Comparison Between Tree Gatekeeping Procedure and Graphical Approach in a Pivotal Clinical Trial with Multiple Objectives and Multiple Endpoints

• Activity Number: 599
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Masakazu Fujiwara*+ and Hideaki Watanabe
• Companies: SHIONOGI & CO., LTD. and SHINOGI CO & LTD
• Keywords: multiplicity ; gatekeeping ; graphical approach ; type I error ; pivotal study ; non-inferiority

Abstract

There are quite a few pivotal clinical trials with multiple objectives and multiple endpoints. If it is indispensable to reflect all of results for those endpoints into drug labeling, multiplicity adjustment is strictly required. In this paper, in a pivotal clinical trial for specific disease, we examined a framework for multiplicity adjustment that only if non-inferiority is shown for each of primary and multiple secondary endpoints, superiority is tested for each of those endpoints in order to obtain admission of the new drug from the authority and drug labeling which would differentiate it from other drugs. In this framework, considering priority of hypotheses and logicality, tree gatekeeping procedure may be one of the appropriate procedures to control the family wise error rate. However, if a hypothesis in hypothesis family is significant, this procedure has difficulty in use the significant level effectively for the other hypotheses within the family. Therefore, we consider graphical approach as other option which could improve the difficulty for tree gatekeeping procedure. In this paper, we compare these two procedures through simulation studies in the above framework.