Relating Self-Reported and Objective Health Indicators to Adult Mortality in Bangladesh

Omar Rahman, Harvard University

This paper tests the relative importance of self-reported and objective health indicators in predicting mortality among older respondents to the Matlab Health and Socioeconomic Study (MHSS). The survey, conducted in rural Bangladeshi region in 1996, is matched to subsequent mortality events through a unique Demographic Surveillance System (DSS). In addition to socioeconomic data, MHSS includes self-reported data on general health, health trajectory, mobility, and morbidity as well as objective measures of physical functioning. We relate each health measure to subsequent mortality risk, addressing several crucial questions: 1) Do self-reported health measures predict mortality as well as objective ones? 2) Do self-reported measurements explain mortality even in the presence of controls for objective health status? 3) To what extent are well-documented age and sex differences in self-reported health status in Bangladesh reflected in subsequent mortality differentials? 4) How would longitudinal mortality models be affected by depending exclusively on self-reported health data?