Ofstedal, Mary Beth, Zachary Zimmer, Grace Cruz, Angelique W. Chan, and Yu-Hsuan Lin. 2003. "Self-Assessed Health Expectancy Among Older Asians: A Comparison of Sullivan and Multistate Life Table Methods." Elderly in Asia Report No. 03-60. March 2003.
Self-assessed health has been found to be a strong predictor of changes in health and of mortality and has been included in many surveys of health and aging around the world. In this paper, we estimate expectancies in self-assessed health and compare these among older adults across four Asian settings (the Philippines, Taiwan, Singapore and Indonesia), making use of data from several national panel surveys conducted in the mid to late 1990s. All of these societies are undergoing rapid population aging and social and economic change, and there is much concern among policymakers about of the potential implications for future disease burden and associated informal and formal care demands. Yet, very little health expectancy research has been conducted in these settings. This paper is the first of a series of planned health expectancy analyses based on these panel surveys that will focus on alternative indicators of physical and mental disability. In the current analysis, self-assessed health is dichotomized into categories reflecting negative health ratings (e.g., poor/not good at all) versus positive or neutral health ratings (excellent to good/average/fair). In the first stage of the analysis we calculate health expectancy using the Sullivan method based on data from a single wave of each survey to compare trends in self-assessed expectancies by age and sex across settings. In the second stage we take advantage of the panel data by calculating health expectancy using multistate life table methods and compare these estimates with the Sullivan estimates. Results suggest that despite differences in the proportion reporting negative health across settings, patterns by age and sex are similar. Sullivan and multistate estimates also compare closely, except for Singapore, where there are very large transition rates from favorable to negative self-assessed health over the survey period.