In reply to: "Validating an HIV risk score".
In reply to: "Validating an HIV risk score".
复制标题
回复：“验证艾滋病毒风险评分”。
DOI：
10.1016/j.ajem.2014.09.016
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
Haukoos,JasonS
中科院分区：
文献类型：
--
作者：
Hsieh,Yu-Hsiang;Rothman,RichardE;Haukoos,JasonS
我们感谢您的来信和对我们学习的兴趣。然而，尽管有这些担忧，我们相信我们的数据和报道的结果证明了经验性开发的临床预测工具(在这种情况下，是丹佛艾滋病毒风险评分的缩写版本)在城市急诊科(ED)识别未诊断艾滋病毒感染风险增加的患者方面的潜在价值。1丹佛艾滋病毒风险评分不是作为一种工具来精确量化患者的艾滋病毒风险，而是将患者分类为不同的风险阶层，以帮助临床医生确定哪些患者应该接受艾滋病毒检测，哪些患者不应该接受艾滋病毒检测。因此，开发这种临床预测工具的一个目标是评估它在不同人群和环境中的表现如何，包括在马里兰州巴尔的摩的两个急诊室。在这些情况下，包括我们在本刊文章中报告的工作，丹佛艾滋病毒风险评分表现得非常好。1-3接下来的合乎逻辑的步骤包括评估有针对性的艾滋病毒筛查的有效性，将评分作为一种工具来识别风险增加的患者。我们不认为评分表现上的细微差异会显著影响它在实际临床实践中的工作方式，这是由如何报告其校准或区分所定义的。4.
We appreciate the correspondence and interest in our study. However, in spite of the concerns raised, we believe our data and findings as reported demonstrate the potential value of an empirically-developed clinical prediction instrument (in this case, an abbreviated version of the Denver HIV Risk Score) for identifying patients at increased risk for undiagnosed HIV infection in an urban emergency department (ED). 1 The Denver HIV Risk Score was not developed as an instrument to precisely quantify a patient's HIV risk, but rather to categorize patients into differing risk strata in order to help clinicians identify patients who should or should not be offered HIV testing. A goal in developing this clinical prediction instrument, therefore, was to assess how well it performed across different populations and settings, including in two EDs in Baltimore, Maryland. In these instances, and including the work reported in our article in this Journal, the Denver HIV Risk Score has performed very well. 1-3 The next logical steps include assessing the effectiveness of targeted HIV screening using the Score as an instrument to identify patients at increased risk. We do not believe the slight differences in how the Score performs, as defined by how its calibration or discrimination is reported, will significantly influence how it works in actual clinical practice. 4