The use of concurrent controls in clinical trials has been advocated for a considerable period of time and by a large number of researchers. Nevertheless, the nature of clinical investigations appears to lead to continuing interest in the use of historical controls. The two opposing views are enunciated in papers by Byar et al1 and Gehan and Freireich.2
When a clinical trial is viewed as a scientific experiment, the use of concurrent controls, and especially randomized concurrent controls, is to be preferred based on long-standing principles of experimental design. There is little doubt that the ethics of clinical experimentation rightly impose constraints on what might otherwise be viewed as ideal experimental design.