Latent Growth Mixture Models for Longitudinal and Discrete Survival Data

• Activity Number: 495
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 8:30 AM to 10:20 AM
• Sponsor: SSC
• Author(s): Depeng Jiang* and Robert Tate
• Companies: University of Manitoba and University of Manitoba
• Keywords: discrete-time survival data ; growth mixture modeling ; multivariate longitudinal data ; Manitoba Follow Up Study

Abstract

This study proposes joint latent growth mixture models for jointly analyzing multivariate longitudinal and discrete-time survival data. Growth mixture modeling represents unobserved heterogeneity between subjects in their development using both random effects (e.g., Laird and Ware, 1982) and finite mixtures (e.g., McLachlan and Peel, 2000). We begin by exploring separate joint trajectory models for each outcome variable and discrete-time survival data by using the discrete-time survival mixture model proposed by Muthen & Masyn (2005). Then this discrete-time survival mixture model was extended to jointly modeling the multivariate longitudinal data (e.g., trajectories of mental health and trajectories of physical health) and discrete-time survival data. We use the Data from the Manitoba Follow-up Study as an illustration. The joint trajectory models help us identify five distinct trajectories of mental health and four distinct trajectories of physical health. The links between mental health and physical health trajectories and differences in the mortalities among these trajectories will be discussed.