急危重症患者院内转运不良事件风险预测模型的构建

Construction of a risk prediction model for adverse events in intra-hospital transfer of critically ill patients

目的探讨急危重症患者院内转运不良事件的危险因素,并构建风险预测模型。方法采用目的抽样法,选取2023年1月至6月乌鲁木齐市某三级甲等医院急诊科收治的342例急危重症患者为研究对象,根据是否发生院内转运不良事件分为发生组和未发生组。采用Logistic回归模型分析相关影响因素,构建预测模型,绘制可视化列线图。使用ROC曲线下面积及校准曲线验证模型的预测效能。结果建模组239例急危重症患者中有75例发生院内转运不良事件,发生率为31.38%。Logistic回归分析显示,收缩压、改良早期预警评分、供氧装置、监护仪、转运班次、转运总时长为急危重症患者院内转运不良事件发生的独立危险因素(P<0.05)。结果显示,建模组ROC曲线下面积为0.943,灵敏度为0.968,特异度为0.875;验证组ROC曲线下面积为0.922,灵敏度为0.903,特异度为0.875。Hosmer-Lemeshow检验显示χ2=7.348,P=0.403,校正曲线显示该列线图模型的预测概率和实际概率具有较好的一致性。结论急危重症患者院内转运不良事件发生率较高,收缩压、改良早期预警评分、供氧装置、监护仪、转运班次、转运总时长是其发生院内转运不良事件的影响因素,基于以上影响因素构建的风险预测模型具有良好风险识别能力,列线图模型具有较好的区分度和校准度。

Objective To explore the risk factors for adverse events in intra-hospital transport of critically ill patients and construct a risk prediction model.Methods A total of 342 patients with acute and critical illnesses in the emergency department of a tertiary hospital in Urumqi City from January to June 2023 were selected for the study by purposive sampling,and they were divided into the occurrence group and the non-occurrence group according to whether they had an adverse event of intra-hospital transport.The Logistic regression model was used to analyze the related risk factors,construct the prediction model,and draw a visual nomogram.The predictive efficacy of the model was verified by using the area under the ROC curve and the calibration curve.Results Seventy-five out of 239 patients with acute and critical illnesses in the modeling group experienced intra-hospital transport adverse events,with an incidence rate of 31.38%.Logistic regression analysis showed that systolic blood pressure,Modified Early Warning Score(MEWS),oxygen supply device,monitor,transfer shift,and total transport duration were independent risk factors for adverse events in intra-hospital transport of critically ill patients(P<0.05).The results showed that the area under the ROC curve in the modeling group was 0.943,with a sensitivity of 0.968 and specificity of 0.875,and the area under the ROC curve in the validation group was 0.922,with a sensitivity of 0.903 and a specificity of 0.875.Hosmer-Lemeshow test showed 2=7.348,P=0.403,and the calibration curve showed good agreement between the predicted and actual probabilities of this nomogram model.Conclusion The incidence of adverse events of intra-hospital transport is higher in critical patients.Systolic blood pressure,MEWS score,oxygen supply device,monitor,transfer shift,and total transport duration are the influencing factors of adverse events in intra-hospital transport,and the risk prediction model constructed on the basis of the above influencing factors has good risk identification ability,and the nomogram model has a better degree of differentiation and calibration.