A design for decision making: construction and connection of knowledge bases for a diagnostic system in medicine

We describe the process of organizing medical knowledge into knowledge bases and designing one architecture for a decision support system in clinical psychiatry. We define a set of knowledge bases that we regard as the necessary and sufficient structures to represent the medical knowledge to provide clinical consultations: disease profiles; frames with semantic relations to represent clinical findings; production rules with probabilities, to relate findings with diagnoses; a hierarchical classification tree, to represent disease categories; heuristic questions, to narrow the diagnostic hypotheses; and diagnostic criteria to conclude the clinical investigation. We propose one new architecture for a support system connecting these knowledge bases in a particular way to simulate medical clinical reasoning.