Patient Illness in Medical Diagnosis

In most medical AI programs, the use of notions such as causal relationships, temporal patterns, and aggregate disease categories has been limited. Yet studies of clinicians' behavior reveal that a diagnostic or therapeutic program must consider a case at various levels of detail to integrate overall understanding with detailed knowledge. To explore these issues, Ramesh Patil, Peter Szolovits, and William Schwartz have applied the knowledge-based approach in a detailed study of consultation for electrolyte and acid-base disturbances. The resulting program, Patil's dissertation work, is known as ABEL (for Acid-Base and ELectrolyte program). ABEL and an earlier M.I. T.lTufts program known as the Digitalis Therapy Advisor (Gorry et ai., 1978) were important departures from other systems in that they both viewed clinical problem solving as a process of constructing an explanation of man~festations, what they have called a patient-specific model. In ABEL, this description includes data about the patient as well as the program's hypothetical interpretations of these data in a multilevel causal network. Proceeding from the lowest level, the concepts and relations gradually shift in content from pathophysiological to syndromic knowledge. The aggregate level of this description summarizes the patient data, providing a global perspective for ~fficient exploration of the diagnostic possibilities. The pathophysiological description provides the ability to handle complex clinical situations arising in illnesses