A Contribution to Adaptive Randomization Designs in Clinical Trials

Abstract

This thesis investigated methods in covariate adaptive randomization in clinical trials. A new method, referred to by the author as critical percentage method (CPM) is introduced. The CPM method is compared with the MIN and RM methods under various sample sizes. The comparison is based on descriptive and analytical approaches as well as simulation experiments. The study showed that the suggested method (CPM) superior to the RM method for all sample sizes in both single layers as well as total assignments. As for the MIN method the CPM revealed superiority only in the case of single layers.