A Bayesian Subgroup Analysis Using an Additive Model

• Activity Number: 462
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Yang Xiao*+ and Siva Sivaganesan and Purushottan Laud and Peter Mueller
• Companies: University of Cincinnati and University of Cincinnati and Medical College of Wisconsin and The University of Texas MD Anderson Cancer Center
• Keywords: additive model ; Bayesian subgroup analysis ; stepwise model selection ; multiplicity

Abstract

Subgroup analysis in clinical trials examines whether a treatment efficacy differs among the sub-population defined by one or a set of baseline characteristics, which will potentially influence future health decision making and research for clinical investigators. However, difficult issues such as multiplicity adjustment, sample size and power need to be addressed and reported correctly. We extend a Bayesian approach in Sivaganesan et. al, by using a single additive model with covariates each at two levels to allow for the treatment-covariate interaction. Posterior probabilities along with carefully chosen threshold values are used to decide on subgroup effects.We account for multiplicity in the Bayesian sense by the use of suitable prior distributions for the inclusion parameter(s), and in the frequentist sense by control of Type I error rate. Frequentist operating characteristics of the approach is investigated under various scenarios considering various error rates specific to subgroup analysis.