Abstract

STUDY OBJECTIVES: To identify socioeconomic and demographic predictors of long term mortality and institutional residence in old age, taking into account changes in socioeconomic and demographic circumstances between the 1971 and 1981 censuses. DESIGN: Multivariate logistic regression modelling of outcomes for 10 year age cohorts of each gender. The outcomes were death by 31 December 1992; being in an institution in 1991. SETTING: Members of the Longitudinal Study (a 1% sample of the British Census): 43,092 men and 50,839 women aged 55-74 in 1971. MAIN RESULTS: Being in rented accommodation and in a household without access to a car carried 35-45% higher mortality rate over 21 years and similar excess risk of being in an institution in 1991. Marital status and living arrangements were weaker predictors of death but being single was a major predictor of moving to an institution for men. Losing household access to a car was a strong factor for mortality for men and for institutionalisation for men aged 55-64 in 1971. The effects were weaker for women. Moving into rented accommodation was a predictor of both outcomes for women and of death for the younger cohort of men. People who started to live alone in the inter-census period were at reduced risk of dying. CONCLUSIONS: These results demonstrate persistence of inequalities in health related outcomes throughout old age, both in those with unfavourable circumstances in mid-life and in those who, in later life, have lost earlier advantages.