What inference for two-stage phase II trials?

摘要: Simon’s two-stage designs are widely used for cancer phase II trials. These methods rely on statistical testing and thus allow controlling the type I error rates, while accounting interim analysis. Estimation after such trials is however not straightforward, several different approaches have been proposed. Different point confidence intervals estimation, as well computation of p-values reviewed compared a range plausible Cases where actual number patients recruited in trial differs from preplanned sample size also considered. For uniformly minimum variance unbiased estimator (UMVUE) bias corrected had better performance than others when was planned. intervals, using mid-p approach yielded coverage probabilities closer to nominal level so-called ’exact’ intervals. When differed UMVUE did perform worse an specifically developed situation. Analysis conditional having proceeded second stage required adapted analysis methods, (UMVCUE) can be used, which performs slightly The use may recommended it exhibited good properties both equal or value. Restricting cases stop early futility valuable, UMVCUE that case.