Alternatives to Bayes?

Recent proposals have suggested that rule-based systems of diagnostic inference are an attractive medium for computer-aided diagnosis, in part because clinicians find their behaviour easy to understand. Bayesian systems have been more prominent in this field to date, but no direct comparison of their clinical abilities has been reported. A rule-based system that was closely modelled on clinical thinking is described and a quantitative comparison with a successful Bayesian system for the diagnosis of »dyspepsia« is presented. The results suggest that the rule-based approach may have considerable potential as an efficient alternative to Bayesian inference.

Kürzlich unterbreitete Vorschläge deuten an, daß regelbezogene Systeme diagnostischer Schlußfolgerung ein attraktives Medium für die Computer-imterstützte Diagnose sind, teilweise weil die Kliniker ihre Verhaltensweise leichter verständlich finden. Bayes’sche Systeme sind bis heute auf diesem Gebiet führend, aber es gibt keine Berichte über direkte Vergleiche ihrer klinischen Fähigkeiten Ein eng nach klinischem Denken modelliertes, regelbezogenes System wird beschrieben, und ein quantitativer Vergleich mit einem erfolgreichen Bayes’schen System zur Diagnose von »Dyspepsie« wird vorgestellt. Die Ergebnisse deuten an, daß der regelbezogene Ansatz beträchtliches Potential als wirksame Alternative zur Bayes’schen Schlußfolgerung haben könnte.

Key-Words:

Computer-aided Diagnosis - Pattern-directed Inference - Bayes Rule - Decision Making

Schlüssel-Wörter:

Computer-unterstützte Diagnostik - »mustergesteuerte« Schlußfolgerung - Bayes-Regel - klinische Entscheidungsfindung

• References

• 1 Croft D. J. Is computerised diagnosis possible? Comput. Biomed. Res. 1972; 5: 351-367.
• 2 de Dombal F. T. Computers and the surgeon — a matter of decision. Surg. Ann 1979; 11: 33-57.
• 3 Fox J. Some observations on fuzzy diagnosis and medical computing. Internat. J. biomed. Comput 1977; 8: 269-275.
• 4 Fox J. Medical diagnosis : inference, recall and a theory of skill. (Manuscript submitted for publication).
• 5 Fox J. Packet memory : a heuristic base for perception and cognition. (Manuscript in preparation).
• 6 Fox J. Towards a reconciliation of fuzzy logic and standard logic. (Manuscript submitted for publication).
• 7 Horrocks J. C, de Dombal F. T. Computer aided diagnosis of dyspepsia. Amer. J. Digest. Dis 1975; 20: 397-496.
• 8 Hoerocks J. C, de Dombal F. T. Diagnosis of dyspepsia from data collected by a physician’s assistant. Brit. med J 1975; III: 421-423.
• 9 Horrocics J. C, de Dombal F. T. Clinical presentation of patients with dyspepsia. Gut 1978; 19: 19-26.
• 10 Johnson-Laird P. N. Models of Deduction. In Joffe- Falmagne R. (Edit.): Reasoning: Representation and Process Hillsdale/N. J.; Lawrence Erlbaum Ass: 1975
• 11 Staudenmayer H. Understanding Conditional Reasoning with Meaningful Propositions. In Joffe-Falmagne R. (Edit.): Reasoning: Representation and Process Hillsdale/N. J.; Lawrence Erlbaum Ass: 1975
• 12 Rogers W, Ryack B, Moeller G. Computer-aided medical diagnosis: literature review. Internat. J. biomed. Comput 1979; 10: 267-289.
• 13 Shortliffe E. H. Computer-based Medical Consultations : MYCIN. New York: American Elsevier; 1976
• 14 Shortluffe E. H, Buchanan B. G. A model of inexact reasoning in medicine, Math. Biosci 1975; 23: 351-379.
• 15 Wagner G, Tautu P, Wolber XT. Problems of Medical Diagnosis. A. Bibliography. Meth. Inform. Med 1978; 17: 55-74.
• 16 Waterman D. A, Hayes-Roth R. Pattern-directed Inference Systems. New York: Academic Press; 1978