Optimize Promising Zone in an Adaptive Design for a Confirmatory Study

• Activity Number: 36
• Type: Contributed
• Date/Time: Sunday, July 29, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Jung Wook Park*+ and Xiaosha Sherman Zhang and Jay Yang and Chaofeng Charles Liu
• Companies: Astellas Pharma Global Development, Inc. and Astellas Pharma Global Development, Inc. and Astellas Pharma Global Development, Inc. and Astellas Pharma Global Development, Inc.
• Keywords: adaptive design ; promising zone ; confirmatory study

Abstract

In recent years, implementation of optimal adaptive designs for confirmatory clinical trials with the limited budget has been of great interest in the pharmaceutical industry. A popular method is to find a fixed range of conditional power (namely promising zone) that would provide most promising results after sample size adaptation such as Mehta and Pocock (2010)(MP hereafter). In this presentation, we will propose an alternative approach to define a promising zone, which maximizes conditional power increase, and introduce an approximate permutation test to evaluate the performance of the promising zone in terms of relative efficiency, compared to a random adaptation and the MP promising zone adaptation. A simulation study will be presented to show the operating characteristics of the proposed promising zone. Then, we will compare statistical properties between using CHW vs. not using CHW in promising zone based adaptive design.