Mortality effects of community socioeconomic status

Abstract

We linked data from the National Longitudinal Mortality Study to census tract information on 239,187 persons to assess 11-year mortality risk among black and white women and women associated with median census tract income, adjusted for individual family income from the Current Population Survey. We stratified Cox proportional hazards models by ages 25-64 years and 65 years and older. We used a robust covariance matrix to obtain standard errors for the model coefficients that account for correlation among individuals in the same census tract. Both income indicators were independently related to all-cause mortality. Among persons age 25-64 years, the rate ratios (RR) for individual family income and the median census tract income, respectively, for low income relative to high income were RR = 2.10 vs. 1.49 for black men, RR = 2.03 vs 1.26 for white men; and RR = 1.92 vs 1.30 for black women and RR = 1.61 vs 1.16 for white women. Among persons age 65 years or greater, only individual family income was associated with mortality, and only for white men. Although family income has a stronger association with mortality than census tract, our results indicate that, more broadly, area socioeconomic status makes a unique and substantial contribution to mortality and should be explored in health policy and disease prevention research.