Reasoning Foundations of Medical Diagnosis Revisited
M Fieschi, J. Gouvernet · Yearbook of Medical Informatics · 1999
InButterworth'sMedicalDictionary ( 1965) diagnosis is defined as the art of applying scientific methods to the elucidation of the problems presented by a sick patient.Thus, it made sense to try and incorporate mathematical methods that would be likely to enrich this art.As soon as computers were available to physicians, the question of arriving at a diagnosis using automatic methods became a major research topic.Naturally, while automatic methods would not provide a tool to replace the physician, they were nevertheless useful in several respects in analyzing how a diagnosis was reached along with the mathematical and statistical tools proposed to formalize the problem.The aim of this pioneering research was usually to arrive at a correct diagnosis by being able to incorporate all the information contained in the medical record.The technology of the period had something to do with this approach.It must be remembered that only batch processing was possible at the time, and the interactivity required to design a step-by-step decisionsupport tool, and that the best strategy was out of reach.We will come back to this point later on.Among all the articles that appeared before 1960, the article by Ledley and Lusted [ 1] is the most important in our view.It presents in perfectly clear 78 Commentary Reas. oning Foundations of Medical Diagnosis RevisitedReflections on R.