Modeling Disability Trajectories among the Oldest-Old in China

Zachary Zimmer, University of Utah
Linda G. Martin, RAND Corporation
Daniel Nagin, Carnegie Mellon University
Bobby Jones, Carnegie Mellon University

This paper identifies and describes common disability and mortality trajectories among oldest-old in China and assesses associations with socio-demographic characteristics. Data are from the China Longitudinal Health Longevity Survey conducted in 22 Chinese provinces. About 9,000 individuals 80 and older were interviewed in 1998. Survivors were followed up in 2000, 2002, and 2005, and mortality was recorded for non-survivors. The outcome is a count of limitations with activities of daily living (ADLs). A semi-parametric, finite mixture model determines typical trajectories in number of limitations, while mortality is modelled simultaneously. Covariates are employed to determine characteristics of those who fall into trajectory groups. Results indicate three trajectories representing “healthy,” “developing,” and “high” disability groups. Women, more than men, are likely to be in the “developing” and the “high” groups. Further analyses examine additional covariates and how drop-out due to mortality associates with trajectories.