Clinical datafiles and expert systems: knowledge evaluation from data analysis

Medical expert systems (MES) are usually built with factual knowledge based on experts´ experience and medical books. Large clinical data files have been designed and are currently used in many medical domains. Data analysis allows numerical knowledge to be extracted from these data files. This numerical knowledge is complementary to the physician´s knowledge. In matters of open question, data analysis makes it possible to analyze the experts´ knowledge and verify how specific parameters answer the purpose of the expert system. This methodology has been used to come to a conclusion about the possibility of building an expert system for the diagnosis of coronary artery disease. An analysis was made of the data from 145 questionnaires consisting of 108 items selected by cardiologists