ABSTRACT
The optimal design minimizes the expected sample size under the null hypothesis, and the minimax design minimizes the maximum of the total sample size. The optimal design is the one where the expected sample size under the null hypothesis is the smallest among all feasible designs. Recall that a sequential design requires a larger maximum sample size than the fixed-size design. Computer programs are also available for calculating Simon's two-stage designs. A common issue of concern with Simon’s design is its rigid rules. In practice, trials are often interrupted for additional review of the study status or end up with deviations from the original design, either fewer patients enrolled due to accrual shortage or more patients enrolled due to over-running. In order to reach the same statistical power, a sequential trial would require a larger maximum sample size than the corresponding fixed sample size for a nonsequential design.
