[What is the role of statistics in medical science?--a critical essay].

In the present paper, the fundamental problems of statistics used in medical science are discussed, based on radical criticism and thorough examination of statistical ideas and methods which are popularly used in the medical field of today. The formal applications of statistical test and estimation theories in mathematical statistics and further belief in double-blind experimental design are frequently found in medical science, where the meaning of population is not made clear and the true characteristics of "statistical test and estimation" and of "double-blind experimental design" are either not understood or are misunderstood. These facts bring many evil influences on medical ideas leading to medical trials. Radical criticism and thorough examination are necessary for the return of medical research to the right path. New statistical ideas and methods must be developed on the basis of the considerations mentioned above. The use of statistics must be made with a modest attitude and exploratory or detective ideas must be adopted, in so far as medical trials are administered, with human mind, to individuals who are intricate and multifarions in reaction. Medical trials and statistics are to be unified in order that desirable results may emerge in human world, enhancing the validity of their contents each other. Here, for example, the exploratory method of multidimensional data analysis and the idea of optimal control with QOL of individuals in medical trials are explained.