Risk factors for hospital-acquired infection after gynecological surgery: development and internal validation of a nomogram

Hospital-acquired infection (HAI) after gynecological surgery is a common complication that worsens outcomes, while clinically applicable tools for evaluating infection risk remain limited. This study aimed to identify factors associated with postoperative HAI and develop a nomogram to support clinical risk stratification. We conducted a retrospective case-control study including 3004 patients who underwent gynecological surgery at our hospital between January and December 2025. Seventy-six patients with postoperative HAI were assigned to the infection group, and 228 non-infected patients were randomly selected as controls at a 1:3 ratio. Univariable and multivariable logistic regression analyses were performed to identify factors associated with postoperative HAI, and a nomogram was subsequently developed and internally validated using bootstrap resampling. Model performance was assessed using the Hosmer-Lemeshow test, calibration curve, receiver operating characteristic (ROC) curve, and decision curve analysis (DCA). The incidence of postoperative HAI was 2.53% (76/3004), with surgical site infection (SSI) being the most common type (56 episodes, 1.86%). Multivariable analysis identified central venous catheter (CVC) dwell time, operative duration, and perioperative transfusion as independent risk factors, while preoperative length of stay and urinary catheter dwell time showed a trend toward increased risk. The nomogram showed good discrimination, with an area under the ROC curve (AUC) of 0.904, and acceptable apparent calibration within the derivation sample (mean absolute error = 0.057; Hosmer-Lemeshow χ² = 14.804, P = 0.063). DCA suggested potential clinical utility across threshold probabilities of 0.04–0.99. The nomogram developed for HAI risk stratification after gynecological surgery showed good discrimination and may help identify patients at higher risk and support targeted surveillance.