英国国家早期预警评分对不同疾病老年急诊患者死亡预测的研究  

A validation of national early warning score （NEWS） in emergency elderly patients with different diseases in China

作　　者：刘芳艳[1] 李春盛[2] 何庆[3] 王秀杰[4] 张国强[5] 郭树彬[6] 沈洪[7] 张新超[8] 郑亚安[9] 秦俭[1] 

机构地区：[1]首都医科大学宣武医院急诊科, 北京100053 [2]首都医科大学附属朝阳医院急诊科 [3]成都市第三人民医院/重庆医科大学附属成都第二临床学院 [4]哈尔滨医科大学附属第一医院急诊科 [5]卫生部中日友好医院急诊科 [6]中国医学科学院北京协和医院急诊科 [7]中国人民解放军总医院急诊科 [8]卫生部北京医院急诊科 [9]北京大学第三医院急诊科

出　　处：《中华急诊医学杂志》2015年第11期1248-1252,共5页Chinese Journal of Emergency Medicine

基　　金：卫生行业科研专项项目基金

摘　　要：目的 本研究探讨了英国国家早期预警评分（national early warning score,NEWS）对我国不同疾病的老年急诊患者死亡的预测价值.方法 多中心前瞻性研究.以我国不同地区的九所三甲医院在2013年10月至2014年5月之间急诊就诊的不同疾病老年患者作为研究对象,现场采集进入急诊的数据.进行NEWS评分,以急诊就诊期间的死亡为观察终点,按转归分为死亡组和存活组.计量资料为非正态分布,两组间比较采用两独立样本秩和检验;计数资料组间比较采用x2检验.绘制NEWS评分对所有患者急诊死亡预测的工作特征曲线（ROC曲线）,计算约登指数,以最大约登指数时的评分为临界值（cut off）.分别绘制各病人群使用NEWS评分进行死亡预测的ROC曲线以评价NEWS评分对不同疾病死亡预测的分辨度,并以Hosmer-Lemeshow拟合优度检验评价NEWS的校准度.以P＜0.05为差异具有统计学意义.结果 共收集了1 528份的进入急诊的数据.研究发现NEWS对我国老年患者的急诊死亡预测的ROC曲线下面积（95％ CI）为0.834（0.778～0.890）,其中对呼吸系统、消化系统疾病患者死亡预测的ROC曲线下面积（95％ CI）分别为0.885 （0.816 ～0.954）、0.858 （0.737～0.979）,而对于心血管系统患者的ROC曲线下面积（95％ CI）则相对较低,为0.798 （0.503～1.000）.结论 NEWS对我国急诊老年患者具有很好的预测价值,其中对呼吸系统疾病患者的预测能力较强,但对于心血管系统疾病患者的预测能力则相对较弱,因此我们还可以进一步增加相应的预测参数以提高早期预警评分的预测能力.Objective To assess the performance of NEWS for emergency elderly patients with different diseases in China.MethodsA prospective cohort study was carried out by recruiting elderly patients with different diseases from nine hospitals across China.Data of elderly patients admitted between 1 st October, 2013 and 31 May, 2014 were collected and the patients were assessed onsite by using NEWS.The in-hospital death was the ultimate observation.The patients with in-hospital death were compared to those alive with regard to their initial physiological parameters by using Mann-Whitney U test, and Chi-square analysis was applied to Yates＆#39; continuity correction.The data were used to create receiver operator characteristic curve （ROC） so that the effectiveness of the scores in determining outcome could be assessed.The different areas under the curve （AUROC） were compared according to the method of Hanley and McNeil.Patients were categorized into different groups according to their primary diseases.Each group＆#39;s ROC curves were created to assess the effectiveness of the scores among different diseases in determining outcome.Calibration was assessed by using the Hosmer-Lemeshow （H-L） test.P 〈 0.05 was considered significant.Results Data from 1 528 elderly patients were collected.The AUROC （95% CI） for the NEWS ability to predict in-hospital death were 0.834 （0.778-0.890）.For patients with respiratory, and gastrointestinal diseases, NEWS had good discrimination, as the AUROCs （95% CI） for NEWS ability to predict mortality were 0.885 （0.816-0.954）, and 0.858 （0.737-0.979）, respectively.While for patients with cardiovascular diseases, the discrimination of NEWS was not so good, as the AUROCs （95% CI） for NEWS ability to predict mortality was 0.798 （0.503-1.000）.Conclusions The performance of NEWS for emergency elderly patients is valid and reliable.For patients with respiratory, gastrointestinal diseases, NEWS had good performance of discrimination to predict mortality.However,

关 键 词：英国早期预警评分 老年 死亡 预测 呼吸系统 心血管系统 

分 类 号：R459.7[医药卫生—急诊医学]