OBJECTIVE To examine general practitioner consultations by demographic and socioeconomic variables and to derive a method of measuring the impact of relative deprivation on general practitioner workload. METHODS The study was based on general practitioner consultations reported in the general household surveys of 1983-7, covering a sample of 129,987 individuals in Great Britain. Odds ratios for general practitioner consultations were obtained for selected variables among children (0-15 years), men (16-64), women (16-64), and elderly people (greater than or equal to 65). These were then used to derive deprivation indices specific to electoral wards for use in general practice. METHODS Great Britain, with particular findings illustrated by English electoral wards and the conurbations of London, Manchester, Merseyside, and the West Midlands. RESULTS Council tenure increased the likelihood of consultation significantly in all four groups. Odds ratios were raised in children, men, and women with no access to a car. Birth in the New Commonwealth or Pakistan yielded high odds ratios in men, women, and elderly people but not in children. Marginally increased consultation rates were evident in the manual socioeconomic groups in women, elderly people, and children with a single parent mother. The deprivation indices for general practice derived using these odds ratios varied substantially among English electoral wards with, for example, anticipated general practitioner consultations in the electoral ward of Hulme, Manchester, being 24% higher than the average ward in England as a result of local attributes, and consultations in the Cheam South ward of Sutton, London, 11% lower than average. CONCLUSIONS This deprivation index for general practice overcomes several shortcomings expressed about the underprivileged area score, which has been adopted in the 1990 contract as a basis for allocating deprivation supplements to general practitioners. The proposed index can be applied nationwide.