Bayesian Analysis for Crossover Trials with Repeated Measurements in Soft Contact Lens Trials

• Activity Number: 245
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 2:45 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Quan Zhang*+ and Youssef Toubouti and Bradley P. Carlin
• Companies: University of Minnesota and Johnson & Johnson and University of Minnesota
• Keywords: Bayesian inference ; crossover trial

Abstract

A crossover trial is a form of longitudinal study in which each patient acts as his own control. Subjects are randomly assigned to an initial study arm, and receive each treatment sequentially. While many frequentist methods (e.g., GEE methods) exist for analyzing such data, Bayesian random effects models are natural but less well-developed in the literature. In this talk, we introduce a Bayesian method using Gibbs sampling. Using simulation, we show our approach can detect a true difference between two treatments with a specific false discovery rate that we can control via the posterior credible interval. We illustrate our Bayesian method using real crossover trial data with multiple endpoints, namely measures of contact lens vision, comfort, and handling. Assuming a multivariate normal distribution, our Bayesian method is able to detect treatment differences in both contact lens efficacy and safety.