Modern Approaches to Clinical Trials Using SAS: Classical, Adaptive, and Bayesian Methods

10.8 Summary

In this chapter we gave an exposition on optimal RAR designs for clinical trials with binary outcomes. These designs use accumulating data from patients in the trial to modify treatment randomization probabilities to achieve an allocation that is optimal for some pre-specified experimental objective(s). A common objective is to minimize the number of treatment failures in the study subject to a constraint on power of a statistical test. Optimal RAR designs can result in a modest reduction in treatment failures in the trial, while maintaining important statistical properties such as power and estimation efficiency. Therefore, such designs can be attractive in clinical trials where outcomes are grave and even a modest reduction in ...