Marital status and mortality in British women: a longitudinal study.

BACKGROUND Most previous studies on marital status and mortality did not adjust for the effect of 'marital selection'. Little research has been done about the relation between marital status and mortality in British women, with the exception of research on bereavement. METHODS Subjects consisted of women aged > or = 35 in a longitudinal study of a nationally representative sample. Marital status and covariates were enumerated at a baseline interview in 1984/85 and a follow-up interview in 1991/92. Death data up to May 1997 were obtained from the National Health Service Central Register. Cox regression was used to estimate hazard ratios (HR) for the single, divorced and widowed states in relation to the married state. RESULTS Having adjusted for age and martial selection factors, being single (HR = 1.45) was significantly associated with higher all-cause mortality. Being divorced and being widowed showed no excess mortality risk (each HR = 1.09). CONCLUSIONS Being single was associated with higher mortality. A causal interpretation is plausible. Being divorced and being widowed were not associated with higher mortality.