Statistics Seminar - Stephen Walter - Statistical Issues in the Decision to Abandon a Clinical Trial before its Completion
Speaker: Stephen Walter (McMaster University)
Title: Statistical Issues in the Decision to Abandon a Clinical Trial before its Completion
Abstract: We will discuss a variety of reasons why investigators might decide to stop a clinical trial early. Then we will examine some of the statistical implications of abandoning the original design following interim analyses; in particular, we will briefly review methods to control the inferential Type I error rate and assess conditional power at interim analyses, and examine recent work on the implications for the estimation of treatment effect sizes.