Machine Learning Analysis of Risk Factors for Catheter-Associated Urinary Tract Infections in Stroke Patients and Their Impact on Healthcare Quality

Lu Lu,1,2 Xiaojun Li,2,3 Donghao Cai,2,4 Chuangchuang Mei,2,4 Xiaojing Hong,2,4 Jiewen Xie2,5 1Department of Lujingdong Clinic, Guangdong Provincial Second Hospital of Traditional Chinese Medicine, Guangzhou, People’s Republic of China; 2Guangdong Provincial Key Laboratory of Research and Development in Traditional Chinese Medicine, Guangdong Provincial Second Hospital of Traditional Chinese Medicine, Guangzhou, People’s Republic of China; 3Department of Nosocomial Infection, Guangdong Provincial Second Hospital of Traditional Chinese Medicine, Guangzhou, People’s Republic of China; 4Department of Laboratory Medicine, Guangdong Provincial Second Hospital of Traditional Chinese Medicine, Guangzhou, People’s Republic of China; 5Baiyun Apartment Department, Guangdong Provincial Second Hospital of Traditional Chinese Medicine, Guangzhou, People’s Republic of ChinaCorrespondence: Jiewen Xie, Baiyun Apartment Department,Guangdong Provincial Second Hospital of Traditional Chinese Medicine, Guangzhou, People’s Republic of China, Email [email protected]: Catheter-associated urinary tract infections (CAUTIs) represent a common complication among hospitalised stroke patients, exacerbating clinical outcomes and increasing healthcare burdens. This study aims to identify key risk factors for CAUTI in stroke patients using a random forest algorithm, construct and validate a clinical prediction model, and investigate the impact of CAUTIs on healthcare quality.Methods: Data were retrospectively collected from 7486 stroke patients with indwelling urinary catheters at Guangdong Provincial Second Hospital of Traditional Chinese Medicine between January 2020 and December 2024. Patients were categorised into infection and non-infection groups based on CAUTI occurrence. Healthcare quality indicators including length of stay and hospitalisation costs were compared between groups. Random forest algorithm screening identified variables influencing infection, with a decision tree model constructed. Model performance was assessed using area under Receiver Operating Characteristiccurve (AUC), calibration curve, and decision curve analysis (DCA).Results: Among 7486 patients, 248 developed CAUTI. Hospitalisation duration and costs were significantly higher in the infection group. Random forest identified three potential predictors: age, days of catheter use (DaysCAU), and time from admission to catheter insertion (TTCAU). Multivariate logistic regression confirmed ageand DaysCAU as independent risk factors for CAUTI. The constructed nomogram model demonstrated an AUC of 0.794 in the training cohort and 0.778 in the validation cohort. Calibration curves indicated good agreement between predicted and actual values, while DCA confirmed significant clinical net benefit within the 5– 20% risk interval.Conclusion: Secondary CAUTI in stroke patients significantly prolongs hospital stays and increases healthcare costs. The predictive model based on age and catheterisation duration demonstrates favourable discriminatory performance and generalisability. It provides a practical tool for clinicians to identify high-risk patients early and formulate individualised prevention strategies, holding significant implications for enhancing healthcare quality.Keywords: stroke, catheter-associated urinary tract infection, nomogram, risk prediction, healthcare quality