Practical Issues in Reviewing Confirmatory Trials with Sequential Design

• Activity Number: 443
• Type: Topic Contributed
• Date/Time: Wednesday, August 6, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Yun Wang*+ and Rajeshwari Sridhara
• Companies: FDA and FDA
• Keywords: Confirmatory trial ; sequential design ; multiple testing ; control type I error

Abstract

Group sequential design has been widely used in confirmatory clinical trials to control multiplicity in testing primary endpoint at different time points, such as interim analysis and final analysis. In addition, many multiple testing procedures have been proposed for testing secondary endpoints. We will discuss some common issues we encounter in testing secondary endpoints in confirmatory studies with group sequential design: What alpha level we should use for testing secondary endpoint(s) if the time of testing secondary endpoint is dependent on when the primary endpoint achieves statistical significance? Reporting secondary endpoints in the labeling is more challenging than handling a bunch of p values. If a study is stopped early due to overwhelming superiority in primary endpoint at interim analysis, how should we report secondary endpoint in the labeling if the data on the secondary endpoint are not mature yet? Our discussion will be based on some illustrative examples from clinical trials we reviewed.