Analysis of risk factors for hospital-acquired infections in emergency intensive care unit patients

Wang Xingyan; Zhou Lisen; Chen Shengkuan; Liang Zhengtong; Peng Huaidong; Li Yanling

doi:10.3969/j.issn.1002-1949.2025.09.005

Abstract：Objective To explore the influencing factors of nosocomial infection in emergency intensive care unit(EICU)patients and to establish a prediction model of nosocomial infection.Methods The retrospective study was conducted to analyze the patients admitted to the EICU of the Second Affiliated Hospital of Guangzhou Medical University from January 2018 to December 2021,including 78 patients with nosocomial infection.218 patients without nosocomial infection were matched in the proportion of 1∶3.The independent risk factors of nosocomial infection in EICU patients were analyzed by multivariate Logistic regression,and the prediction model of nosocomial infection in EICU patients was constructed.Results Multivariate Logistic regression analysis showed that age(P=0.011,OR=1.026,95％CI 1.006-1.047),endotracheal intubation(P＜0.001,OR=5.739,95％CI 2.771-11.887),the combined use of antibiotics(P＜0.001,OR=3.127,95％CI 1.649-5.930)were independent risk factors for nosocomial infection in EICU patients.The final prediction model was P=0.026 × A+1.747 × B+1.140 × C-4.730(A was age;when there was endotracheal intubation,B was assigned as 1,otherwise it was 0;When antibiotics were combined,the value of C was 1,otherwise it was 0).When the predictive model value was＞0,the risk of nosocomial infection increased.The AUC of the model was 0.787,the sensitivity was 0.910,the specificity was 0.586,the optimal cutoff value was 0.19,and the prediction accuracy was 74.80％.Conclusions Age,the combined use of antibiotics and endotracheal intubation are independent risk factors for nosocomial infection in EICU of our hospital.The prediction model of EICU nosocomial infection constructed based on above-mentioned indexes has good prediction effect.