Statistics for diagnostic procedures. I. How sensitive is "sensitivity"; how specific is "specificity"?

Abstract

The terms of medical decision making (i.e., accuracy, sensitivity, specificity, predictive value, false positive/negative rates) are often seen in the radiologic literature. The interrelationships among these terms are explained, with special emphasis on the following: (1) How useful a procedure is for detecting/excluding disease depends on sensitivity, specificity, and prevalence of disease, and thus varies from practice to practice, and (2) "Accuracy" is often very misleading.