[Comparative clinical trials]

The therapeutic advantages of a newly developed cancer chemotherapy should be evaluated in comparison with the standard (or control) treatment. Therefore the comparative clinical trials should be designed so that the only differences observed among treatment groups are directly made by treatment program. This requires comparability of patients throughout the trial. The preferred method of evaluation is randomized controlled trials which can eliminate conscious or unconscious bias in assigning treatment. Before patients are randomized into the treatments of the trial, it is advisable to stratify patients according to known prognostic factors (stratified randomization) in an attempt to achieve equal distribution of these factors within each stratum. There are many difficulties with historical control study in the conduct of the comparative clinical trials, and its use may be acceptable only when randomized controlled trials are not feasible. The method of administration of anticancer drugs, patient eligibility, response parameters, toxicity of anticancer drugs, size and duration of the study, and advantages and disadvantages of multi-institution trials were also discussed within the framework of the comparative clinical trials.