Causal expert systems supporting medical decision making and medical education: explanations based on simulated situations

The experience gained in developing CHECK, a diagnostic expert system combining heuristic and causal knowledge, is discussed. This experience has made it possible to use pathophysiological information in assessing the compatibility and consistency of diagnostic conclusions. In addition it has provided the opportunity of using the causal network to inquire directly about the domain, entering the system at different levels (findings, states and hypotheses) independently from patient data.<<ETX>>