Bayesian Approach for Interim Analysis with Failure Time Endpoint

• Activity Number: 506
• Type: Contributed
• Date/Time: Wednesday, August 3, 2016 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Zailong Wang* and Chen Gao and Peter Mesenbrink
• Companies: Novartis Pharma and University of Minnesota and Novartis Pharma
• Keywords: Hierarchical Bayesian Model ; Interim Analysis ; Failure Time Endpoint ; Dirichlet Process ; Gibbs Sampling ; Nonparametric Approach

Abstract

Semiparametric and Nonparametric methods for estimation of failure-time distributions are widely used. With limited data available at an interim analysis for a clinical trial study, it is difficult to use these methods to predict the distribution at new time points or the end of the trial. Several Bayesian semiparametric and nonparametric methods have been proposed to model the failure times. First, the baseline distribution in the accelerated failure-time (AFT) is modeled as a mixture of Dirichlet processes for right-censored data. Second, the Weibull distribution is considered in a hierarchical Bayesian model with a mixture of Dirichlet process. Gibbs sampling method is used to obtain posterior distributions in each case. These methods will be illustrated and compared with simulations and real examples.