Self-rated health and mortality over three decades - Results from a time-dependent covariate analysis

Previous longitudinal studies assessing relative mortality risks associated with fair or poor self-rated health have differed in the extent to which observed relative risks are explained by disease burden and health risk factors. Gender and ethnic differences have rarely been assessed. The authors used proportional hazards models with time-dependent covariates to examine associations between fair or poor self-rated health and mortality over 28 years for 5,976 Alameda County Study respondents age 21 to 94 at baseline. Adjustments for a number of demographic variables, chronic conditions, mobility impairment, and health risk factors accounted for about half of the unadjusted relative risk. No gender or ethnicity differences in relative risk comparisons were found, but consistent with other studies, lower relative risks associated with increasing age were found. The authors conclude that self-rated health is a deceptively simple variable that likely measures a great deal more than disease burden.