Designs for Bayesian Adaptive Bioequivalence Trials

• Activity Number: 343
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Cheng Zheng*+ and Jixian Wang
• Companies: Novartis and Novartis
• Keywords: Adaptive Design ; Interim Analysis ; Bioequivalence ; Pharmacokinetic ; Bayesian Predictive Probability

Abstract

Bioequivalence (BE) trials play an important role in drug development for demonstrating the bioequivalence between test and reference formulations. The key statistical analysis for BE trials is the two one-sided tests (TOST). By assuming no interim looks, the sample size of the trials is often determined to achieve a given level of power based on TOST. However, some key parameters for sample size calculation could be quite uncertain. Allowing for one interim analysis with the possibility of sample size adjustment or early stopping for futility may be beneficial. In this work, we propose a simple yet efficient approach called Bayesian adaptive TOST (BATOST) based on p-value combination test for designing and analyzing BE studies with one interim look. Using Markov chain Monte Carlo (MCMC), Bayesian predictive probability is applied for sample size re-estimation and the calculation of the futility boundary at the interim look. The simulation results show BATOST controls the type I error rigorously while keeps the average sample size low. The loss of power is very small compared with either designs with fixed sample size or flexible designs without a strong control over type I error.