Optimal allocation of sample sizes to multicenter clinical trials.

In this article, we discuss an approach for optimal sample size allocation in designing multicenter clinical trials. The method we studied was adapted from a stratified sampling survey design. The sample size allocated to centers is a function of the center's treatment cost, the standard deviation of the endpoint, and the availability of patients. We illustrate our approach using two hypothetical scenarios derived from our experiences in designing and conducting multicenter clinical trials. Simulation results are also presented.