A0557
Title: Accounting for delayed responses in group sequential clinical trial designs
Authors: Carolin Herrmann - Heinrich Heine University Duesseldorf (Germany) [presenting]
Abstract: Group sequential clinical trial designs are applied frequently and come with the advantage of potentially earlier decisions and reduced sample sizes. In their interim analyses, decisions are made on whether the trial is terminated early with a potential rejection of the null hypothesis or whether the trial is continued with another stage. Often, there occurs a time gap between enrolling a patient into a trial and measuring the respective outcome. This potential delay in the availability of the outcome measurement can affect the statistical analyses, especially when it comes to interim analyses. Some statistical methods that account for those delayed responses have already been proposed. The idea is often twofold: First, a decision is made on an irreversible recruitment stop, followed by a hypothesis test in case of stopping. Modifications of this idea are proposed. Moreover, results from a simulation study are presented where different options of accounting for delayed responses are compared with the option of not accounting for them. Results are discussed with respect to power and sample size. There will be a special focus on potential performance differences as well as the influence of the amount of pipeline data and interim information.
