Predictive Power of Self-Rated Health for Subsequent Mortality Risk During Old Age: Analysis of Data From a Nationally Representative Survey of Elderly Adults in Taiwan

BACKGROUND Previous research has investigated differences in the predictive power of self-rated health (SRH) for mortality based on socioeconomic status (SES). However, these studies mainly assessed adults in the general population and did not focus specifically on elderly adults. In addition, this predictive power has never been evaluated using subjective SES, which is an important measure of SES in elderly adults. METHODS This study used data from the Survey of the Health and Living Status of the Middle Aged and the Elderly in Taiwan (SHLS) conducted by the Bureau of Health Promotion, Taiwan. The SHLS is a 15-year longitudinal survey based on a nationally representative sample. It was initiated in 1989 with 4049 respondents aged 60 years or older. Both education and subjective financial satisfaction were used as SES measures in the present study. A Cox regression model was used to estimate the interaction between SRH and SES for 3829 individuals without missing data. RESULTS As compared with those who reported their health as good, those who reported their health as poor and their education as high had a higher hazard ratio (hazard ratio = 1.97, 95% confidence interval = 1.35-2.88) for 6-15-year mortality, after adjusting for depressive symptoms, activities of daily living, and instrumental activities of daily living. This HR was significantly higher than those for adults with middle (1.16, 0.93-1.44) and low (1.27, 1.05-1.54) education, based on the χ(2) test (P < 0.05 for both comparisons). A similar pattern was observed when financial satisfaction was used as the SES measure. However, the pattern was attenuated when using 5-year mortality from baseline. CONCLUSIONS The use of SRH as a single health measure in elderly adults may yield inconsistent results across different SES groups, especially when used as a predictor of a longer-term mortality. This is true regardless of whether objective or subjective measures of SES are used, where both are important measures of SES in elderly adults.