Two-Stage Sample Size Reassessment Using Perturbed Unblinding

Sample size reassessment (SSR) is an increasingly popular strategy for designing and conducting clinical trials. In particular, SSR based on updating the variance estimate is a prudent practice accepted by the regulatory authorities to assure adequate power for a study. Since its development in the early 1990s, however, debate has continued over whether a treatment-blinded or unblinded approach should be used for SSR based on the variance estimate. A blind procedure is preferred from the regulatory standpoint, because it better preserves the study integrity; however, it does not provide the best-unbiased estimate of the variance. On the other hand, the usual unblinded analysis reveals the treatment effect, which leads to controversy regarding the interpretation of the targeted effect size as well as concerns of inflating the Type I error and possibly biasing the trial. In this article, we devise a novel solution to this problem, one that uses perturbed unblinding to estimate the variance but still keeps the treatment effect masked. We then give a bias-corrected final test, which preserves the Type I error rate. We also discuss several points of consideration, with a focus on the issues of SSR, that were raised at the recent workshop on adaptive designs held jointly by the U. S. Food and Drug Administration (FDA) and the Pharmaceutical Research and Manufacturers of America (PhRMA). In particular, we propose a switch of paradigm from SSR to a two-stage design for clinical trials to alleviate the concern of possible change of behavior due to change of sample size.