急诊重症监护室患者分流安全性评估模型的构建与实证检验

Construction and empirical test of shunt safety evaluation model for patients in emergency intensive care unit

目的探究影响急诊重症监护室(emergency intensive care unit,EICU)患者分流安全性的影响因素,构建分流安全性评估模型,并评估其预测效能,为EICU安全分流提供理论依据。方法收集并分析2019年8月1日0时0分-2021年5月31日23时59分由四川大学华西医院EICU就诊后分流至普通病房患者的人口学资料、生命体征、实验室检查等指标,以转出EICU后短期预后不良为终点事件。随机选取患者中70%作为模型构建队列,30%为模型验证队列。在模型构建队列中采用多因素logistic回归分析筛选分流安全性的影响因素,并构建EICU患者的分流安全性评估模型。在验证队列中,采用受试者操作特征曲线分析该模型对EICU患者分流安全性的评估效能。结果共纳入患者582例,其中短期预后不良者59例(10.1%)。多因素logistic回归分析显示,患者离开EICU时的呼吸频率[比值比(odds ratio,OR)=0.863,95%置信区间(confidence interval,CI)(0.794,0.938),P=0.001]、格拉斯哥昏迷量表得分[OR=1.575,95%CI(1.348,1.841),P<0.001]、白蛋白[OR=1.137,95%CI(1.008,1.282),P=0.036]、凝血酶原时间[OR=0.956,95%CI(0.914,1.000),P=0.048]是分流安全性的影响因素。基于上述指标创建EICU患者的分流安全性评估模型,该分流安全性评估模型预测短期预后不良的曲线下面积为0.815,最佳截断值为4分,灵敏度为93.3%,特异度为61.5%。结论患者离开EICU时的呼吸频率、格拉斯哥昏迷量表得分、白蛋白及凝血酶原时间是影响EICU分流安全性的因素,基于其构建的分流安全性评估模型能较好地预测由EICU分流至普通病房患者中短期预后不良的发生。

Objective To explore factors affecting the shunt safety of patients in emergency intensive care unit(EICU), construct a shunt safety evaluation model, and evaluate its prediction effectiveness, so as to provide a theoretical basis for the decision-making of shunt safety in EICU. Methods The demographic data, vital signs, laboratory examinations and other indicators of patients transferred to the general ward from the EICU of West China Hospital of Sichuan University from 0:00 on August 1, 2019 to 23:59 on May 31, 2021 were collected and analyzed. The short-term poor prognosis after being transferred out of the EICU was regarded as the end-point event. Of the patients, 70% were randomly selected as the model construction cohort, and 30% were the model validation cohort. In the model construction cohort, multivariate logistic regression analysis was used to screen the influencing factors affecting shunt safety, and the shunt safety evaluation model of patients in EICU was constructed. In the validation cohort, receiver operating characteristic curve was used to evaluate the effectiveness of the model in evaluating the shunt safety of patients in EICU. Results A total of 582 patients were included, of whom 59 patients(10.1%) had a poor shortterm prognosis. Multivariate logistic regression analysis showed that the patients’ respiratory rate when leaving the EICU [odds ratio(OR)=0.863, 95% confidence interval(CI)(0.794, 0.938), P=0.001], Glasgow Coma Scale scores[OR=1.575, 95%CI(1.348, 1.841), P<0.001], albumin [OR=1.137, 95%CI(1.008, 1.282), P=0.036], prothrombin time[OR=0.956, 95%CI(0.914, 1.000), P=0.048] were the influencing factors of shunt safety. Based on the above indicators,a shunt safety evaluation model for patients in EICU was created. The area under the curve for the shunt safety assessment model to predict poor short-term prognosis was 0.815, the best cut-off value was 4 points, the sensitivity was 93.3%, and the specificity was 61.5%. Conclusions The patients’ respiratory rate when leaving EICU, Glasgow Coma Scale scores,albumin and prothrombin time are factors affecting the shunt safety for patients in EICU. The shunt safety assessment model can better predict the short-term poor prognosis of patients transferred from EICU to general ward.