Multiple Endpoints Analysis with Latent Variable via Pseudolikelihood

• Activity Number: 409
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Juanmei Liu*+ and Minglei Liu
• Companies: Medtronic and Medtronic
• Keywords: multiple endpoints ; latent variable ; Structural equation model

Abstract

The consideration of multiple endpoints of important outcomes is a critically important issue in the design of clinical trials. Specifying multiple endpoints for the evaluation of treatment efficacy, however, can increase the rate of false positive conclusions about the efficacy of a treatment.

We describe the use of latent variables in the design, analysis, and interpretation of clinical trials, by consideration of correlation/covariance structure of the multiple outcomes. Moreover, as categorical multiple endpoints are prevalent in clinical trials, a computational effective method - pairwise likelihood, was developed by authors to reduce dimensions and thus avoid complexity, resulting in ease of computation and interpretation.