A new expert system for evaluating cardiosurgical patients inintensive care units: preliminary validation

摘要

A new expert system for the analysis of cardiosurgical patients inIntensive Care Units (ICUs) is described and preliminary validationsperformed. The system employs a hybrid inference engine. Whichintegrates both quantitative and qualitative reasoning in an originalway. Causal knowledge of the main relationships among hemodynamic,metabolic and respiratory quantities is used to make predictions aboutthe patient status starting from the value of a few quantities monitoredin ICUs. Examples concerning six real patients are presented anddiscussed. In all cases, results of the expert systems substantiallyagree with those previously obtained using more traditional statisticalclassification techniques. As an essential improvement, the expertsystem also provides information on deep physiological quantities andcausal relationships which may affect the patient status