A self-adjusting randomization plan for allocation of patients into two treatment groups

This paper presents a self-adjusting randomization plan which is thought to be useful in controlled stratified randomized clinical trials. The method assures an optimal balance of treatments within each subgroup of patients. An imbalance caused by a high number of dropouts in one treatment group will be counteracted by the self-adjusting nature of the method. The method has been tested in computer-simulated trials and has been shown to produce better results than conventional plans for stratified randomized clinical trials. The method does not involve any complicated calculations and its use in a running clinical trial has shown that it is easy to handle.