COMPARISON OF A SCREENING TEST AND A REFERENCE TEST IN EPIDEMIOLOGIC STUDIES .2. A PROBABILISTIC MODEL FOR COMPARISON OF DIAGNOSTIC TESTS
COMPARISON OF A SCREENING TEST AND A REFERENCE TEST IN EPIDEMIOLOGIC STUDIES .2. A PROBABILISTIC MODEL FOR COMPARISON OF DIAGNOSTIC TESTS
复制标题
DOI：
10.1093/oxfordjournals.aje.a120610
复制
发表时间：
1966-01-01
影响因子：
5
通讯作者：
BUCK, AA
中科院分区：
文献类型：
--
作者：
GART, JJ;BUCK, AA
在不同人群或同一人群的不同年龄组中比较两种诊断测试时，可能会发现测试之间的一致性指数差异很大。我们在本文中介绍了一个概率模型，它表明，这种变化的关联可能是一个数学的神器，只是由于不同的人口之间的疾病的患病率进行比较。也就是说，如果该模型适用于特定的一对诊断测试，则它们的诊断质量可能是稳定的，而同时，它们的共阳性和共阴性可能随着患病率而变化。当模型拟合时，如果知道另一种测试的灵敏度和特异性，则可以从这些数据中获得一种测试的灵敏度和特异性的估计值。
In comparing two diagnostic tests in different populations or in various age groups of the same population, it may be found that the indices of agreement between the tests vary greatly. We have introduced in mis paper a probabilistic model which shows that this variation in association may be a mathematical artifact, simply due to the difference in the prevalence of the disease among the populations being compared. That is, if this model holds for a particular pair of diagnostic tests, their diagnostic qualities may be stable, while at the same time, their co-positivity and co-negativity may vary with the prevalence. When the model fits, it is possible to obtain from such data estimates of the sensitivity and specificity of one test, if one knows the sensitivity and specificity of the other test.