Large variation in predictors of mortality by levels of self-rated health: Results from an 18-year follow-up study

Abstract Objectives To analyze the variation in factors associated with mortality risk at different levels of self-rated health (SRH). Study design Retrospective cohort study. Methods Cox regression analysis was used to examine the association between mortality and demographic, socioeconomic and health-related predictors for respondents with good, average, and poor SRH in a longitudinal data set from Estonia with up to 18 years of follow-up time. Results In respondents with good SRH, male sex, older age, lower income, manual occupation, ever smoking, and heavy alcohol consumption predicted higher mortality. These covariates, together with marital status, illness-related limitations, and underweight predicted mortality in respondents with average SRH. For poor SRH, only being never married and having illness-related limitations predicted mortality risk in addition to older age and male sex. Conclusions The predictors of all-cause mortality are not universal but depend on the level of SRH. The higher mortality of respondents with poor SRH could to a large extent be attributed to health problems, whereas in the case of average or good SRH, factors other than the presence of illness explained outcome mortality.