Rael T Lange1,2,3,4, Sara M Lippa1,2,3. 1. a Defense and Veterans Brain Injury Center , Bethesda , MD , USA. 2. b Walter Reed National Military Medical Center , Bethesda , MD , USA. 3. c National Intrepid Center of Excellence , Bethesda , MD , USA. 4. d Department of Psychiatry , University of British Columbia , Vancouver , Canada.
Abstract
OBJECTIVE: The purpose of this paper is to highlight the risks and pitfalls of interpreting sensitivity and specificity in isolation when evaluating the clinical utility of a new test/measure to predict a target behavior, disease, or condition. METHOD: This paper (a) provides a primer for a set of metrics that can be used to appropriately examine the clinical utility of a test/measure (i.e. test operating characteristics - TOC), and (b) provides a discussion regarding various interpretative considerations when using TOCs that supports the tenet of this paper. CONCLUSIONS: There are two primary conclusions. First, sensitivity and specificity should never be interpreted in isolation as a means for evaluating the clinical utility of a test/measure. When sensitivity and specificity are used in this manner, more often than not, this will result in erroneous and/or incomplete conclusions regarding the clinical utility of a new test/measure. Second, sensitivity and specificity values are important, but they must be interpreted in conjunction with other more relevant clinical utility metrics (e.g. positive predictive power and negative predictive power).
OBJECTIVE: The purpose of this paper is to highlight the risks and pitfalls of interpreting sensitivity and specificity in isolation when evaluating the clinical utility of a new test/measure to predict a target behavior, disease, or condition. METHOD: This paper (a) provides a primer for a set of metrics that can be used to appropriately examine the clinical utility of a test/measure (i.e. test operating characteristics - TOC), and (b) provides a discussion regarding various interpretative considerations when using TOCs that supports the tenet of this paper. CONCLUSIONS: There are two primary conclusions. First, sensitivity and specificity should never be interpreted in isolation as a means for evaluating the clinical utility of a test/measure. When sensitivity and specificity are used in this manner, more often than not, this will result in erroneous and/or incomplete conclusions regarding the clinical utility of a new test/measure. Second, sensitivity and specificity values are important, but they must be interpreted in conjunction with other more relevant clinical utility metrics (e.g. positive predictive power and negative predictive power).
Entities:
Keywords:
Test operating characteristics; predictive power; psychometrics; sensitivity; specificity
Authors: Leontien C C Toussaint-Duyster; Monique H M van der Cammen-van Zijp; Dick Tibboel; Saskia Gischler; Joost van Rosmalen; Hanneke IJsselstijn Journal: Acta Paediatr Date: 2020-02-22 Impact factor: 2.299