改良早期预警评分预测急诊患者死亡风险的前瞻性队列研究

The Value of the Modified Early Warning Scoring in Prediction of the Outcome of Patients in Emergency Department:a Prospective Cohort Study

目的评价急诊患者入院时改良早期预警评分(MEWS)对其48 h死亡风险的预测价值,并尝试增加年龄评分来提高预测能力.方法前瞻性采集2010年9月1日至2011年8月31日在红河州第一人民医院急诊科留院观察患者的年龄、体温、脉搏、呼吸、收缩压、AVPU评分,计算入院时MEWS分值,随访主要终点为48小时内死亡,次要终点为院内全因死亡.用Logistic回归评价各参数与终点相关性,受试者工作特征曲线下面积(AUROCC)和Hosmer-Lemeshow拟合优度检验进行分辨度和校准度的评价.结果 1 837例纳入研究,平均年龄(47.0±20.18)岁,男性1183例,占64.4%,女性654例,占35.6%,死亡399例,占21.7%,其中48 h时内死亡353例,占观察病例的19.2%.Logistic回归显示年龄(OR=1.012,95%可信区间:0.997～1.027)、体温(OR=0.827,95%可信区间:0.618～1.107)、脉搏(OR=1.004,95%可信区间:0.994～1.015)、呼吸(OR=1.045,95%可信区间:1.001～1.091)、收缩压(OR=1.000,95%可信区间:0.995～1.005)、AVPU评分(OR=3.981,95%可信区间:2.921～5.425)与患者48 h内死亡相关.增加年龄评分前后的MEWS在评估急诊患者48 h内死亡风险的AUROCC分别为0.838(95%可信区间:0.802～0.881)和0.841(95%可信区间:0.796～0.881),二者差异没有统计学意义(P=0.9196).Hosmer-Lemeshow拟合优度检验分别为P=0.058和P=0.205.结论MEWS对于评估急诊患者入院后48 h内死亡风险具有中度的分辨度,拟合度较好,评分有效性好,可行性高.增加年龄评分后分辨度的提高没有统计学意义.

Objective To assess the prognostic value of the Modified Early Warning Score（MEWS） in predicting mortality in 48 hours after admission to an Emergency department（ED）,and to improve the predicative value by addition of age scoring.Methods A prospective observational cohort study was performed ont the consecutive adult patients（≥16 yrs）admitted in emergency department in The First People＇s Hospital of Honghe State from September 1,2010 to August 31,2011.The data of patients including age,temperature,puls rate,respiratory rate,systolic pressure,AVPU score were collected and the MEWS were calculated.The primary endpoint was death during first 48 hours in the emergency department,and secondary endpoint was overall death in-hospital.A multivariate logistic regression analysis was performed to identify independent predictors of inhospital mortality.Hosmer-Lemeshow good of fit test and area under receiver operator characteristic curves（AUROCC） were used to illustrate and compare before and after addition of age scoring to MEWS.Results During the observation period,data were collected from 1837 admissions.The mean age of the population was 47.0±20.18.Overall,1183（66.4%）of patients were male,and 654（35.6%）were female.During follow-up,399（21.7%）patients died and 353（19.2%）cases of deaths occurred within 48 hours after hospital admission.Logistic regression showed age（OR=1.012,95% CI： 0.997-1.027,P=0.004）,temperature（OR=0.827,95%CI： 0.618-1.107,P=0.002）,pulse rate（OR=1.004,95% CI：0.994-1.015,P=0.00）,respiratory rate（R=1.045,95% CI： 1.001-1.091,P=0.004）,systolic pressure（OR=1.000,95% CI： 0.995-1.005,P=0.025）and AVPU score（OR=3.981,95% CI： 2.921-5.425,P0.001）had correlation with the death of patients in 48 hours after admission.The AUROCC before and after addition of age scoring to MEWS on first 48 hours in the emergency department were 0.838（95% CI： 0.802-0.881）and 0.841（95% CI： 0.796-0.881）,respectively,and there was no statistically significant difference.Using Hosmer-Lemeshow statistics,goodness of fit test showed P=0.058 and P=0.205 respectively,Conclusions MEWS has moderate ability in predicting the risk of death in emergency during first 48 hours,and has good fitting degree and efficiency.Addition of age scoring dose not significantly improve the predictive value of MEWS scores.