A method for the quantification of the decision-making process in a computer-oriented medical world.

The aim of this paper is to suggest an original approach to the decision-making process in a computer-oriented medical record, now in use on an experimental basis, at a major District Hospital in Sesto S. Giovanni, near Milano, in a general medicine department. Two types of algorithms are introduced in order to 'capture' more effectively the different steps of the decisional process on the basis of a differential diagnosis approach. The first refers to the field of symbolic reasoning, close to the area of Artificial Intelligence and uses score variables, while the other can handle qualitative expressions since it follows the fuzzy-set approach. A comparison among the performances offered by such algorithms, the physician's decision and the Bayes Rule is then carried on. Further developments should clarify the problems of parameter sensitivity as regards the successive steps to the final diagnosis. The present implementation is suitable for individual hospital departments and educational purposes.