An Application of Bayes' Rule to Diagnostic Test Evaluation

Abstract

It is often the case that a clinician has diagnostic values such as sensitivity and specificity available for a certain diagnostic test but not the positive or negative predictive values. Because the clinician uses these predictive values to make decisions concerning the wellbeing of the patient, it is important to be able to compute them from the sensitivity and specificity. This article presents a well-established theorem called Bayes' rule for doing this. A brief, intuitive development of Bayes' rule and the framework for this application is given with a minimum of mathematics and without proofs. Several examples are provided. © 1990, Sage Publications. All rights reserved.