HORMONE REPLACEMENT THERAPY - CHARACTERISTICS OF USERS AND NONUSERS IN A BRITISH GENERAL-PRACTICE COHORT IDENTIFIED THROUGH COMPUTERIZED PRESCRIBING RECORDS

Study objective - To assess the feasibility of recruiting a cohort of women, including long term users of postmenopausal hormone replacement therapy (HRT), through computerised general practice prescribing reco rds, and to compare clinical and demographic characteristics of users and non-user controls. Design - Cross sectional analysis of questionna ire data. Setting - Subjects were recruited through 17 general practic es in the Oxfordshire, south west Thames, and north west Thames region s that contributed to the VAMP Research Database. Participants - A tot al of 2964 women aged 45-64 years were identified. Altogether 1482 wer e long term (>1 year) users of HRT and 1482 were non-user controls: 10 37 (70%) of the users and 819 (55.3%) of the controls agreed to partic ipate and provided questionnaire data. Main results - Users of HRT wer e more likely to have undergone hysterectomy than controls. Most women with a history of hysterectomy used unopposed oestrogen, while those with intact uteri generally used a combination of oestrogen and a prog estagen. Among women who had undergone hysterectomy, HRT users did not differ significantly from controls over a range of demographic and cl inical characteristics but they were more likely to be past users of o ral contraceptives. Among women with intact uteri, users were similar to controls in terms of reported clinical characteristics, but were of higher social class and were more likely to be past users of oral con traceptives and to have had a mammogram after the age of 50. Compared with the general population, all categories of women recruited to the study were of higher social class and exhibited more health conscious behaviours. Conclusions - Electronic general practice prescribing reco rds provide a feasible and efficient method for recruiting women to a cohort of HRT. Women who agreed to participate in this study were not representative of the general population, emphasising the importance o f internal controls in such a study. Among participants, HRT users who had not undergone hysterectomy showed evidence of better health than non-users on some dimensions. In the whole sample, however, there were no appreciable differences in social class and self reported health i ndicators between users and controls.