神经内科重症监护室多重耐药菌感染预测模型构建

Establishment of prediction model for multidrug-resistant organisms infections in patients of intensive care unit of neurology department

目的分析多重耐药菌(MDR)感染在神经内科重症监护患者的发生情况及病原菌分布,并构建分险预测模型。方法纳入2020年1月-2023年1月连云港市第一人民医院神经内科重症监护室收治的598例患者进行研究,根据检出MDR情况,分为MDR感染组42例、MDR未感染组556例,分析检出的MDR菌株分布及MDR感染的危险因素,构建预测模型并评估其预测效能。结果598例神经内科重症监护患者,共检出42株MDR,MDR感染发生率7.02%;高龄、留置导管、住院时间≥10 d、联合应用广谱抗菌药物为神经内科重症监护患者MDR感染的危险因素(P<0.05);据此构建的风险预测模型经Hosmer-Lemeshow拟合度检验结果显示,模型有较好的拟合优度(χ^(2)=5.785,P=0.671);校准曲线结果显示,预测概率与实际概率接近,提示模型具有良好的区分、校准和预测能力;受试者工作特征(ROC)曲线结果显示,曲线下面积(AUC)值为0.702,95%CI为0.616~0.788。模型诊断敏感度为70.00%,特异度为77.50%,提示模型区分度较好。结论神经内科重症监护患者MDR感染临床发生率较高,引起MDR感染危险因素较多,据筛选的危险因素构建的预测模型可为预防措施的制定提供依据。

OBJECTIVE To observe the prevalence of multidrug-resistant organisms(MDR)infections among the patients of intensive care unit(ICU)of neurology department,analyze the distribution of pathogens and establish the prediction model.METHODS A total of 598 patients who were treated in ICU of neurology department of the First People′s Hospital of Lianyungang City from Jan 2020 to Jan 2023 were enrolled in the study and were divided into the MDR infection group with 42 cases and the no MDR infection group with 556 cases according to the status of MDR infection.The distribution of isolated MDR and risk factors for the MDR infections were analyzed,the prediction model was established,and the predictive efficiency was evaluated.RESULTS Totally 42 strains of MDR were isolated from the 598 ICU patients of neurology department,and the incidence of MDR infection was 7.02%.The old age,catheter indwelling,length of hospital stay no less than 10 days and combined use of broad spectrum antibiotics were the risk factors for the MDR infections in the ICU patients of neurology department(P<0.05).The risk prediction model was established based on the risk factors,the result of Hosmer-Lemeshow test of goodness of fit showed that the model had favorable goodness of fit(χ^(2)=5.785,P=0.671);the result of calibration curve indicated that the predicted probability was close to the actual probability,showing that the model had favorable capability of differentiation,calibration and prediction.The result of receiver operating characteristic(ROC)curve analysis showed that area under curve(AUC)was 0.702,with 95%CI 0.616-0.788.The sensitivity of the model was 70.00%for diagnosis,with the specificity 77.50%,indicating that the model had favorable discrimination degree.CONCLUSION The incidence of MDR infection is high among the ICU patients of neurology department,there are a variety of risk factors for the MDR infections.The prediction model that is established based on the screened risk factors can provide basis for formulating prevention measures.