基于Logistic回归分析构建ICU脓毒症多器官功能障碍综合征风险预测模型

A risk prediction model of multiple organ dysfunction syndrome in ICU sepsis based on Logistic regression analysis

目的:基于Logistic回归分析构建重症监护室(ICU)脓毒症多器官功能障碍综合征(MODS)风险预测模型,并进行效果检验。方法:采用便利抽样法选取本院2020年1月—2022年1月ICU收治的脓毒症病人220例,根据有无出现MODS分为对照组和MODS组,采用单因素分析、多因素Logistic回归分析筛选ICU脓毒症病人出现MODS的独立危险因素,据此拟合风险预测模型的回归方程,检验模型预测效果。结果:ICU脓毒症病人发生MODS 38例。多因素分析结果显示,合并慢性病、qSOFA评分>2分、APACHEⅡ评分升高、降钙素原升高、肌酐升高、肿瘤坏死因子-α升高等均为ICU脓毒症病人MODS的独立危险因素(P<0.05)。根据独立预测因子拟合预测模型回归方程,模型Hosmer-Lemeshow检验显示,χ^(2)=4.561,P=0.683,提示预测结果与实际情况相符,模型受试者工作特征曲线(ROC)曲线分析显示,ROC曲线下面积(AUC)为0.805[95%CI(0.729,0.877)],最佳风险截断值为0.347,最大约登指数为0.673,对应灵敏度、特异度分别为0.823、0.850,预测准确率为85.45%。结论:本研究在ICU脓毒症病人发生MODS的危险因素基础上,建立的风险预测模型具有良好拟合程度和区分能力,且准确度较高,能为预测ICU脓毒症病人发生MODS风险提供有效工具。

Objective:An ICU sepsis risk prediction model for multiple organ dysfunction syndrome(MODS)was constructed based on Logistic regression analysis and the effect was tested.Methods:Using the convenience sampling method to select 220 septic patients admitted to ICU from January 2020 to January 2022,who were divided into control group and MODS group according to the presence of MODS,univariate and multivariate Logistic regression analysis was used to screen independent risk factors for multiple organ dysfunction syndrome,from which the regression equation of the risk prediction model was fitted to test the effect of the model prediction.Results:38 cases with MODS.According to multivariate analysis,combined chronic diseases,qSOFA score>2,elevated APACHEⅡscore,PCT,creatinine and TNF-αwere all independent risk factors for MODS in ICU sepsis patients(P<0.05).According to the independent predictor fitting prediction model regression equation,model Hosmer-Lemeshow test showed thatχ2=4.561,P=0.683>0.05,suggesting that the prediction results were consistent with the actual situation,model ROC curve analysis showed that AUC was 0.805(95%CI 0.729-0.877),the optimal risk cut-off of 0.347,the maximum Yoden index of 0.673,corresponding sensitivity and specificity of 0.823 and 0.850 respectively,and the prediction accuracy of 85.45%.Conclusion:Based on the risk factors of MODS in ICU sepsis patients,the risk prediction model has good fit and differentiation ability,and high accuracy,which can provide an effective tool to predict the risk of MODS dysfunction syndrome in ICU sepsis patients.