Bayesian enhancement two‐stage design for single‐arm phase II clinical trials with binary and time‐to‐event endpoints. Issue 3 (21st February 2018)

• Record Type: Journal Article
• Authors: Shi, Haolun; Yin, Guosheng
• Is Part Of: Biometrics. Volume 74:Issue 3(2018)
• Page Start: 1055
• Page End: 1064
• Publication Date: 2018-02-21
• Subjects: Bayesian adaptive design -- Highest posterior density interval -- Phase II clinical trial -- Posterior probability -- Simon's design -- Survival endpoint; Biometry -- Periodicals; 570.15195
• DOI: 10.1111/biom.12864
• Languages: English
• ISSNs: 0006-341X

Abstract

Summary: Simon's two‐stage design is one of the most commonly used methods in phase II clinical trials with binary endpoints. The design tests the null hypothesis that the response rate is less than an uninteresting level, versus the alternative hypothesis that the response rate is greater than a desirable target level. From a Bayesian perspective, we compute the posterior probabilities of the null and alternative hypotheses given that a promising result is declared in Simon's design. Our study reveals that because the frequentist hypothesis testing framework places its focus on the null hypothesis, a potentially efficacious treatment identified by rejecting the null under Simon's design could have only less than 10% posterior probability of attaining the desirable target level. Due to the indifference region between the null and alternative, rejecting the null does not necessarily mean that the drug achieves the desirable response level. To clarify such ambiguity, we propose a Bayesian enhancement two‐stage (BET) design, which guarantees a high posterior probability of the response rate reaching the target level, while allowing for early termination and sample size saving in case that the drug's response rate is smaller than the clinically uninteresting level. Moreover, the BET design can be naturally adapted to accommodate survival endpoints. We conduct extensive simulation studies to examine the empirical performance of our design and present two trial examples as applications.