Historical and methodological aspects of computer-assisted medical history-taking

The purpose is to draw, in a historical perspective, the main outlines of a theoretical and methodological approach of computer-assisted anamnesis (CAA). At the University of Liège (Belgium), studies on CAA were started in 1966, as an extension of previous works in the field of computer-assisted instruction (CAI). A short retrospect of the SIAM-DOCEO Project is given. Various factors contribute to reducing the place physicians effectively assign to history-taking, in spite of its essential role. Non-interactive self-administered questionnaires cope only very partially with that situation. For about 20 years, a number of CAA projects have been set up in many countries. Four basic questions dealing with the diffusion of CAA techniques are examined. (a) It is crucial not to put man-machine interaction and human dialogue on the same footing. A computer is quite unable to understand anything, but it can help health professionals in gathering medical history data. (b) As a rule, a CAA program should intervene in a limited and well defined pathological field, i.e. not too far from the diagnostic stage. (c) Acceptance by patients is very encouraging. (d) In contrast, reactions are mixed among physicians, essentially because of subjective factors. Accuracy and reliability, completeness and relevancy are, as a whole, undeniable qualities of CAA data, but the computer is--hopefully only for a short time--still often perceived more as a competitor than as a tool. The bibliography comprises 84 references.