Anderson Cancer Center, Houston, Texas.
Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire.
J Bronchology Interv Pulmonol. 2023 Jul 1;30(3):207-222. doi: 10.1097/LBR.0000000000000920.
Diagnostic testing is fundamental to medicine. However, studies of diagnostic testing in respiratory medicine vary significantly in terms of their methodology, definitions, and reporting of results. This has led to often conflicting or ambiguous results. To address this issue, a group of 20 respiratory journal editors worked to develop reporting standards for studies of diagnostic testing based on a rigorous methodology to guide authors, peer reviewers, and researchers when conducting studies of diagnostic testing in respiratory medicine. Four key areas are covered, including defining the reference standard of truth, measures of dichotomous test performance when used for dichotomous outcomes, measures of multichotomous test performance for dichotomous outcomes, and what constitutes a useful definition of diagnostic yield. The importance of using contingency tables for reporting results is addressed with examples from the literature. A practical checklist is provided as well for reporting studies of diagnostic testing.
诊断测试是医学的基础。然而，呼吸医学中的诊断测试研究在方法、定义和结果报告方面存在显著差异。这导致了结果常常相互矛盾或不明确。为了解决这个问题，一组 20 名呼吸期刊编辑致力于制定基于严格方法的诊断测试研究报告标准，以指导作者、同行评审员和研究人员在进行呼吸医学中的诊断测试研究时使用。涵盖了四个关键领域，包括定义真实的参考标准、用于二分类结果的二分类测试性能的度量、用于二分类结果的多分类测试性能的度量，以及构成有用的诊断收益定义的内容。本文还通过文献中的示例说明了使用列联表报告结果的重要性。同时还提供了一份用于报告诊断测试研究的实用清单。
