Adaptive and Repeated Cumulative Meta-Analyses for Safety Signal Detection during a New Drug Development Process

• Activity Number: 390
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Hui Quan*+ and Yingqiu Ma and Yan Zheng and Meehyung Cho
• Companies: Sanofi and Sanofi and Sanofi and Sanofi
• Keywords: conditional power ; combined test ; Type I error rate control ; fixed and random effect models ; adaptive Bayesian design ; non-inferiority

Abstract

During a new drug development process, it is desirable to use cumulative data from all completed studies sequentially to timely detect potential safety signals. For this purpose, repeated meta-analyses can be performed on combined data from multiple completed studies. Moreover, if data from the originally planned program are not enough for ensuring power to test a specific hypothesis, adaptation in total sample size to increase the amount of safety data by adding new studies into the program can also be conducted. Without appropriate adjustment, Type I error rate will be inflated because of the repeated analyses and sample size adaptation. In this paper, we provide a systematic discussion on potential issues associated with adaptive and repeated cumulative meta-analyses conducted during a drug development process. We consider both frequentist and Bayesian approaches. Simulation results are provided to compare the performances of different methods. A new drug development example is used to demonstrate the application of the methods.