Adaptive Designs and Sequential Monitoring

University Of Virginia Main Campus, Charlottesville VA

Investigators

Abstract

Clinical trials are complicated and usually involve multiple objectives such as controlling type I error, increasing power of detecting treatment difference, assigning more patients to better treatment, and more. In literature, both adaptive designs (by changing design procedure sequentially) and sequential monitoring (by changing analysis procedure sequentially) have been proposed to achieve these objectives to some degree. In this project, the investigator combines these two sequential procedures and studies the advantages of sequential monitoring response-adaptive randomized clinical trial. The investigator first derives the asymptotic distribution of the sequential test statistics of the combined procedure. Based on the asymptotic properties, the investigator selects appropriate boundaries for the combined procedure to achieve multiple objectives. Further, he investigates the implementation of the combined procedure to some real clinical trials. Clinical trials usually involve multiple competing objectives such as detecting clinical difference among treatments, minimizing total cost and protecting more people from possibly inferior treatments. Adaptive designs dynamically use sequentially accruing data in decisions for collecting future data in order to achieve these objectives. In conducting clinical trials, sequential monitoring is a standard technique to balance the ethical and financial advantages of stopping a trial early against the risk of an incorrect conclusion. This proposal is concerned with combining adaptive designs and sequential monitoring. The investigator will first investigate sequential monitoring of a response-adaptive randomized clinical trial and will then study the advantages of this combined procedure. Upon completion of this project, one can apply sequential monitoring to an adaptive randomized clinical trial to achieve multiple objectives. The research projects will produce advanced statistical tools for analyzing data which is sequentially collected. These tools may be applied in many fields including drug development, medical studies, industrial experiments, economics and finance.