A Conditional Adaptive Weighted Test Method for Confirmatory Trials

The authors consider a statistically valid method that has little concern caused by down weighting in stage 2 or slight loss in efficiency and possibly improves power performance for a trial design that allows for adaptation of statistical information. In addition to the need for a priori sound data-based planning that accounts for limited uncertainty, the criteria for statistical information adaptation are scientifically sound and can foster transparent trial logistics that are necessary for such adaptation in confirmatory trials. A major rationale of the proposed conditional adaptive weighted test method is to increase the probability of success of an adequate and well-controlled confirmatory trial that may otherwise fall short of statistical significance despite initial rigorous planning for statistical information.