Fundamentals of Controlled Clinical Trials^{[*] [**]}

Summary

The result of therapeutic research should, if possible, also benefit the recovery of the patient. The basic rule for every therapeutic-clinical trial must involve a comparison of therapeutic approaches. In acute conditions, such as acute infectious diseases, infarcts, etc., comparisons should be made between two or more groups: the collective therapeutic comparison, i.e. the between-patients trial. The formation of groups to be compared one with the other can be justified only if one is reasonably sure that a pathogenic condition does indeed exist. In chronic diseases, which extend essentially unchanged over a lengthy period but are nevertheless reversible, therapeutic comparisons may be made between two or more time intervals within the course of the disease in the same individual. This type of therapeutic trial rests primarily upon a (refined!) type of specious reasoning and secondarily upon modified statistics: the individual therapeutic comparison, i.e. the within-patient trial. The collective therapeutic comparison on the one hand and the individual therapeutic comparison on the other hand overlap somewhat in scope. The immediate therapeutic effect is not always an indication of its true value which may become evident only upon long-term treatment. The short-term trials of therapeutic regimens in an individual must, therefore, be frequently supplemented by long-term trials which can only be carried out by comparing two groups. For many clinical investigations the joint efforts of numerous hospitals are, therefore, absolutely necessary.

The second basic rule of therapeutic research is the elimination of secondary causes. The difficulties introduced by these secondary considerations are far greater in therapeutic trials carried out on ambulatory patients than has hitherto been realized. In order to remove subjective secondary causes, the author demanded in 1931 the use of hidden or illusory media (placebos, dummies), i.e. unconscious causative agents. The double blind procedure is indispensible in all experiments, especially where psychological processes and problems are of foremost importance. Therefore, it is absolutely necessary that convincing tests be finally carried out with the double blind procedure on the one hand and the simple blind procedure on the other hand in order to determine the range and limitations of both. Therapeutic research of the future is threatened by the hazard of self-satisfaction which produces a sense of accomplishment merely as the result of obeying formal conditions and of following formal standards of accuracy.

^* Introductory lecture held at the International Seminar for Medical Documentation and Statistics, 16-28 October 1961, Berlin. Originally published in German language in “Methods of Information in Medicine”, Vol.1 (1962), pp. 1-5.

^** Translated by Mrs. L. Blumenthal