Institute of Applied Health Research, University of Birmingham, UK.
Institute of Cardiovascular and Medical Sciences, University of Glasgow, UK.
Age Ageing. 2018 May 1;47(3):349-355. doi: 10.1093/ageing/afy023.
Diagnostic Test Accuracy (DTA) describes a field of research that aims to assess how well a test is able to detect or exclude a condition of interest. Although geriatric medicine is not as reliant on investigations as other medical disciplines, almost all patient encounters with older adults will involve some form of diagnostic assessment. Thus, understanding the terminology and methods of DTA is essential for any clinician. In this review we use examples based around the diagnosis of dementia to highlight issues in DTA research. Some of these are generic to any DTA research and some are particularly pertinent to older adults. One can apply a test accuracy framework to a clinical question by defining four key components: the condition of interest; the index test(s) (i.e. the assessment(s) of interest); the reference standard (the best available method for assessing the condition of interest) and the population or healthcare setting in which testing takes place. Test accuracy is often described using complementary measures of sensitivity and specificity. However, many other metrics to describe test accuracy are available; in clinical practice predictive values may have greater utility. These and other descriptive statistics can be derived from a two by two table that cross-classifies the index test results with the reference standard results. Test performance and utility is not only determined by accuracy, other measures such as feasibility and acceptability should be considered and may be of particular importance when describing test performance in older adults with physical and cognitive impairments.
诊断测试准确性（DTA）描述了一个旨在评估测试检测或排除感兴趣疾病的能力的研究领域。尽管老年医学不像其他医学学科那样依赖于检查，但几乎所有与老年人的患者接触都会涉及某种形式的诊断评估。因此，了解 DTA 的术语和方法对于任何临床医生来说都是必不可少的。在这篇综述中，我们使用围绕痴呆症诊断的示例来突出 DTA 研究中的问题。其中一些问题是任何 DTA 研究都通用的，而另一些问题则特别与老年人有关。人们可以通过定义四个关键组成部分来将测试准确性框架应用于临床问题：感兴趣的疾病；索引测试（即感兴趣的评估）；参考标准（评估感兴趣疾病的最佳方法）以及进行测试的人群或医疗保健环境。测试准确性通常使用敏感性和特异性的补充措施来描述。但是，还有许多其他描述测试准确性的指标可用；在临床实践中，预测值可能具有更大的实用性。这些和其他描述性统计数据可以从交叉分类索引测试结果和参考标准结果的四格表中得出。测试性能和实用性不仅取决于准确性，还应考虑其他措施，例如可行性和可接受性，并且在描述身体和认知障碍的老年人的测试性能时可能特别重要。
