Covariates Adaptive Randomization Designs in Clinical Trials: A Comparative Study

摘要

This paper investigated in covariate adaptive randomization designs, which are used to reduce covariate variables imbalance between treatments in clinical trials. Critical percentage and imbalance minimization methods are compared each one to another, and both are compared with pure randomization method in term of imbalance. The comparison is intended to show which method has minimum imbalance at three covariate variables with twelve single layers and three sample sizes 10, 20 and 100. The results which carried out from the simulation experiment clearly shown that the performance of critical percentage approach is closely similar to imbalance minimization method in full balance case as well as maximum imbalance at all sample sizes. And pure randomization method has the maximum imbalance compared to others at each sample size.