Risk factor analysis and prediction modeling of nosocomial infections associated with extracorporeal membrane lung oxygenation
Objective:To explore the risk factors of nosocomial infections associated with patients treated with Extracorporeal Membrane Oxygenation(ECMO)and establish a risk prediction model.Methods:The clinical data of 153 patients who received ECMO treatment from December 2020 to June 2023 in our hospital were retrospectively analyzed.According to the diagnostic criteria of postoperative infection,they were divided into the infected group(71 cases)and the uninfected group(82)cases.Baseline data and ECMO-related parameters were collected and compared between the two groups.Multiple logistic regression analysis was applied to study the independent risk factors for nosocomial infections in ECMO patients and establish a prediction model.Propensity score matching and marginal analysis were applied to assess the economic losses of patients and hospitals caused by nosocomial infections.Results:Hospital-acquired infections occurred in 71(46.41％)of 153 ECMO patients.The results of multifactorial logistic regression analysis showed that Acute Physiology and Chronic Health Score version Ⅱ score,total hospitalization time,ECMO running time,and mechanical ventilation time before ECMO evacuation were risk factors for hospital-acquired infections in ECMO patients(P<0.05).A risk prediction model was established based on the results of logistic regression,and the fit of the model was high by the Hosmer-Leme show test(χ2=4.124,P=0.776).The results of the ROC curve analysis showed that the risk prediction model for ECMO-associated infections predicted that the AUC for the occurrence of nosocomial infections after treatment with ECMO was 0.911(95％CI(0.852,0.970),P<0.001).Conclusion:The prediction model constructed in this study can be used as a reference for developing effective measures to prevent ECMO-related nosocomial infections.
NETL
NSTL
万方数据
