基于循证理念构建ICU患者气管插管非计划拔管风险预警模型研究  

Derivation and External Validation of An Evidence-Based Risk Warning Model for Unplanned Endotracheal Extubation in Critically Ill Patients

作　　者：卜晓兰 李振刚 张亚荣 祁进芳 袁媛[1] Bu Xiaolan;Li Zhengang;Zhang Yarong(Department of Critical Care,the First Affiliated Hospital of Xinjiang Medical University,Urumqi,Xinjiang 830092;Xinjiang Medical University,Urumqi,Xinjiang 830092,China)

机构地区：[1]新疆医科大学第一附属医院,新疆乌鲁木齐830092 [2]新疆医科大学,新疆乌鲁木齐830092 [3]新疆医科大学第六附属医院,新疆乌鲁木齐830092

出　　处：《四川医学》2025年第4期427-433,共7页Sichuan Medical Journal

摘　　要：目的建立并验证重症监护病房(ICU)患者气管插管非计划拔管(UEE)风险预警模型。方法在9个中、英文数据库中检索UEE影响因素的相关文献。对文献进行系统评价确定UEE影响因素及其效应值,将效应值转换为回归系数β构建Logistic风险预警模型。通过分析风险预警模型对2023年1月至12月入住我院ICU的312例患者UEE发生风险的预测效能和临床应用价值对模型进行外部验证。结果纳入13篇文献共报告34个可分析影响因素。提取其中合并后效应值具有统计学意义且文献报告次数≥3次的4个影响因素,构建风险预警模型Logit(P)=0.0594+2.5233×夜班+1.8703×RASS评分+0.0100×心率+1.2238×气管插管留置时间。风险预警模型预测312例ICU患者UEE发生风险的受试者工作特征曲线下面积(AUC)、敏感度、特异度分别为0.839(95%CI 0.702~0.976,P<0.001)、0.909、0.694;模型预测概率与实际概率拟合效果良好(Hosmer-Lemeshow检验:X^(2)=8.947,P=0.347);决策曲线显示阈值概率为20%~44%时临床收益较高。结论基于循证理念的风险预警模型具有较好的预测效能和应用价值,可用于评估ICU患者UEE风险。Objective To develop and validate a risk warning model for unplanned endotracheal extubation(UEE)in intensive care unit(ICU)patients based on evidence-based concepts.Methods We searched nine Chinese and English databases for studies related to the risk factors of UEE.A Meta-analysis was conducted to determine the risk factors of UEE.These effect sizes were then converted into regression coefficients(β)to construct a logistic risk warning model.The model was externally validated using data from 312 patients admitted to the ICU of the First Affiliated Hospital of Xinjiang Medical University from January to December 2023,to evaluate its predictive performance and clinical application value.Results A total of 13 studies reporting 34 risk factors were included in the meta-analysis.Four risk factors with statistically significant pooled effect sizes and reported in at least three studies were selected to build the risk warning model:Logit(P)=0.0594+2.5233×night shift+1.8703×richmond agitation-sedation scale(RASS)score+0.0100×heart rate+1.2238×endotracheal tube dwell time.The area under the receiver operating characteristic curve(AUC)of the risk warning model was 0.839(95%CI 0.702~0.976,P<0.001),and the sensitivity and specificity were 0.909 and 0.694,respectively.The predicted probability of the model calibrated well with the actual probability(Hosmer-Lemeshow test:X^(2)=8.947,P=0.347).The decision curve showed that the model provided clinical benefit when the threshold probability was 20%~44%.Conclusion The evidence-based risk warning model demonstrated good predictive performance and application value for assessing the risk of UEE in ICU patients.

关 键 词：非计划拔管 气管插管 风险预警模型 重症监护 

分 类 号：R47[医药卫生—护理学]