16 Naish J, Sturdy P, Bobby J, Pereira P. The association between Asian ethnicity and ... 17 McColl A, Roderick P, Gabbay J, Smith H, Moore M. Performance.
Family Practice © Oxford University Press 2001
Vol. 18, No. 2 Printed in Great Britain
Prescribing rates for psychotropic medication amongst east London general practices: low rates where Asian populations are greatest Sally A Hull, Jo Cornwell, Charlotte Harvey, Sandra Eldridge and Portia Omo Barea Hull SA, Cornwell J, Harvey C, Eldridge S and Omo Bare P. Prescribing rates for psychotropic medication amongst east London general practices: low rates where Asian populations are greatest. Family Practice 2001; 18: 167–173. Objectives. The aim of this study was to examine the contribution of Asian ethnicity to the variation in rates of practice prescribing for antidepressant and anxiolytic medication, taking into account other population and practice organizational factors. Methods. A practice-based cross-sectional survey was carried out of the prescribing of antidepressants and anxiolytics (daily defined dosages) in 164 general practices. The study was set in East London and the City Health Authority, which includes the multiethnic inner London boroughs of Hackney, Tower Hamlets, Newham and the City of London. The main outcome measures were the annual prescribing rates for each group of drugs, calculated as the total annual daily defined dosages divided by the practice population, and the ratio of antidepressant/ anxiolytic annual prescribing rates. Results. Prescribing rates for antidepressants showed a 25-fold variation between practices; this was greater for anxiolytics. The median annual prescribing rate for all antidepressants combined was 4.13 (interquartile range 2.50–5.88). For all anxiolytics and hypnotics combined the median annual prescribing rate was 3.55 (interquartile range 1.71–6.36). Univariate analysis showed that Asian ethnicity alone accounted for 28% of the variation in antidepressant prescribing and 20.5% of the variation in the anxiolytic prescribing. A backwards multiple regression model using 10 explanatory practice and population variables accounted for 47.7% of the variance in antidepressant prescribing and 34% of the variance in the anxiolytic prescribing. Conclusion. In practices where the proportion of Asian patients is high, both antidepressant and anxiolytic prescribing is low. This is important for understanding interpractice prescribing variation and for setting levels of drug budgets. This study confirms that the low rates of non-psychotic disorders presented by Asian populations is not a selective feature of access to secondary care, but is evident in the prescribing behaviour of GPs. Uncertainty remains as to how much this is due to a lower prevalence rate, ‘culture-bound syndromes’ or practical difficulties in diagnosis and management within the general practice setting. Keywords. Asian ethnicity, depression, general practice.
there is less certainty within the field of mental health. Debate continues on whether the apparent lower rates of depression and anxiety in Asian groups are due to a genuine difference in prevalence, or whether presentations may be masked by differences in symptomatic experiences leading to ‘culture-bound syndromes’ which are not well captured by western psychiatric concepts and instruments.1,2 Most of the studies which have suggested lower rates of mental disorders amongst Asian populations are based on hospital populations.3 The few studies based in general practice have shown
Introduction Whilst a considerable amount is known about ethnic differences in the incidence of major physical disorders,
Received 3 March 2000; Revised 23 August 2000; Accepted 30 October 2000. Department of General Practice and Primary Care, Queen Mary and Westfield College, Medical Sciences, Mile End Road, London E1 4NS and aEast London and the City Health Authority, London, UK.
167
168
Family Practice—an international journal
that Asian groups consult their GPs more frequently than the white population except for mental disorders, where there are low consultation rates particularly amongst women.4 A practice-based study in east London demonstrated that Asian (predominantly Bangladeshi) patients with depression presented more frequently than white patients with somatic symptoms, and were treated for a shorter period of time with lower doses of antidepressants.5 Depression is a common condition with an overall adult annual prevalence of 5%.6 The majority of identified cases are managed within general practice, and antidepressant prescribing plays a major part in treatment. Recent guidelines have provided advice on the goals of antidepressant prescribing including advice on dose and length of treatment.7 There is evidence that antidepressant prescribing in general practice rarely meets these criteria.8,9 In areas where there are large ethnic minority populations, GPs may face additional practical challenges in recognizing and managing depression. These difficulties may include language and culture barriers, access to and expectations of treatment, and the time constraints of current practice. A previous study in Cambridge linked higher levels of psychotropic prescribing to practice rates of women over 65 and temporary patients.10 There is also evidence for an association between permanent sickness rates and antidepressant prescribing.11 This study set out to examine the relationship between practice prescribing of antidepressants and anxiolytics, calculated using the daily defined dose system, and the proportion of patients with Asian names on practice lists.
Methods Prescribing, population and practice data were obtained for all 164 general practices in East London and the City Health Authority (ELCHA). This includes inner city boroughs with high levels of social deprivation, and an ethnically diverse population, the largest grouping being of South Asian origin.
Prescribing data The World Health Organization-defined daily dose system of medication was used for this study.12 This enables the total volume of prescriptions for each medication to be converted into a standardized daily drug dose. Aggregation of all the defined daily doses for a group of related drugs can provide a measure of the notional days of treatment for a therapeutic drug group. Data from the Prescription Pricing Authority (PPA) for all 164 practices within the ELCHA was obtained for the period January–December 1996. For each drug in the British National Formulary Chapters 4.1 (hypnotics
and anxiolytics) and 4.3 (antidepressant drugs), the total number of defined daily doses prescribed during the year was calculated for each practice.13,14 The defined daily doses were then aggregated into two broad classes; anxiolytics with hypnotics (excluding barbiturates) and antidepressants (including tricyclics, selective serotonin re-uptake inhibitors and monoamine oxidase inhibitors). For each practice, an annual prescribing rate for each of these two groups of drugs was calculated by combining the total annual daily defined dosages for each group, and dividing by the practice population.
Practice data and population characteristics The East London General Practice Database was used as a framework for this project. This holds information on all practices in the district, collated from the health authority, the 1991 census and the PPA. A selection of variables available on the GP database were chosen for inclusion in the analysis. These included factors reflecting practice resources such as the number of doctors, list size and staffing details. Three factors which reflect the organizational efficiency of the practice were included; training status, the proportion of eligible women on the list who have had cervical cytology and the asthma bronchodilator/preventor items prescribing ratio. Population characteristics included in the analysis were the Jarman underprivileged areas score for each practice, the borough locality (Newham, Hackney or Tower Hamlets) the percentage of children under 4 years of age on the practice list, the percentage of those over 65 years of age and the proportion of the practice list with Asian names. This variable was devised using the names of patients on each practice list, and assigning each name to an ethnic group. This method has been shown to be accurate for the attribution of South Asian ethnicity, but underestimates the attribution of Afro-Caribbean ethnicity.15,16
Analysis The unit of analysis was the general practice. One singlehanded practice which was an extreme outlier for all the prescribing outcome variables was excluded from the analysis. The prescribing outcome variables had nonnormal distributions, requiring transformation prior to statistical analysis. Categorical variables were analysed using the first category as the baseline against which the other categories were compared. Univariate analysis between practice and population characteristics and the main prescribing outcome variables were examined using descriptive statistics and simple linear regression models. Multivariate models were fitted using backwards stepwise regression models with entry and exit criteria of P = 0.049 and P = 0.05, respectively. All statistical analysis was undertaken using STATA.
169
Prescribing rates for psychotropic medication amongst east London general practices
where there are low proportions of Asians in the practice population. Higher rates of prescribing were also associated with greater numbers of patients over 65 years of age, with smaller practices and with non-training status (see Tables 3 and 4). The ratio of antidepressant/anxiolytic annual prescribing rates has been proposed as a marker of appropriate prescribing for mental health conditions comparable with the asthma prophylaxis to bronchodilator items ratio for respiratory prescribing.17 In this analysis, 35.8% of the variation between practice ratios could be accounted for by the asthma prescribing ratio, the size of partnership, the percentage of Asians on the practice list, the percentage of those over 65 years and the practice locality.
Results Table 1 shows the variation in prescribing of antidepressants and anxiolytics/hypnotics. Rates of prescribing showed considerable variation between practices, with a greater range in anxiolytic prescribing than for antidepressants. For all antidepressants combined, the median defined daily dose per practice population was 4.13 (range 0.5–12.4). For all anxiolytics and hypnotics combined, the median defined daily dose per practice population was 3.55 (range 0.02–25.0). Multivariate analysis The explanatory variables used in the multivariate analysis, and their distribution between practices, are shown in Table 2. The proportion of Asian names on practice lists varied from 1.5 to 98% of the practice population. The multivariate analysis was undertaken on the 138 practices for which a full data set was available (see Table 4). For antidepressant prescribing, 47.7% of the variation between practices can be explained by the proportion of patients with Asian names on the practice list, the size of the partnership, the presence of a practice manager and the practice locality. Antidepressant prescribing is significantly higher in practices with low proportions of Asians on the list, in large practices and where there is a practice manager. The relative contribution of these factors to the explanation of variance can be assessed from Table 3, which shows the results of univariate analysis. Here the association between higher antidepressant prescribing and low levels of Asians in the practice population alone can be seen to account for 27.9% of the variance between practices, with practice size accounting for 17.9% of the variance. Multivariate analysis also showed that anxiolytic prescribing is higher TABLE 1
Discussion Prescribing patterns The range of psychotropic prescribing between practices was large. For antidepressants, there was 25-fold variation between the highest and lowest prescribers; for anxiolytics this range was even greater. However, the median annual prescribing rate of antidepressant prescribing at 4.13 is close to the figure quoted by Pharoah (4651/1000 population) for Cambridge practices in 1995.10 The median annual prescribing rate for anxiolytics and hypnotics combined was 3.55; this is about half the figure quoted by Pharoah (hypnotics 6043/1000 population, anxiolytics 1714/1000 population). These findings do not support the view that inner city practitioners are prescribing anxiolytics at higher rates than their rural colleagues, as has been suggested previously.18 Larger partnerships were associated with higher rates of antidepressant
The range of antidepressant and anxiolytic prescribing: annual prescribing rates per practice population in 163 practices in east London
Variable
Median
Minimum
Maximum
Interquartile range
No. of practices
Tricyclic antidepressants: annual prescribing ratea
1.79
0.25
7.64
1.2–2.7
163
Selective serotonin re-uptake inhibitors: annual prescribing rate
2.1
0.24
7.8
1.38–3.24
163
Monoamine oxidase inhibitors: annual prescribing rate
0.004
0
1.4
0–0.07
163
All antidepressants: annual prescribing ratea
4.13
0.50
12.4
2.50–5.88
163
Hypnotics: annual prescribing rate
2.37
0.006
15.4
1.17–4.3
163
Anxiolytics: annual prescribing rate
1.0
0.02
18.5
0.49–2.05
163
All anxiolytic and hypnotics: annual prescribing ratea
3.55
0.02
25.0
1.71–6.3
163
Ratio of antidepressant/anxiolytic and hypnotics: annual prescribing rate
1.22
0.11
23.18
0.67–2.08
163
aSum
of annual daily defined dosages for that group of drugs/practice population.
170
Family Practice—an international journal TABLE 2
Distribution of explanatory variables used in the univariate and multiple regression analysis
Variable
No. of practices
Median
Minimum
Maximum
Interquartile range
Continuous variables List size/full-time equivalent GP
163
2121
593.5
Percentage of practice list 0–4 years
163
8
3
16
6–9
Percentage of practice list .65 years
163
10.5
2.0
27.6
2.9–20.1
Percentage of eligible women on list receiving cervical cytology
163
75.7
30.1
99.2
63.5–85.5
Asthma prophylaxis to bronchodilator prescribing ratio
161
0.4
Underprivileged area score
154
41.4
17.5
59.7
38.2–45.1
Percentage of practice population with Asian names
146
13.8
1.5
98.0
6.4–30.1
0.12
6010
1713–2612
0.79
0.39–0.55
Categorical variables Practice locality
Hackney Newham Tower Hamlets
56 65 42
Partnership size
1 2 or 3 4 or more
71 58 35
Training status
Yes No
19 144
Practice manager
Yes No
100 58
TABLE 3
Univariate regression analysis for total antidepressant and anxiolytic prescribing and the antidepressant/anxyolitic ratio, by general practice and population characteristics No. of practices
Adjusted R2 (%)
Constant
B coefficient (95% CI)
P-value
Annual prescribing rate for all antidepressants combineda Of practice population with Asian names Partnership size 4 or more Presence of a practice manager Tower Hamlets locality
146 163 158 163
27.9 17.9 5.5 3.7
2.355 1.94 1.87 1.99
–0.013 (–0.02 to –0.009) 0.62 (0.414 to 0.83) 0.295 (0.11 to 0.48) 0.25 (0.05 to 0.45)
,0.0001 ,0.0001 0.002 0.01
Annual prescribing rate for anxyolitics and hypnoticsa % of practice population with Asian names % of practice list .65 years Proportion of women on list receiving cytology screening Partnership size 4 or more Training status
146 163 163 163 163
20.5 17.9 4.8 0.2 1.6
2.40 1.09 3.00 2.01 2.02
–0.11 (–0.020 to –0.011) 8.13 (5.47 to 10.79) –0.014 (–0.023 to –0.004) –0.204 (–0.55 to 0.14) –0.40 (–0.82 to –0.02)
,0.0001 ,0.0001 0.003 0.2 0.059
Ratio of antidepressant/anxiolytic annual prescribing rates % of practice population with Asian names Proportion of children under 4 years on practice list Partnership size of 4 or more Partnership size of 2 or 3 Asthma prophylaxis to bronchodilator prescribing ratio Tower Hamlets locality Newham locality
146 163 163 163 161 163 163
6.4 6.0 7.8 0.3 9.9 2.2 1.2
–0.0355 –0.47 0.10 0.26 –0.87 0.15 0.13
0.008 (0.003 to 0.01) 8.87 (3.65 to 14.09) 0.63 (0.31 to 0.96) –0.095 (–0.37 to 0.18) 2.29 (1.24 to 3.35) 0.29 (–0.009 to 0.59) 0.24 (–0.03 to 0.51)
0.001 0.001 0.0002 0.49 ,0.0001 0.06 0.08
Variable
aTotal
annual daily defined dosages for that group of drugs/practice population.
171
Prescribing rates for psychotropic medication amongst east London general practices TABLE 4
Multivariate associations between prescribing of antidepressants, and anxiolytics, and the ratio of antidepressant/anxiolytics and practice characteristics in 138 practices in east London
Model Annual prescribing rate for all antidepressants % Asians on practice list Partnership size 4 or more Presence of a practice manager Tower Hamlets locality
Adjusted B coefficient (95% CI)
P-value
–0.011 (–0.013 to –0.008) 0.356 (0.168 to 0.544) 0.155 (0.004 to 0.307) 0.357 (0.197 to 0.517)
,0.0001 ,0.0001 0.004 ,0.0001
–0.02 (–0.021 to –0.01) 4.04 (1.15 to 6.94) 0.35 (–0.666 to –0.03) –0.48 (–0.88 to –0.09)
,0.0001 0.007 0.03 0.02
0.01 (0.001 to 0.013) –4.51 (–7.28 to –1.74) 2.02 (0.92 to 3.11) 0.70 (0.36 to 1.03) 0.26 (0.01 to 0.52) 0.43 (0.13 to 0.72) 0.41 (0.08 to 0.74)
0.016 0.002 ,0.0001 ,0.0001 0.04 0.005 0.016
Adjusted R2 = 47.7% constant = 2.069 F = 32.23 n ,0.0001 Annual prescribing rates for all anxiolytics and hypnotics % Asians on practice list % of practice list .65 years Partnership size 4 or more Training status Adjusted R2 = 3.5.3% constant = 3.17 F = 19.67 n ,0.0001 Ratio of antidepressant/anxiolytic prescribing % Asians on practice list % of practice list .65 years Asthma P/B prescribing ratio Partnership size 4 or more Partnership size 2 or 3 Newham locality Tower Hamlets locality Adjusted R2 = 37.5% constant = –0.949 F = 12.74 P , 0.0001
prescribing, lower rates of anxiolytics and a higher antidepressant/anxiolytic ratio. Within east London, where there are above average rates of single-handed practices, larger practices have a greater share of staffing support, smaller list sizes and include the majority of training practices, all factors which may influence prescribing habits. Previous studies have shown that overall practice prescribing rates are higher for practices with a greater proportion of Asian patients on the list,16,19 which is consistent with findings of higher consultation rates4 and a greater prevalence of chronic disease such as diabetes and ischaemic heart disease.20,21 This study, in showing low rates of psychotropic prescribing amongst Asian populations, demonstrates the importance of analysing drug groups separately, and contributes to the knowledge base required to set practice budgets and understand interpractice variation. Study limitations Cross-sectional studies such as this, based on routinely available information, are prone to a number of limitations. The data may be of variable quality and introduce a source of bias. We chose not to use the 1991 census variables for Asian ethnicity because the information has become outdated. In addition, attributing census-based variables to practice populations depends on the assumption that there is no selection bias of general practices by ward populations.22 In the case of Asian populations within east London, this assumption
is likely to be false; many Asian patients will seek out an Asian doctor in order to have a common culture and language. Our finding of practices with up to 98% Asian names supports this view. This in turn might be thought to improve the recognition of mental health problems, which would tend to decrease the differences in practice variation which we have found. This does not appear to be translated into overall prescribing differences, Gill and others found no systematic difference in total prescribing costs, frequency or generic prescribing rates between Asian-trained and British-trained Asian and non-Asian doctors.23 The disadvantage of using the naming system to allocate ethnicity is that groups other than Asians were not captured for this study; it is possible that the sizeable Afro-Caribbean population may also be an influence on prescribing rates. The choice of explanatory variables is also open to debate. Age banding of the practice populations was not available to us; as a proxy, we used the proportion of children under 4, and adults over 65 years old on the list, recognizing that the age distribution of the practice population is an important predictor for prescribing costs.10 We did not include details of counsellors or psychologists attached to practices, but note that previous studies have not found this to affect prescribing rates or costs.24 Nevertheless, in spite of these potential limitations, we were able to account for nearly half the variation in antidepressant prescribing and over a third of anxiolytic prescribing
172
Family Practice—an international journal
between practices. The major factors contributing to this variation was the size of the Asian population and the practice partnership size. Asian ethnicity and mental health problems This study confirms that the low rate of non-psychotic disorders presented by Asian populations is not a selective feature of access to secondary care, but is evident in the prescribing behaviour of GPs. This is the case even where there may be selective registration with same culture doctors. The finding of lower rates of anxiolytic prescribing and a higher ratio of antidepressant/ anxiolytic medication does not suggest that substitution of anxiolytics for antidepressants is occurring in these populations. Findings from the fourth national survey of ethnic minorities help to illuminate these findings.2 In this community survey, rates of both psychotic and nonpsychotic mental illness among Asians were lower than for white groups, but there was some evidence that the screening instruments performed differently for Asians, suggesting that morbidity may not have been well recognized. Rates of depression and psychosis amongst south Asians educated in Britain approximated to rates for their white peers. Suicide rates are also low in south Asian populations, apart from young women, who have a rate 2–3 times the national average.25 Where does this leave practitioners within primary care? If existing screening instruments and related clinical enquiries do not pick up depression in this group, it is hardly surprising that even the best motivated practitioners have difficulties in doing so. Undoubtedly, further untangling of the symptomatic expression of these disorders in south Asians is required, both through community surveys and through understanding the meanings and clinical significance of symptoms expressed to doctors. Meanwhile, practical investments in the primary care setting need to concentrate on an expansion of access to health advocates in order to minimize the language and culture gaps, and an emphasis on training in depression awareness and transcultural sensitivity to consolidate the existing skills of practitioners.
• The prevalence of common mental disorders may be lower amongst Asian populations, but underpresentation and difficulties with diagnosis will contribute to the low rates of prescribing.
References 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Key messages • There was a 25-fold variation in annual prescribing rates for antidepressant medication between practices in east London. Variation was greater for anxiolytics. • In a multiple regression analysis, 47.7% of the variation between practices in the prescribing of antidepressants could be accounted for by the percentage of Asians on the practice list, a partnership size of four or more, the presence of a practice manager and Tower Hamlets locality. • Prescribing of both antidepressants and anxiolytics is lower amongst general practices which have high numbers of Asians in their practice population.
18
19
20
21
22
23
Ineichen B. The mental health of Asians in Britain. Br Med J 1990; 300: 1669–1670. Nazroo JY. Ethnicity and Mental Health: Findings from a National Community Survey. London: Policy Studies Institute, 1999. Stern G, Cottrell D, Holmes J. Patterns of attendance at child psychiatry outpatients with special reference to Asian families. Br J Psychiatry 1990; 156: 384–387. Gillam SJ, Jarman B, White P, Law R. Ethnic differences in consultation rates in urban general practice. Br Med J 1989; 299: 953–957. Cornwell J, Hull S. Do general practitioners prescribe antidepressants differently for South Asian patients? Fam Pract 1998; 15: S16–S18. Blacker CVR et al. Depressive disorders in primary care. Br J Psychiatry 1987; 150: 737–751. Paykel ES, Priest RG. Recognition and management of depression in general practice: consensus statement. Br Med J 1992; 305: 1198–1202. Johnson D. A study of the use of antidepressant medication in general practice. Br J Psychiatry 1974; 125: 186–192. Catalan L et al. General practice patients on long term psychotropic drugs: a controlled investigation, Br J Psychiatry 1988; 152: 399–406. Pharoah P, Melzer D. Variation in prescribing of hypnotics, anxiolytics and antidepressants between 61 general practices. Br J Gen Pract 1995; 45: 595–599. Mackenzie I, Buckingham K, Wankowski J, Wilcox M. Morbidity deprivation and antidepressant prescribing in general practice. Br J Gen Pract 1999; 49: 884–886. World Health Organisation Collaborating Centre for Drug Statistics Methodology. Guidelines for Defined Daily Doses. Oslo, Norway: World Health Organisation, 1991. Maxwell M, Heaney D, Howie JGR, Noble S, General practice fundholding: observations on prescribing patterns and costs using the daily defined dose method. Br Med J 1993; 307: 1190–1194. British Medical Association, Royal Pharmaceutical Society of Great Britain. British National Formulary. London: BMA and the Pharmaceutical Press, 1997. Nicoll A, Bassett K, Ulijaszek SJ. What’s in a name? J. Epidemiol Community Health 1986; 40: 364–368. Naish J, Sturdy P, Bobby J, Pereira P. The association between Asian ethnicity and prescribing rates in east London general practices: a database study. Health Informatics J 1998; 100–105. McColl A, Roderick P, Gabbay J, Smith H, Moore M. Performance indicators for primary care groups: an evidence based approach. Br Med J 1998; 317: 1354–1360. Holm M, Olesen F. Factors affecting prescription of psychotropic drugs in general practice. Scand J Prim Health Care 1988; 6: 169–173. Gill P, Scrivener G, Lloyd D, Dowell T. The effect of patient ethnicity on prescribing rates. Health Trends 1995; 27: 111–113. McKeigue PM, Shah B, Marmot MG. Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancet 1991; 337: 382–386. Balarajan R. Ethnicity and variations in the nation’s health. Health Trends 1995; 4: 114–119. Scrivener G, Lloyd D. Allocating census data to general practice populations: implications for study of prescribing variation at practice level. Br Med J 1995; 311: 163–165. Gill PS, Dowell A, Harris C. Effect of doctors’ ethnicity and country of qualification on prescribing patterns in single handed general practices. Br Med J 1997; 315: 1590–1594.
Prescribing rates for psychotropic medication amongst east London general practices 24
Sibbald B, Addington-Hall J, Brenneman D, Freeling P. Investigation of whether on site general practice counsellors have an impact on psychotropic drug prescribing rates and costs. Br J Gen Pract 1996; 46: 63–67.
25
173
Soni-Raleigh V. Suicide patterns and trends in people of Indian subcontinent and Caribbean origin in England and Wales. Ethnicity and Health 1996; 1.
