OBJECTIVE To investigate the risk factors for healthcare-associated infections in cardiovascular surgery patients and their predictive models, and to evaluate the predictive models.

METHODS Totally 5 364 patients in the department of cardiovascular surgery of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology from Jun. 2020 to Jun. 2023 were selected for retrospectively analysis, and the patients were randomly divided into a modeling group (4023 cases) and a validation group (1341 cases) according to the ration of 3∶1, the logistic regression analysis was applied to identify the risk factors for healthcare-associated infections in the modeling group, the value of each risk factor was assigned based on the correlation coefficient β, and the risk assessment model of healthcare-associated infections was constructed, and the discrimination, calibration and clinical practicality of the model were evaluated by using the receiver operating characteristic (ROC) curve, the Hosmer-Lemeshow goodness-of-fit test, the calibration curve and the decision curve analysis.

RESULTS In this study, there were 321 cases of healthcare-associated infections among cardiothoracic and vascular surgery patients, with a healthcare-associated infection incidence rate of 5.98%, and the number of hospital-acquired infections cases was 343, with a hospital-acquired infection incidence rate of 6.39%. The results of binary logistic regression analysis showed that the length of hospitalization (OR=2.970, 95%CI: 1.588-5.552, P=0.001), the number of surgical procedures (OR=2.706, 95%CI: 1.757-4.167, P < 0.001), the duration of surgery (OR=2.143, 95%CI: 1.491-3.080, P < 0.001), the application duration of antimicrobial agent (OR=2.433, 95%CI: 1.675-3.543, P < 0.001), the combined application of antimicrobial agents (OR=2.228, 95%CI: 1.403-3.541, P=0.001) and the employment of ventilator (OR=4.095, 95%CI: 2.408-6.964, P < 0.001) were risk factors for healthcare-associated infections in patients undergoing cardiovascular surgery. Values were assigned to each risk factor, the study subjects were divided into low-, medium- and high-risk groups, with the incidence of hospital-acquired infections increasing as the risk level increased. The area under the ROC curve (AUC) was 0.845 (95%CI: 0.825-0.865) in the modeling group and 0.823 (95%CI: 0.784-0.862) in the validation group, suggesting that the established risk assessment model had a good predictive value.

CONCLUSION The risk assessment model for healthcare-associated infections in patients undergoing cardiovascular surgery constructed in this study has a good prediction accuracy, which can be used to identify high-risk patients and facilitate early prevention and intervention, thereby effectively reducing the risk of healthcare-associated infections in patients undergoing cardiovascular surgery.