Causal Inference with Latent Growth Mixture Modeling

• Activity Number: 684
• Type: Contributed
• Date/Time: Thursday, August 5, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Booil Jo*+
• Companies: Stanford University
• Keywords: causal inference ; principal stratification ; growth mixture modeling ; latent variable modeling

Abstract

In preventive intervention trials, longitudinal outcome patterns may hold important treatment implications and useful information for better strategizing future trials. In this context, growth mixture modeling (GMM) has been gaining its popularity, but little is known about the possibility of causal interpretation based on GMM analyses. In this study, we will focus on recently proposed GMM methods and systematically examine the feasibility/practicality of using GMM as a tool for estimating causal effects. Specifically, we will expand the conceptual framework of principal stratification (PS) so that it can accommodate exploratory components of GMM. Within this broader framework, we will examine the performance of GMM and examine how GMM and PS are connected. We utilize three different GMM procedures to improve GMM's ability to capture causal effects and to improve interpretability.