ABSTRACT
This chapter introduces Pocock's test, O'Brien and Fleming's test, Wang and Tsiatis' test, and the inner wedge test for clinical trials with group sequential designs. It discusses the application of these tests to discrete study endpoints such as binary responses and time-to-event data. It outlines the concept of alpha-spending function in group sequential methods. The chapter describes procedures for sample size reestimation at a given interim analysis without unblinding. It derives conditional powers at interim analyses for the cases when comparing means and proportions and also discusses some practical issues. Conditional power at a given interim analysis in group sequential trials is defined as the power of rejecting the null hypothesis at the end of the trial conditional on the observed data accumulated up to the time point of the planned interim analysis. In addition, group sequential methods and their followed estimation procedures so far are concentrated only on the population average.
