All Times EDT
Keywords: Trial monitoring, Conditional power, Bayesian predictive power, Predictive probability
Interim monitoring for efficacy in clinical trials offers opportunities to assess the projected probability of success (PoS) of the trial so that informed decisions can be made regarding resource planning for sponsors. Commonly used approaches to assess PoS at interim analyses include conditional power (CP), Bayesian predictive power (BPP), and predictive probability (PP). While these statistical methods have been used in numerous studies, the properties and assumptions of these methods are not clear to many trialists, including study statisticians. More importantly, variation in interpretations of the measures calculated from these methods can be challenging when it comes to communication to the study team. This poster provides a summary of CP, BPP, and PP, as well as a theoretical comparison among the three. A comprehensive simulation study is performed to evaluate the operating characteristics of each approach assuming different efficacy scenarios. A discussion surrounding the implications and interpretations of the decisions using these measures at the interim analysis will also be included.