Twenty year follow-up of a cohort based in general practices in 24 British towns

Background A national prospective study of cardiovascular disease (CVD) was set up in 1978 to explain the reasons for the marked geographical variation in CVD rates in Great Britain. A total of 7735 males, aged 40-59 years (born between 1919 and 1939) randomly selected from one general practice age-sex register in each of 24 towns, responded to a screening invitation from their general practitioner (GP) and were examined in 1978-1980. We describe the methods used and the contact maintained after following a cohort for 20 years. Methods The established system of patient registration with a GP was used for tracing and maintaining contact with a low-mobility cohort through local area health authority networks and the National Health Service Central Register. Results By 31 December 1997, there were 1856 recorded deaths (14 known to have occurred abroad), 66 emigrations/living overseas/lost from follow-up. In addition, 1500 study subjects had registered with new GPs who, every 2 years, provided information on both fatal and non-fatal cardiovascular events to complement information supplied by the original practices. Information was obtained on all cardiovascular events and deaths for 99.5 per cent of the surviving sample. Questionnaires mailed to surviving subjects 5 years after recruitment (1983-1985) and again in November 1992 and 1996 were returned by 98, 90 and 88 per cent, respectively, providing information on lifestyle changes, new cardiovascular symptoms and new diagnoses. Conclusion Using the NHSCR framework for primary care registration procedures, maximum follow-up has been maintained. Brief and structured enquiry forms have gained and maintained co-operation from subjects and their GPs with considerable success. Mortality reporting from dual sources identified 5 per cent more cases.