Adaptive Testing with Multiple Sample Size Adjustment
Lu Cui, Division of Biometrics I, OB, CDER, FDA, Rockville, MD 20852
Sample size determination is an important but often difficult issue in planning clinical trials. Due to lack of prior knowledge of the treatment effect of interest in a clinical trial, the sample size needed to achieve a desired statistical power may be underestimated, leading to an insignificant trial result. One way to solve this problem is to make the trial more flexible so that sample size reestimation and modification can be made at an interim stage of a clinical trial. In this study, a new adaptive statistical testing procedure is proposed. This new procedure allows one to adjust sample size at each of a sequence of analyses based on the information accumulated up to that analysis. The new procedure is easy to use and guarantees a controlled type I error rate. With the increase of sample size, the power of the new procedure increases. The new test can be viewed as an extension of a two-stage test and an extension of a group sequential test with a fixed sample size.