased from 336 (34%) in 1994 to 400 (51%) in 1998 and more patients had their height at some time, although the proportion was small 76 (10%) in Use of urinalysis did not change significantly between 1994 and There was a significant increase in the proportion of all subjects with advice on diet or exercise in their notes between 1994 and 1998 (Table 1). Among those subjects who were treated with antihypertensive drugs, there were substantial changes in patterns of drug utilisation between 1994 and 1998 (Table 2). In 1994, 191 (31%) were treated with Brinerdin but by 1998 this proportion had fallen to 29 (6%). In 1998 more use was made of thiazide and potassium sparing diuretic combinations, particularly Dyazide (triamterene with hydrochlorothiazide) increasing from 47 (8%) to 105 (22%). There was also an increase in the use of betablockers from 103 (16%) in 1994 to 132 (27%) in 1998, and ACE inhibitors from 52 (8%) in 1994 to 98 (20%) in There was also a slight increase in the use of calcium channel antagonists from 28 (4%) in 1994 to 39 (8%) in After excluding 51 cases with no blood pressure records available, the proportion of cases with hypertension was estimated according to two definitions (Table 3). The proportion with hypertension was 1176/1708 (69%) according to the WHO definition and 1337/1708 (78%) according to the more stringent definition of 140/90 mm Hg (Table 3). A high proportion of subjects with hypertension were treated with antihypertensive drugs: 1111/1176 (94%) according to WHO definition and 1111/1337 (83%) according to the JNC-VI definition with no differences between years. Among subjects with hypertension according to the WHO definition, 338 (51%) had adequate control (BP 160/95 mm Hg) in 1994 and 297 (58%) in 1998 (Table 3). According to the criterion of 140/90 mm Hg, 110 (15%) of subjects had adequate blood pressure control in 1994 and 76 (13%) in Differences between study years were not significant even after adjusting for age, sex and the duration of follow-up at the clinic Table 2 Drug utilisation in subjects prescribed antihypertensive drugs. Figures are frequencies (per cent of column total). Note categories are not mutually exclusive 1994 (625) 1998 (486) P value Brinerdin 191 (31) 29 (6) Thiazide diuretics 66 (11) 97 (20) Thiazide and potassium 47 (8) 105 (22) sparing diuretic combination Beta-blockers 103 (16) 132 (27) ACE inhibitor 52 (8) 98 (20) Calcium antagonist 28 (4) 39 (8) Methyl dopa 206 (33) 162 (33) Other antihypertensive 21 (3) 7 (1) drugs Note: six cases with no BP records excluded. Table 3 Indicators of blood pressure control, in subjects with blood pressure recordings 1994 (945) 1998 (763) P value a BP 160/95 mm Hg or 662 (70) 514 (67) on drugs on antihypertensive 625 (94) 486 (95) drugs b received diet advice b 139 (21) 185 (36) received exercise 36 (5) 99 (19) advice b BP controlled 338 (51) 297 (58) ( 160/95) b BP 140/90 mm Hg or 744 (79) 593 (78) on drugs on antihypertensive 625 (84) 486 (82) drugs c diet advice c 143 (19) 194 (33) exercise advice 36 (5) 105 (18) c BP controlled 110 (15) 76 (13) ( 140/90) c a Adjusted for age, sex and duration of follow-up. b Per cent of cases with BP 160/95 or on drugs. c Per cent of cases with BP 140/90 or on drugs. (Table 3). The adjusted relative odds of blood pressure being controlled (WHO definition) in 1998 compared with 1994 were 1.39 (95% confidence interval 0.96 to 2.00). Discussion Background There is little information available concerning the management of hypertension in middle- and lowincome countries in spite of an increasing awareness of the importance of cardiovascular diseases in association with epidemiological transition. 21 Some health policy analysts have argued that hypertension management should receive a low priority for investment in middle and low income countries 22 but few studies have provided objective data concerning the cost-effectiveness of hypertension treatment in these settings. 23 Many studies have considered variations in hypertension prevalence, 24 but few studies have examined the problems of health care delivery. Selection of appropriate drugs is of great importance because of wide variations in cost. For example, in Mexico in year s supply of enalapril cost 40 times as much as 1 year s supply of hydrochlorothiazide. 25 Quality of care issues are also particularly important in middle and low income countries because trained staff may be few in number and difficult to retain, supplies of pharmaceuticals may be limited, and cultural differences between patients and health care staff may be substantial. Poor quality care tends to reduce the effectiveness of intervention without reducing costs. Relevance of main findings Our data may provide relevant information to those concerned with hypertension control in middle- and low-income countries. First, this report shows that 457

4 458 periodic national surveys of the primary care management of hypertension can be carried out quite efficiently in a small, middle-income country. Secondly, the results document some of the current achievements in the government primary care setting, including the high proportion of attenders in whom the blood pressure is measured, and the high proportion of patients requiring treatment who receive it. Thirdly, our data show that an audit used to inform policy and practice, linked with educational activities, may bring about large changes in drug utilisation and some changes in processes of care over a relatively short time period. Finally, the results document some of the problems that remain, in particular the low rates of control in those treated for hypertension. The changes in drug utilisation which we observed were likely to be the result of changes in drug procurement which is managed at national level. The use of Brinerdin was almost eliminated during this time and there was an increase in the use of diuretics and beta-blockers. However, the increased use of ACE inhibitors and calcium antagonists may give cause for concern because of the high cost of these agents. We also documented changes in processes of care with an increase in the recording of weight and height, and greater documentation of advice on diet and exercise. Rates of hypertension control were relatively low. In their population-based study in Barbados, St Lucia and Jamaica, Freeman et al 26 found that control of blood pressure (using the criterion 160/95 mm Hg) was achieved in 58% in Barbados, 55% in Jamaica and 33% in St Lucia. Similar results have been reported from high-income countries. 19 However, because of the high prevalence of hypertension it is clear that government primary care services will be contributing to a reduction in the population attributable risk by detecting the condition and achieving control in significant proportion. It is not surprising that improved rates of control were not detected in this study. Even randomised controlled trials show small changes in diastolic pressure of approximately 5 mm Hg. In our clinic settings blood pressures are often measured to the nearest 10 mm Hg, making detection of small changes unlikely. Limitations of study Our study included data from a large number of cases, obtained from a large number of health centres. We used standard methods to collect data but we only collected data that were in case notes. Thus we studied the care that was and not the care that was delivered. In a single group before and after study, the effect of intervention will be confounded by the effect of history and a range of external influences may have contributed to the changes which we observed. 27 These include, for example, health sector reforms, pharmaceutical company promotions or changes in staffing. In a pilot study we found that it was difficult to investigate the quality of diagnosis in hypertension. In Trinidad and Tobago, there are both public and private primary care providers and doctor shopping is common. As a result the current place of treatment is often different from the one where a diagnosis of hypertension was made. Conclusion Our data show that an audit study used to inform health care policy, staff training and clinical practice may contribute to improving standards of hypertension management in a middle-income country like Trinidad and Tobago. In common with the experience of high-income countries, 28,29 the changes in quality of care which we documented were of modest size and the data illustrate some of the problems which remain. Acknowledgements The authors thank the Chief Medical Officer of Trinidad and Tobago for permission to report this work. We also thank the staff of the Nutrition Division, the PMO(CS), the staff of the community services and the Commonwealth Caribbean Medical Research Council for assistance. The 1994 survey was supported by a grant from the UK Overseas Development Administration. References 1 Miller GJ et al. Adult male all cause, cardiovascular and cerebrovascular mortality in relation to ethnic group, systolic blood pressure and blood glucose concentration in Trinidad, West Indies. Int J Epidemiol 1988; 17: INTERSALT Cooperative study group. Appendix tables. Centre-specific results by age and sex. J Hum Hypertens 1989; 3: Pan American Health Organisation. Health conditions in the Americas. Volume edition. Washington DC: PAHO/WHO, Gulliford MC. Epidemiological transition in Trinidad and Tobago, West Indies 1953 to Int J Epidemiol 1996; 25: D Mahabir, L Bickram, MC Gulliford. Stroke in Trinidad and Tobago: burden of illness and risk factors. Pan Am J Public Health 1998; 4: Collins R et al. Blood pressure stroke and coronary artery disease. Part 2, short term reductions in blood pressure: overview of randomised drug trials in their epidemiological context. Lancet 1990; 335: Strasser T, Wilhelmsen L. Assessing hypertension control and management. WHO Regional Publications, European Series number 47. World Health Organisation: Geneva, World Development Report. Investing in Health. Oxford University Press: Oxford, Republic of Trinidad and Tobago. Office of the Prime Minister. Central Statistical Office. Annual Statistical Digest 1994/5. Number 41. Central Statistical Office: Port of Spain, Murray CJL, Lopez AD. Global comparative assessments in the health sector. Disease burden, expenditures and intervention packages. World Health Organisation: Geneva, Mahabir D, Gulliford MC. Medical practitioners views on the management of hypertension in Trinidad and Tobago. West Indian Med J 1997; 46: Alert CV, Fraser HS, Taylor J. Quality of blood press-

5 ure monitoring and control in primary care in Barbados. West Indian Med J 1996; 45 (Suppl 2): 24 (abstract). 13 Mahabir D, Bickram L, Picou D, Gulliford MC. Quality of blood pressure monitoring and control in primary care in Trinidad and Tobago. West Indian Med J 1995; 44 (Suppl 2): 15 (abstract). 14 R Wilks et al. Quality of care of hypertension in three clinical settings in Jamaica. West Indian Med J 1998; 47 (Suppl 2): 18 (abstract). 15 Commonwealth Caribbean Medical Research Council. Managing hypertension in primary care in the Caribbean. CCMRC: Port of Spain, Diehr P et al. Optimal survey design for community intervention evaluations: cohort or cross-sectional? J Clin Epidemiol 1995; 48: Feldman HA, McKinlay SM. Cohort versus cross-sectional design in large field trials: precision, sample size and a unifying model. Statist Med 1994; 13: World Health Organisation. Hypertension control. Report of a WHO Expert Committee. Number 862. WHO: Geneva, The Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. The Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 1997; 157: Stata Statistical Software. Release 5. Volume 3. College Station, Stata Press: Texas, 1997, pp Beevers DG, Prince JS. Some recent advances in noncommunicable diseases in the tropics 1. Hypertension an emerging problem in tropical countries. Trans Roy Soc Trop Med Hygiene 1991; 85: Pearson TA, Jamison DT, Trejo-Gutierrez J. Cardiovascular disease. In: Jamison DT, Mosley WH, Measham AR, Bobadilla JL (eds). Disease control priorities in Developing Countries. Oxford University Press: Oxford, Cooper RS et al. Hypertension treatment and control in sub-saharan Africa: the epidemiological basis for policy. BMJ 1998; 316: Cooper RS et al. The prevalence of hypertension in seven populations of West African origin. Am J Public Health 1997; 87: Calvo-Vargas CG et al. Changes in the costs of antihypertensive medications in a developing country: a study in Mexico comparing 1990 and Am J Hypertens 1998; 11: Freeman V et al. A comparative study of hypertension prevalence, awareness, treatment and control rates in St Lucia, Jamaica and Barbados. J Hypertens 1996; 14: Cook TD, Campbell DT. Quasi-experimentation. Design and analysis issues for field settings. Rand McNally College Publishing Company: Chicago, Grimshaw JM. Russell IT. Effect of clinical guidelines on medical practice: a systematic review of rigorous evaluations. Lancet 1993; 342: Oxman AD, Thomson MA, Davis DA, Haynes RB. No magic bullets: a systematic review of 102 trials of interventions to improve professional practice. CMAJ 1995; 153: