Evaluating the Validity of a Bayesian Program for Predicting Stage in Hodgkin’s Disease^{[*] [**]}

A Bayesian model that provides probabilistic information about the spread of malignancy in a Hodgkin’s disease patient has been developed at the Tufts New England Medical Center. In assessing the model’s reliability, it seemed important to use it to make predictions about patients other than those relevant to its construction. The accuracy of these predictions could then be tested statistically. This paper describes such a test, based on 243 Hodgkin’s disease patients of known pathologic stage. The results obtained were supportive of the model, and the test procedure might interest those wishing to determine whether the imperfections that attend any attempt to make probabilistic forecasts have gravely damaged their accuracy.

Im Tufts New England Medical Center ist ein Bayes’sches Modell entwickelt worden, das probabilistische Information über den Grad der Bösartigkeit bei Patienten mit Morbus Hodgkin liefert. Um die Zuverlässigkeit des Modells zu ermitteln, schien es wichtig, Vorhersagen über andere Patienten zu machen als diejenigen, die für seine Erstellung maßgebend waren. Die Genauigkeit dieser Vorhersagen wurde statistisch getestet. In dieser Arbeit wird ein solcher auf 243 Fällen von Morbus Hodgkin mit bekanntem pathologischem Stadium basierender Test beschrieben. Die erzielten Ergebnisse bestätigen das Modell. Das Testverfahren könnte diejenigen interessieren, die feststellen wollen, ob Unvollkommenheiten, die zu jedem Versuch probabilistischer Vorhersagen gehören, deren Genauigkeit schwerwiegend beeinträchtigen.

Key-Words:

Bayesian Analysis - Medical Diagnosis - Hodgkin’s Disease - Model Validity .

Schlüssel-Wörter:

Bayes’sche Analyse - ärztliche Diagnose - Morbus Hodgkin - Modellvalidität

^* Supported by Grants A-19122 and CA-19619 from the National Cancer Institute, National Institutes of Health, and by Grant 1P 41 RR 01096-02 from the Division of Research Resources of the National Institutes of Health.

^** Address for reprints:Dr. Jane F. Desforges, Division of Hematology-Oncology, Box 841, Tufts New England Medical Center, 171 Harrison

• References

• 1 Carbone P. P, Kaplan H. S, Musshoff K. et al. Report of the Committee on Hodgkin’s Disease Staging Classification. Cancer Res 1971; 31: 1860-1961.
• 2 Feinstein A. R. Clinical Biostatistics XXXIX. The Haze of Bayes, the Aerial Palaces of Decision Analysis, and the Computerized Ouija Board. Clin. Pharmacol. Ther 1977; 22: 482-496.
• 3 Fryback D. Bayes’ Theorem and the Conditional Nonindepend-ence of Data in Medical Diagnosis. Comput. Biomed. Res 1978; 11: 423-434.
• 4 Rifkin R. D, Hood W. B. Bayesian Analysis of Electrocardiographic Exercise Stress Testing. New Engl. J. Med 1977; 297: 681-686.
• 5 Rutherford C. J, Desforges J. F, Davies B, Barnett A. The Decision to Perform Staging Laparotomy in Symptomatic Hodgkin’s Disease. Brit. J. Haem 1980; 44: 347-358.
• 6 Safran C, Tsichlis P. N, Bluming A. Z. et al. Diagnostic Planning Using Computer-Assisted Decision Making for Patients with Hodgkin’s Disease. Cancer 1977; 39: 2426-2434.
• 7 Schwartz W. B, Gorry G. A, Kassirer J. P. et al. Decision Analysis and Clinical Judgment. Amer. J. Med 1973; 88: 459-472.
• 8 Wagner G, Tautu P, Wolber U. Problems of Medical Diagnosis — A Bibliography. Meth. Inform. Med 1978; 17: 55-74.