Applications of Bayesian Model Selection for Clinical Safety Data

• Activity Number: 493
• Type: Topic Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Bradley McEvoy*+ and Rajesh R Nandy and Ram C. Tiwari
• Companies: FDA/CDER and University of California at Los Angeles and FDA/CDER/OTS/OB
• Keywords: Drug Safety ; Bayesian Model Selection ; Multiplicity adjustment

Abstract

In drug safety, development of methods for multiplicity adjustment, when attempting to identify adverse events that exhibit a differential in risk effects between treatment and comparator arms, have exploited potential relationships among adverse events according to underlying medical features. This is a non-trivial problem resulting from the multi-dimensional characterization of an adverse event. In order to preserve the complex relationships inherent in drug safety data, we propose an alternative framework based on Bayesian model selection. This framework is shown to be simple yet extremely flexible, with numerous possible extensions. The methodology is illustrated using existing data.