Influence of Prior on Bayesian Sample Size Calculation

• Activity Number: 259
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Kevin Hou* and Gongfu Zhou and Yufan Zhao
• Companies: Incyte Corporation and Incyte Corporation and Incyte Corporation
• Keywords: Prior ; Sample Size ; Robust

Abstract

Bayesian designs have been increasingly used in clinical development, especially in phases I and II trials. In this article, we focus on sample size calculation for single arm trials with response rate as the primary endpoint. We implement the predictive expectation criterion to account for all possible trial outcomes, therefore avoid the need to hypothesize the trial outcome. We evaluate the property of our approach under similar setting as Mayo and Gajewski (2004). Our result is comparable with that of Mayo and Gajewski (2004), and seems slightly more robust.