Assessing the success probability of a Phase III clinical trial based on Phase II data.

Assessing the probability that a Phase III clinical trial will demonstrate clinically relevant efficacy based on Phase II data is an important topic in clinical drug development. An accurate estimate of how likely a Phase III trial will succeed based on available data will inform the decision on whether to move an experimental medicine forward to Phase III testing. Bayesian and likelihood methodologies have been developed in the literature to assess the probability of reproducibility in clinical trials for parametric models. A class of approaches that combines the Bayesian and likelihood approaches is proposed to evaluate the success probability of a Phase III trial based on Phase II data, which applies to the parametric, semi-parametric, and non-parametric settings and includes the Bayesian and likelihood approaches as special cases.