Significant Design Components in General Two-stage Adaptive Procedures

• Activity Number: 361
• Type: Contributed
• Date/Time: Wednesday, August 10, 2005 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Tatsuki Koyama*+ and Allan R. Sampson and Leon J. Gleser
• Companies: Vanderbilt University and University of Pittsburgh and University of Pittsburgh
• Keywords: conditional power function ; sample size adjustment ; adaptive design

Abstract

Because of a growing interest in these flexible procedures, a number of two-stage adaptive designs have been proposed for use in phase III clinical trials. Rather than treating numerous types of designs separately, we introduce a common framework that provides a simple method for controlling many aspects of two-stage procedures, including type-I and type-II error rates and the maximum total sample size. The framework also enables us to compare different procedures under the same formulation and to learn the effects of different design specification on the performance characteristics of the resulting design. Some of the two-stage designs utilize conditional stage II characteristics on the stage I results and allow the stage II design to be chosen after stage I; others use unconditional characteristics to choose a stage II design a priori. Although there are some practical differences between these approaches, the proposed framework encompasses both and highlights their similarities.