The 2-in-1 design allows the possibility of seamlessly expanding a phase II study to confirmatory phase III study and controls type I error without multiplicity adjustment. In this paper, we applied the CHW method to the 2-in-1 design strategy, and compared it with the unweighted conventional test statistics. It shows that when the interim decision threshold is high enough, the CHW method is slightly more powerful. Otherwise, results based on the CHW method can be difficult to interpret when the estimated treatment effects differ notably between interim and final analysis, which may be avoided by using the conventional test statistic.
展开▼