An efficient multiple-stage procedure for phase II clinical trials that have high response rate objectives.

Gehan's two-stage procedure is commonly used in phase II clinical trials to enable one to stop a trial early if the observed response rate in the first stage does not warrant further study and to obtain an adequate estimate of response rate if the trial passes the first stage. However, when the response rate of interest is relatively high (0.3 or more), Gehan's procedure can be inefficient since the trial may hardly be halted even though the actual response rate is low. In this paper, we extend Gehan's two-stage procedure to a more efficient multiple-stage procedure, which allows early termination of a phase II trial at multiple stages and provides precise estimate of the response rate if the trial reaches the final stage. The proposed multiple-stage procedure is suitable for phase II clinical trials that have high response rate objectives.