老年脓毒症患者继发慢性危重症风险预测模型构建与验证

Construction and validation of a model for predicting progression of sepsis to chronic critical illness in elderly patients

目的构建并验证老年脓毒症患者继发慢性危重症的风险预测模型,为医护人员早期识别并干预提供依据。方法将来自MIMIC-Ⅳ数据库的5527例老年脓毒症患者作为训练组,采用Lasso回归和logistic回归分析确定慢性危重症的影响因素并构建列线图风险预测模型。分别使用受试者工作特征曲线(ROC)、决策曲线分析和校准曲线对模型的区分度、临床实用性和校准度进行评估。回顾性收集合肥市1所三甲医院的134例老年脓毒症患者作为外部验证组进行验证。结果年龄,入ICU 24 h内的呼吸频率、体温、红细胞比容、红细胞分布宽度、血尿素氮、乳酸、部分凝血活酶时间、有无机械通气和使用抗生素是老年脓毒症患者继发慢性危重症的影响因素(均P<0.05)。基于以上因素构建的列线图风险预测模型在训练组和外部验证组的ROC曲线下面积分别为0.733(95%CI:0.722~0.755)、0.817(95%CI:0.746~0.887)。决策曲线分析显示,当阈值概率在8%~99%时,模型有较好的临床效益;校准曲线显示,预测概率与实际概率接近,校准度较好。结论构建的10个因素老年脓毒症患者慢性危重症风险预测列线图模型有一定的预测价值,可作为医护人员识别高危患者的工具。

Objective To construct and validate a risk prediction model for chronic critical illness after sepsis in elderly patients,and to provide reference for early identification and intervention by healthcare professionals.Methods A total of 5,527 elderly sepsis patients from the MIMIC-Ⅳdatabase were used as the training cohort.Lasso regression and logistic regression were employed to identify the influencing factors of chronic critical illness and construct nomogram prediction model.The model′s discrimination,clinical applicability,and calibration were evaluated using receiver operating characteristic(ROC)curve,decision curve analysis,and calibration curve.Data of 134 elderly sepsis patients from a tertiary hospital in Hefei were retrospectively collected for external validation.Results Age,respiratory rate within 24 hours of ICU admission,body temperature,hematocrit,red blood cell distribution width,blood urea nitrogen,blood lactate,activated partial thromboplastin time,mechanical ventilation,and antibiotic use were identified as influencing factors for the development of chronic critical illness in elderly sepsis patients(all P<0.05).The nomogram prediction model constructed based on these factors achieved an area under the ROC curve of 0.733(95%CI:0.722-0.755)in the training cohort and 0.817(95%CI:0.746-0.887)in the external validation cohort.Decision curve analysis indicated that the model had good clinical applicability when the threshold probability was between 8%and 99%.The calibration curve showed that the predicted probabilities of the model were close to the actual probabilities in both the training cohort and the external validation cohort.Conclusion The 10-factor nomogram risk prediction model is valuable for identifying elderly sepsis patients at risk of developing chronic critical illness.