ABSTRACT
Modifications of hypotheses of on-going clinical trials based on accrued data can certainly have an impact on statistical power for testing the treatment effect with the pre-selected sample size. Modifications of hypotheses of on-going trials commonly occur during the conduct of a clinical trial due to the following reasons that (i) an investigational method has not yet been validated at the planning stage of the study, (ii) information from other studies is necessary for planning the next stage of the study, (iii) there is a need to include new doses, and (iv) recommendations from a pre-established data monitoring committee (DMC). In addition, to increase the probability of success, the sponsors may switch a superiority hypothesis (originally planned) to a non-inferiority hypothesis. In this chapter, we will refer to adaptive hypotheses as modifications of hypotheses of on-going trials based on accrued data. Depending upon the clinically meaningful difference (e.g., effect size, non-inferiority margin or equivalence limit) to be detected, sample size is necessarily adjusted for achieving the desired power when adaptive hypotheses are applied during the conduct of a clinical trial.
