Futility analyses are becoming more commonly used in clinical trials. There are statistical issues that arise from the use of futility analyses that warrant a collaborative discussion between the FDA and Industry. Some, but not all, of these issues include the following: type I error adjustment, power (particularly in terms of pediatric studies), use of Bayesian predictive probability and conditional power.

Questions for discussion: 1.) Why adjust the type I error at all since there is no scientific basis for adjusting for type I error inflation in the context of binding futility analyses? 2.) In the context of pediatric studies where there is often limited power but are conducted as part of a pediatric written request is the utility of futility analyses viewed differently from a regulatory standpoint? 3.) What are the benefits/risks of different strategies for optimally conducting futility analyses (e.g., Bayesian approaches, timing, etc.)?