STROKE DEATH AND UNEMPLOYMENT IN LONDON

Study objective-The aim was to investigate the relationship between social factors and stroke mortality in men and women aged between 45 and 74 years using census and mortality data from 32 London boroughs in 1971 and 1981. Design-Census data from 1971 and 1981 on type of accommodation, density of room occupation, male unemployment rate, and proportion of households without a car were linked with stroke mortality available for each London borough. Setting-32 London boroughs excluding the City of London. Measurements and main results-The main outcome measures were the association between age adjusted mortality from stroke and the proportion of households with no car, non-ownership of home, in council housing or rented accommodation, male unemployment rate, and living density of more than 1.5 people per room. There was no strong correlation between social variables and stroke mortality in 1971, but strong correlations were found for male stroke mortality in 1981. The highest correlation was with male unemployment (r = 0.64, p < 0.001) even after adjusting for the proportion of the population born in the Caribbean and Africa (r = 0.56, p < 0.01). Other social variables were also highly correlated with male stroke mortality: households without a car (r = 0.63, p < 0.001), living density of more than 1.5 people per room (r = 0.053, p < 0.001), council housing (r = 0.45, p = 0.01), and rented accommodation (r = 0.36, p = 0.05). After regressing male mortality on unemployment rate the other social variables were no longer significantly correlated with male stroke mortality. In women, the only significant correlation was found in 1981 between stroke mortality and the proportion of families living in council housing (r = 0.34, p = 0.05). Conclusions-Social factors are important indicators of stroke mortality. The major increases in unemployment over the decade may explain the generally stronger association in 1981 compared with 1971. Male stroke mortality increased by 0.062/1000 for every one percent increase in male unemployment (0.054/1000 after adjusting for place of birth).