A Mixture Model for the Joint Analysis of Latent Developmental Trajectories and Survival

• Activity Number: 300
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Rinke Klein Entink*+ and Jean-Paul Fox and Ardo van den Hout
• Companies: TNO and University of Twente and Institute of Public Health
• Keywords: Bayesian ; MCMC ; Mixture modeling ; multilevel item response modeling

Abstract

A joint modeling framework is proposed that integrates a mixture multilevel item response component to model latent developmental trajectories, given polytomous response data, with a survival component for continuous time survival data. The joint model is illustrated in a real data setting, where the utility of longitudinally measured cognitive function as a predictor for survival is investigated in a group of elderly persons. Time-dependent cognitive function is measured using the generalized partial credit model given occasion-specific mini-mental state examination (MMSE) response data. The mixture model identifies subpopulations that are relatively homogenous in their latent growth trajectories of cognitive function. A parametric survival model is stratified on these subpopulations, and cognitive function as a continuous latent variable is included as a time-varying explanatory variable, along with other covariates. Within the Bayesian framework, a Markov chain Monte Carlo algorithm is developed for simultaneous estimation of the joint model parameters. Practical issues as model building and assessment are addressed using the DIC and various posterior predictive tests.