Using the Community Health Index, general practitioner records and the National Health Service Central Registry for a 14 year follow-up of a middle-aged cohort in the west of Scotland.

Record linkage provides an opportunity for relatively inexpensive follow-up in cohort studies. Most studies have used hospital records. However, the Community Health Index, which is a computer-held list of all patients registered with General Practitioners (GPs), now offers opportunities for measuring morbidity at the primary care level. Searching for patients individually on this computer-held Index, we identified 96% of a middle-aged cohort recruited 14 years earlier in a longitudinal study in the West of Scotland. Of these, information was obtained for 78% through direct postal survey and for 87% through contact with GPs. In total, follow-up information was obtained on 903 (92%) of those thought to be alive. The National Health Service Central Registry records data and cause of death and will pass this information to special studies which have 'flagged' individuals' records. Seven per cent of the deaths in this sample had been missed by this process. This may need to be taken into account in previous studies of mortality for this cohort. New studies could expect successful long-term follow-up through record linkage to the Index or the Registry, and automatic record linkage with the Index would be worth investigating.