Duke University School of Medicine, Durham, NC, United States.
Health Services Research Centre, Singapore Health Services, Singapore; Centre for Quantitative Medicine, Duke-NUS Medical School, Singapore.
Am J Emerg Med. 2018 Feb;36(2):185-192. doi: 10.1016/j.ajem.2017.07.054. Epub 2017 Jul 17.
Current triage methods for chest pain patients typically utilize symptoms, electrocardiogram (ECG), and vital sign data, requiring interpretation by dedicated triage clinicians. In contrast, we aimed to create a quickly obtainable model integrating the objective parameters of heart rate variability (HRV), troponin, ECG, and vital signs to improve accuracy and efficiency of triage for chest pain patients in the emergency department (ED).
Adult patients presenting to the ED with chest pain from September 2010 to July 2015 were conveniently recruited. The primary outcome was a composite of revascularization, death, cardiac arrest, cardiogenic shock, or lethal arrhythmia within 72-h of presentation to the ED. To create the chest pain triage (CPT) model, logistic regression was done where potential covariates comprised of vital signs, ECG parameters, troponin, and HRV measures. Current triage methods at our institution and modified early warning score (MEWS) were used as comparators.
A total of 797 patients were included for final analysis of which 146 patients (18.3%) met the primary outcome. Patients were an average age of 60years old, 68% male, and 56% triaged to the most acute category. The model consisted of five parameters: pain score, ST-elevation, ST-depression, detrended fluctuation analysis (DFA) α1, and troponin. CPT model>0.09, CPT model>0.15, current triage methods, and MEWS≥2 had sensitivities of 86%, 74%, 75%, and 23%, respectively, and specificities of 45%, 71%, 48%, and 78%, respectively.
The CPT model may improve current clinical triage protocols for chest pain patients in the ED.
目前胸痛患者的分诊方法通常利用症状、心电图（ECG）和生命体征数据，需要由专门的分诊临床医生进行解释。相比之下，我们旨在创建一个快速获取的模型，整合心率变异性（HRV）、肌钙蛋白、心电图和生命体征的客观参数，以提高急诊科胸痛患者分诊的准确性和效率。
2010 年 9 月至 2015 年 7 月，方便地招募了急诊科出现胸痛的成年患者。主要结局是在急诊科就诊后 72 小时内出现血运重建、死亡、心脏骤停、心源性休克或致死性心律失常的复合结局。为了创建胸痛分诊（CPT）模型，进行了逻辑回归，其中潜在的协变量包括生命体征、心电图参数、肌钙蛋白和 HRV 测量值。本机构当前的分诊方法和改良早期预警评分（MEWS）被用作对照。
共纳入 797 例患者进行最终分析，其中 146 例（18.3%）符合主要结局。患者平均年龄为 60 岁，68%为男性，56%分诊至最急性类别。该模型由五个参数组成：疼痛评分、ST 抬高、ST 压低、去趋势波动分析（DFA）α1 和肌钙蛋白。CPT 模型>0.09、CPT 模型>0.15、当前分诊方法和 MEWS≥2 的敏感性分别为 86%、74%、75%和 23%，特异性分别为 45%、71%、48%和 78%。
CPT 模型可能会改善急诊科胸痛患者目前的临床分诊方案。
