Joint Modeling of Latent Group-Based Trajectory Models with Subdistributions

• Activity Number: 194
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Nilesh Shah*+ and Chung-Chou H. Chang and John A. Kellum
• Companies: University of Pittsburgh and University of Pittsburgh and University of Pittsburgh
• Keywords: joint model ; subdistribution ; trajectory analysis ; latent variable ; biomarker

Abstract

In clinical research, patient care decisions are often easier to make if patients are classified into a manageable number of groups based on homogeneous risk patterns. Investigators can use latent group-based trajectory models to discover latent behavioral patterns within a population. These models were designed to group longitudinal trajectories over time. Oftentimes, researchers are interested in uncovering behavioral groups based on event time outcomes as well as longitudinal measures. We propose a method to estimate these groups under the joint modeling framework. The proposed joint model involves three submodels: the first one models the latent risk trajectory groups; the second one models the longitudinal pattern of biomarkers conditional on a specific risk group; and the third one models the subdistribution function conditional on a specific risk group. The model is applied to a study of patients suffering from acute kidney injury. We used Interleukin-6 measurements as longitudinal biomarker measures and recovery of acute kidney injury as the outcome of interest. The analysis showed two distinct behavioral patterns based on the longitudinal and survival measures.