Exploring Bayesian Analogs to Two Frequentist Methods in Noninferiority Testing Through Examples

• Activity Number: 222
• Type: Topic Contributed
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Margaret Gamalo*+ and Ram Tiwari
• Companies: FDA/CDER/OB and FDA

Abstract

The frequentist framework has two widely used methods to determine non-inferiority -- fixed margin and the synthesis methods. In the Bayesian framework, there are procedures that are analogous to these methods. For example, a hierarchical Bayes procedure that allows historical information to be incorporated on the active control (C) via the use of informative prior but uses the fixed margin paradigm for the decision criterion based on the posterior distribution of the comparison of treatment effects of the test treatment (T) and active control (C) (Gamalo et al., 2013) is analogous to the fixed margin approach although it can also be extended to be analogous to the synthesis approach. On the other hand, a network meta-analysis of the data from the historical trials and the non-inferiority trial which is used to predict the difference between putative placebo (P) and the test treatment (T) (Schmidli et at, 2013) is analogous to the synthesis approach. We will explore these methods in some examples vis-à-vis the frequentist methods.