Statistical control in medical administration

The development of medicine as the science of health, in contradistinction to medicine as the art of healing has hitherto relied mainly on animal experiments supplemented by autopsies on the pathological dead and, more recently, by bio- chemical tests on the body fluids of patients and on bioautopsies. The limitations of animal experiments have been recognized at least since the time of John Hunter. While they can vastly stimulate the pace at which clues emerge and indeed furnish clues which could never emerge from experience of human beings alone, the advances made possible by them depend on some method of validating their applicability to the human situation. From time to time, attempts to solve this problem have been made by studying human reactions directly under labor- atory conditions. Thus Hunter infected himself experimentally with syphilis. When inoculation by direct infection was being tried out in the 18th century, before JennerTs discovery of vaccination as a method of immunization against small -pox, a trial was made on six convicted felons at Newgate prior to inoculating two members of the Royal Family. More recently, a well -known experiment, tried out on human volunteers in the Amerioan Army, incriminated the mosquito Aedes aegypti as the vector of yellow fever. There have been many others.This approach has several limitations, of which it suffices to mention one. Human beings differ individually far more than mice or guinea -pigs. In general, validification of laboratory discoveries must therefore rest on the experience of large numbers of cases. For such validification, albeit reinforced by biochemical tests and the study of morbid anatomy, medicine has in large measure relied on the physician's judg- ment to interpret the slow accumulation of clinical experience. This must be at best a lengthy process; the result will lack scientific exactitude, and it is clear that medicine can reach its full stature as an applied science only when it builds on a more secure foundation than personal judgment, however well informed. That the analysis of statistical material offers an alternative has been recognized from time to time during the past hundred years. In 1381 Sir John Simon, medical adviser to the Privy Council and later to the Local Government Board stated:"The experiments which give us our teaching with regard to causes of disease are of two sorts: on the one hand we have the carefully pre -arranged and comparatively few experiments which are done for us in our Pathological Laborator- ies, and for the most part on animals other than man; on the other hand, we have the experiments which Accident does for us, and, above all, the incalculably large amount of crude experiment which is popularly done by man on man under our present ordinary conditions of social life, and which gives us its results for our interpretation.