Chest. 2020 Jul;158(1):S103-S112. doi: 10.1016/j.chest.2020.03.006.
Chest
Jennifer A Bullen
PMID: 32658645 DOI: 10.1016/j.chest.2020.03.006
Medical tests are procedures intended to detect, diagnose, characterize, or monitor a specific medical condition. Understanding the accuracy of a medical test is a critical part of informed decision-making in patient management, as it allows clinicians to appreciate the types of errors a medical test might be prone to making and how often it makes them. Designing a study to assess the performance of a medical test, however, presents unique challenges, from acquiring a reference standard to dealing with the complexities that arise when the test involves an interpretation by a human reader. This article provides an overview of design considerations in this context, including common biases and how to avoid them, statistical considerations, and reporting guidelines. A short list of questions is also provided, which can serve as a quick reference for anyone designing, implementing, or reviewing a study that intends to assess the performance of a medical test.
Copyright © 2020 American College of Chest Physicians. Published by Elsevier Inc. All rights reserved.
Keywords: diagnostic accuracy; statistics; study design