Overestimation of the effect size in group sequential trials

Group sequential designs (GSD), which provide for interim monitoring of efficacy data and allow potential early trial termination while preserving the type I error rate, have become commonplace in oncology clinical trials. Although ethically appealing, GSDs tend to overestimate the true treatment effect size at early interim analyses. Overestimation of the treatment effect may exaggerate the benefit of a drug and provide imprecise information for physicians and their patients about a drug's true effect. The cause and effect of such a phenomenon are generally not well understood by many in clinical trial practice. In this article, we provide a graphical explanation for why the phenomenon of overestimation in GSDs occurs. The potential overestimation of the magnitude of the treatment effect is of particular concern in oncology, in which the more subjective endpoint of progression-free survival has increasingly been adopted as the primary endpoint in pivotal phase III trials.