All Times EDT
Keywords: Conditional Power, Current Trend, Auxiliary Data
This session will focus on the reliability of the conditional power (CP) when used to inform decision making such as sample size re-estimation. Although the CP based on any assumption for the data yet to be observed is always different from the true underlying CP, in practice many study designs treat it as if they were the same. In this work we evaluate the difference between the estimated and true CPs and its impact on the decisions in general. To account for long term follow-up or population shift, we also propose alternative methods to calculate the CP in the presence of auxiliary data. We will provide recommendations based on simulation studies.