Adaptive Designs for Dose-Ranging Studies Based on Group-Sequential Methods

• Activity Number: 686
• Type: Contributed
• Date/Time: Thursday, August 8, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yevgen Tymofyeyev*+ and Jose Carlos Pinheiro
• Companies: Janssen RD of J&J and Janssen R&D
• Keywords: adaptive designs ; dose-ranging studies ; group sequential methods

Abstract

In recent years, many adaptive design methods have been proposed for dose-ranging studies. These designs utilize complex algorithms and adaptation features, such as frequent changes in treatment allocation ratio, interplayed with early stopping rules and treatment arm selection. As a result, the control of Type I error rate often needs to be shown via simulation and, because of it, such studies (or study portions) would hardly be accepted as pivotal in regulatory submission. We consider a design based on the well-accepted group sequential methodology, utilizing spending functions in a dose-finding framework and compare its performance to other adaptive dose-finding designs reported in the literature. We also explore the possibility to include sample size re-estimation approaches without jeopardizing the theoretical strong control of Type I error rate.