[Skip to Navigation]
[Skip to Navigation Landing]
February 1993

Black-Box Mathematics and Medical Practice

Author Affiliations

From the Section of Otolaryngology, Department of Surgery (Dr Piccirillo), Department of Medicine (Dr Feinstein), and Clinical Epidemiology Unit (Drs Piccirillo and Feinstein), Yale University School of Medicine, New Haven, Conn. Dr Piccirillo is now with Washington University School of Medicine, St Louis, Mo.

Arch Otolaryngol Head Neck Surg. 1993;119(2):147-155. doi:10.1001/archotol.1993.01880140029005

Two recent articles in the Archives proposed mathematical models for a rational diagnostic and therapeutic approach to patients with diseases of the head and neck. In the first article,1 decision analysis was used to analyze choices of treatment for patients who have American Joint Committee on Cancer stage I floor-of-mouth carcinoma. In the second article,2 Bayes' theorem was used to construct a diagnostic algorithm for patients who were suspected of having cervical tuberculous lymphadenitis. In both studies, the clinical problem was approached with various computations obtained from a mathematical model, and the recommendations in each report diverged from current clinical practice.

In this report, we describe decision analysis and Bayes' theorem, present alternative interpretations of the two analyses, and suggest what clinicians can do to improve the quality of the data generated in clinical research as used in these models. The goal of this study review is to