Keywords: adaptive design, sample size re-estimation, type-1 error control
In the face of uncertainty about expected treatment effects, adaptive sample size modifications at an interim analysis have become a popular tool when planning a clinical trial. However, a major regulatory concern exists around control of the type-1 error rate, the risk of concluding a beneficial treatment effect when in fact none exists. Generally, the concern arises because study data are unmasked and analyzed multiple times during the course of the study without appropriate statistical adjustments and preplanning. Several statistical methods have been proposed over the last two decades to address this issue. In this talk, we will discuss the sources of error rate inflation as well as review available options for type-1 error control, and provide advice on when and how to use them.