The Design and Testing of a New Approach to Computer-aided Differential Diagnosis^[*]

An algorithm is presented for making diagnostic inferences on the basis of a causal network model of medical knowledge. The algorithm is based on Bayes Rule, but is unique in the way that it accounts for the presence of conditional non-independence of observations and for the presence of multiple diseases in the same patient. An evaluation of the system is performed on a database of patients with chest pain. In this evaluation, the diagnostic accuracy of the system is found to be inferior to that of a logistic regression model and comparable to that of a linear discriminant function. In a review of selected cases from this database, the system can be shown to provide inferences that are not possible with other simpler statistical models. The practicality of this and other computer aids to medical diagnosis is discussed.

Es wird ein Algorithmus vorgestellt, der es erlaubt, auf der Grundlage eines kausalen Netzwerkmodells medizinischen Wissens diagnostische Schlüsse zu ziehen. Der Algorithmus basiert auf der Bayes’schen Regel, ist jedoch nicht einmalig in der Weise, daß er die Existenz bedingter Nicht-Unabhängigkeit von Beobachtungen und das Vorhandensein multipler Krankheiten beim selben Patienten begründet. Eine Bewertung des Systems wird anhand einer Datenbank von Patienten mit Brustschmerzen vorgenommen. Bei dieser Auswertung wird festgestellt, daß die diagnostische Genauigkeit des Systems derjenigen eines logistischen Regressionsmodells unterlegen und derjenigen einer linearen Diskriminanzfunktion vergleichbar ist. In einer Übersicht über ausgewählte Fälle aus dieser Datenbank kann gezeigt werden, daß das System Rückschlüsse zuläßt, die bei anderen, einfacheren statistischen Modellen nicht möglich sind. Die Praktikabilität dieser und anderer Computerhilfen für die ärztliche Diagnose wird erörtert.

Key-Words

Computer-aided Diagnosis - Bayesian Approach - Heuristics - Probability Theory - Knowledge Bases

Schlüssel-Wörter

Computerunterstützte Diagnose - Bayes’ Ansatz - Heuristik - Wahrscheinlichkeitstheorie - Faktenbanken

^* This work was supported by NIH grant LM 03376 from the National Library of Medicine.

• References

• 1 Blois M. S., Tuttle M. S., Sherertz D. D.. RECONSIDER: A Program for Generating Differential Diagnoses. Proceed. Fifth Ann. Symp Comput. Appl. Med. Care (Nov 1981; p. 263-268.
• 2 De Dombal F.. Surgical Diagnosis Assisted by a Computer. Proc. roy. Soc. Lond. B 1973; 184: 433-440.
• 3 Diamond G., Forrester J.. Analysis of Probability as an Aid in the Clinical Diagnosis of Coronary-Artery Disease. New Engl. J. Med 1979; 300: 1350-1358.
• 4 Gardner D.. ADVISOR—Clinical Suggestions by Computer. Technical Report No. 58, Section on Medical Information Sciences. University of California at San Francisco 1981
• 5 Gorry G., Kassirer J., Essig A., Schwartz W.. Decision Analysis as the Basis for Computer-Aided Management of Acute Renal Failure. Amer. J. Med 1973; 55: 473-484.
• 6 Greenfield S., Nadler M., Morgan M., Shine K.. The Clinical Investigation and Management of Chest Pain in an Emergency Department: Quality Assessment by Criteria Mapping. Med. Care 1977; 15: 898-905.
• 7 Gustafson D., Kiestly J., Ludke R., Larson F.. Probabilistic Information Processing: Implementation and Evaluation of a Semi PIP Diagnostic System. Comput. biomed. Res 1973; 6: 355-370.
• 8 Gustafson D., Shukla R., Delbecq A., Walster G.. A Comparative Study of Differences in Subjective Likelihood Estimates made by Individuals, Interacting groups, Delphi Groups, and Nominal Groups. Organ. Behavior Hum. Perform 1973; 9: 280-291.
• 9 Health Sciences Computing Facility University of California. Los Angeles. BMDP Biomedical Computer Programs 1979. University of California Press 1979
• 10 IBID. p. 840.2 A. 26.
• 11 Kelly C, Barclay S.. A General Bayesian Model for Hierarchical Inference. Organ. Behavior Hum. Perform 1973; 10: 388-403.
• 12 Ludwig D.. INFERNET—A Computer-Based System for Modeling Medical Knowledge and Clinical Inference. Proceed. Fifth Ann. Symp. Comput. Appl. Med. Care, Nov 1981; p. 243-252.
• 13 McNeil B., Keeler E., Adelstein S.. Primer on Certain Elements of Medical Decision Making. New Engl. J. Med 1975; 293: 211-215.
• 14 Pauker S., Gorry G., Kassirer J., Schwartz W.. Towards the Simulation of Clinical Cognition: Taking a Present Illness by Computer. Amer. J. Med 1976; 60: 981-996.
• 15 Patrick E., Margolin G., Sanghvi V., Uthurusamy R.. Pattern Recognition Applied to Early Diagnosis of Heart Attacks. Proceed. Medinfo 77, p 203-207 Amsterdam: North-Holland Publishing Company 1977;
• 16 Pipberger H., Klingeman J., Cosma J.. Computer Evalu- ation of Statistical Properties of Clinical Information in the Differential Diagnosis of Chest Pain. Meth. Inform. Med 1968; 7: 79-92.
• 17 Pople H., Myers J., Miller R.. DIALOG: A Model of Diagnostic Logic for Internal Medicine. Proceed. 4th Int. Joint Conf. Artific. Intell., Cambridge, MA 1977; p. 1030-1037.
• 18 SAS Supplemental Library Users Guide: SAS Institute, Inc. Cary, N.C. 1980
• 19 Shortliffe E.. Computer-Based Medical Consultations: MYCIN. New York: Elsevier/North-Holland 1976
• 20 Szolovits P., Pauker S.. Categorical and Probabilistic Reasoning in Medical Diagnosis. Artif. Intell 1978; 11: 115-144.
• 21 Weiss S., Kulikowski C, Amarel S., Safir A.. Model-Based Method for Computer-Aided Medical Decision Making. Artif. Intell 1978; 11: 145-172.