Section of Hospital Medicine, Division of General Internal Medicine, Weill Cornell Medical College, 525 E 68th Street, Box 331, New York, NY, 10065, USA.
Department of Clinical Epidemiology and Biostatistics, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, Canada.
J Gen Intern Med. 2019 Sep;34(9):1892-1898. doi: 10.1007/s11606-019-05125-0. Epub 2019 Jul 3.
Clinical decision-making relies heavily on making a correct diagnosis. Clinicians have a responsibility to understand the full spectrum of the diagnostic information conveyed by a physical exam finding, laboratory test result, or imaging. Many laboratory tests, such as troponin and B-type natriuretic peptide (BNP), are continuous tests with many possible results. Yet, there is a tendency to dichotomize tests into positive and negative, and use sensitivity and specificity to describe the test characteristics. This approach can lead to waste of important diagnostic information and substandard clinical decision-making. The aim of this paper is to demonstrate the role of ROC curves in developing a more comprehensive understanding of diagnostic information portrayed by continuous tests to augment clinical decision-making.
临床决策在很大程度上依赖于正确的诊断。临床医生有责任了解体格检查结果、实验室检查结果或影像学检查所传达的全部诊断信息。许多实验室检查,如肌钙蛋白和 B 型利钠肽(BNP),都是具有许多可能结果的连续检查。然而,人们倾向于将检查分为阳性和阴性,并使用敏感性和特异性来描述检查特征。这种方法可能导致重要诊断信息的浪费和临床决策的不达标。本文的目的是展示 ROC 曲线在更全面地理解连续检查所描绘的诊断信息以增强临床决策方面的作用。