Li Yin-xian, Ye Hong-mei
The Third People's Hospital of Mianyang City, Mianyang, Sichuan, China.
Zhongguo Wei Zhong Bing Ji Jiu Yi Xue. 2008 Jul;20(7):419-21.
To study the prediction value and practicality of modified early warning score (MEWS) grade for death of the patients in intensive care unit (ICU).
According to the predetermined criteria, the clinical data of patients who were admitted to ICU during January 2005 to August 2006(patients with ICU stay exceeding 2 hours were enrolled for study) were collected, and MEWS grade and acute physiology and chronic evaluation II (APACHE II) scores were assessed (90 days as the end point of the observation period). MEWS and APACHE II's receiver operator characteristic (ROC) curve, and their predication index were calculated and compared.
The MEWS's and APACHE II's score in non-survivors were higher than the survivors. This difference had statistically significant difference (both P<0.01). In prediction of death of the patients in ICU, the MEWS grade > or =5 scores, Sensitivity was 89.66%, Specificity was 86.21%, accuracy rating was 87.93%, Youden index (YI)=0.758 6 and area under curve (AUC)=0.911 4. However for the APACHE II, they were grade > or =16 scores, Sensitivity was 96.55%, Specificity was 79.31%, accuracy rating was 87.93%, YI=0.758 6 and AUC=0.898 9.
MEWS grade is a useful index to predict the death of the patients in ICU. It is simple and practical, therefore it should be recommended in clinical practice.
研究改良早期预警评分(MEWS)分级对重症监护病房(ICU)患者死亡的预测价值及实用性。
按照预定标准,收集2005年1月至2006年8月入住ICU的患者临床资料(入住ICU时间超过2小时的患者纳入研究),评估MEWS分级及急性生理与慢性健康状况评分系统II(APACHE II)评分(观察期终点为90天)。计算并比较MEWS和APACHE II的受试者工作特征(ROC)曲线及其预测指数。
非存活者的MEWS和APACHE II评分高于存活者。差异具有统计学意义(均P<0.01)。在预测ICU患者死亡方面,MEWS分级≥5分时,灵敏度为89.66%,特异度为86.21%,准确率为87.93%,约登指数(YI)=0.758 6,曲线下面积(AUC)=0.911 4。而对于APACHE II,分级≥16分时,灵敏度为96.55%,特异度为79.31%,准确率为87.93%,YI=0.758 6,AUC=0.898 9。
MEWS分级是预测ICU患者死亡的有用指标。它简单实用,因此应在临床实践中推荐使用。