Study objective: To develop a community based stroke register to assess the magnitude of the problem of stroke in an entire health district in a high risk area for stroke.

Design: Community based stroke register from general practice data.

Setting: East Lancashire Health Authority with a 1995 population of 534,287.

Patients: The stroke register was developed and maintained for one calendar year in East Lancashire between 1 July 1994 and 30 June 1995. Efforts were made to include all patients who had a stroke during this period from participating general practices, using several sources of referral.

Main results: Of the district's 118 general practices, 93 (79%) participated fully, covering a population of 405,272. A total of 932 strokes, including 642 first ever cases, were cross checked and confirmed, with only 50% from any single source, mainly the practices. The total stroke incidence rate was 1.60 per 1000 per year, adjusted for the England and Wales 1991 census population. The rate increased considerably with age from 0.88/1000 for ages 50-54 to 20.56/1000 for ages 85-89 years. From 50-74 years, the age specific incidence was higher in men, but overall it was higher in women (1.87; 95% confidence interval 1.67, 2.04 per 1000) than in men (1.31; 1.15, 1.47 per 1000), and slightly lower than in Oxford a decade earlier. The rate also varied in different localities, with higher rates in the central towns of Hyndburn (2.05/ 1000), Blackburn (1.63/1000), and Burnley (1.80/1000) and lowest values in rural areas (1.18/1000 in Pendle). Case fatality from stroke at 28 days was 34% and the hospital admission rate was high at 70%.

Conclusions: The multiple source registration method is required for a stroke register. Stroke incidence in this area was still high and there was considerable variation across the district. Case fatality rates were similar to those in previous studies.