The problem of aiding physician in early diagnosis and management of arterial hypertension has intensively been addressed in the form of indications, standards of management, recommendations, protocols, algorithms, etc. All those aids, however, relate more or less to the ‘average patient’ hardly met in medical practice. Using computeraided consultation system that allows for dialogue, provides quite new opportunity to switch from the ‘average’ to a real case that may present with a constellation of features and clinical manifestations.
Building a knowledge based computer consultation system (called HERMES - Hypertension Explanation by a Rule-based Medical Expert System) to aid primary care physician in explanation of causes for arterial hypertension with a respective expert knowledge base was the main objective of the study. The system may help doctors in putting well justified diagnostic hypotheses to be differentiated as well as in planning steps that could lead to diagnosis in a most effective way.
Numerous expert knowledge rules has been acquired as a basic material for HERMES development. A previously developed ELSA method of knowledge acquisition and processing (Expert Lattice Structured Acquirements) as well as the ELSA-Shell system for building medical knowledge bases have been applied to complete the project. The medical knowledge base on arterial hypertension diagnostic management includes definitions of 17 causes for arterial hypertension that might be taken as diagnostic hypotheses, 98 symptoms, signs and laboratory findings, 48 diagnostic questions (tests) and 13,328 medical knowledge rules. The knowledge base has also been applied for development of other aids to medical practice and training as the knowledge based model of patient problem-oriented history taking and extended assumptions for a textbook on differential diagnosis of arterial hypertension aimed at the primary care physician training.