UNDERSTANDING THE SENSITIVITY, SPECIFICITY, AND PREDICTIVE VALUES USED IN DIAGNOSTIC TESTS

Abstract

Numerous diagnostics and screening tests have been routinely used for screening diseased or infected from healthy and normal individuals. Clinicians rely on diagnostic and screening test results to make decisions on the diagnosis and initiate the treatment procedure. However, the diagnostic test outcomes vary from different tests and not always the outcome is 100 percent accurate. Hence the tests showing more accuracy and high sensitivity and specificity are given high priority by the clinicians. To evaluate the performance of dichotomous binary outcomes obtained from diagnostic test results, several statistical measures have been used. They are accuracy, sensitivity, specificity, positive predictive value, and negative predictive value and are intimately connected with probability calculations. Very often interpreting concepts such as false positive, and true negative are quite intuitive, but many students and even health professionals have difficulties in assessing the associated probabilities.  In this article, we will explain the terms and the statistical measures and how to relate and interpret them used in diagnostic tests.