A useful testing strategy is proposed for a confirmatory phase III clinical trial. It consists of a combined test of superiority and test of equivalence, and it is easy to apply. By introducing the strategy, we can perform apost hocanalysis in a confirmatory experiment. Thus a more flexible decision will be possible than the usual single testing method provides. It is shown that the procedure needs no adjustment for multiplicity from the point of view of the closed testing procedure. The relationship between this strategy and a confidence interval is also discussed.
展开▼