|
Abstract:
|
In an adaptive phase IIa/IIb study to explore multiple doses as well as to establish efficacy and safety on the right dose(s), traditionally, we use an interim analysis to select dose arms that are expected to have high benefit and low risk or to drop dose arms that are expected to have low benefit and high risk compared with placebo. In the case of a time to event primary clinical endpoint, we may encounter a problem that when we collect enough events for an interim analysis to make a decision, all or many of the planned subjects have already been randomized. Therefore, in order to expedite a decision at the interim, we may use a surrogate endpoint that may be available sooner at interim. This research explores the risks, for example, of either false positive or false negative occurring, that may occur when a surrogate endpoint is used at interim analysis for decision making. The usage of surrogate endpoints in interim analyses is illustrated for a potential adaptive study design. Simulations are presented assuming different concordance and correlation between the clinical outcome and the surrogate endpoint to elucidate the risks.
|
