Shifting coresidence near the end of life: comparing decedents and survivors of a follow-up study in China.

What we know about transitions in coresidence of elders in China is based on panel data involving survivors. This article examines the tendency to and determinants of shifts in coresidence with adult children among the very old, comparing survivors of an intersurvey period with those who died (decedents). Data come from the Chinese Longitudinal Healthy Longevity Survey Baseline and follow-up surveys indicate shifts in coresidence, defined as change from not living with an adult child to living in the same household as an adult child, and the converse. Rates of shifting are adjusted for time to follow-up. Regressions examine predictors of shifts among four groups: baseline coresident and noncoresident survivors and decedents. Decedents and noncoresidents are more likely to shift than survivors and coresidents. Covariates related to physical and material need as well as marital status are the strongest predictors of shift. Thus, the needs of a very old person dominate coresidential shifts and stability, lending support to an altruistic notion of living arrangement decision making. In the end, we conclude that the period nearing the end oflife is a time offlux in living situation and that coresidential shifts are underestimated when those who die during afollow-up study are ignored.