A decade of caring for drug users.

Sir, O’Reilly et al (October Journal)1 report a much-needed description of the epidemiology of patients who have been removed from a general practitioner’s (GP’s) list. The authors do not, however, consider ethnicity of patients as a possible variable. It could be hypothesized that patients of South Asian origin would be more likely to be removed — judging by evidence from a postal survey of GPs in a northern industrial town with a large Asian population, which suggested more unfavourable attitudes towards Asian patients compared with non-Asians.2 Asians were thought to require longer consultations, to be less compliant, and make excessive and inappropriate use of health services. Clearly, if Asian patients were to be removed more frequently than non-Asians, it would have serious implications for equity. We report here, a preliminary analysis of data on patient removals collected by Leicestershire Health Authority over the six-month period from 1 April 1998 to 30 September 1998, using the same definitions of removed patients as O’Reilly et al.1 Ethnicity was defined on the basis of surname/forename analysis.3 One-third of Leicestershire’s population of approximately 900 000 live in the city of Leicester, and 23.7% of the city’s population can be classified as South Asian on the basis of the 1991 Census (22.3% Indian, 1% Pakistani, and 0.4% Bangladeshi).4 Over the period studied, there were 408 removal events, 403 removed patients, and 207 removal decisions made by GPs. Over one-fifth 106/500 (21%) of Leicestershire GPs made a removal decision. On average, those GPs who removed patients made two removal decisions (range 1–18). Of the removal decisions, 145/207 (70%) were individual removals, 60/207 (29%) were household removals (more than one individual at the same address), and 2/207 (1%) were nursing home removals. The most frequent age group for removal was 20–39 years (140/396, 35%); 207/403 (51%) of removals were male and 196/403 (49%) female. The majority of patients removed (327/403, 81%) lived in Leicester. The proportion of patients removed in Leicester who were South Asians (90/327) did not differ significantly from the proportion in the population (28% versus 24%; χ2 = 2.11; 1 df; P = 0.15). We conclude that our preliminary analysis does not suggest that Asian patients are more likely to be removed from GPs’ lists than non-Asian patients in Leicester area. What is now required is to move beyond epidemiological description and produce a more detailed understanding of the processes that lead to removal for all patients.