Study: Predictive model improves nosocomial infection risk assessment after colon cancer surgery

Editor's Note

A newly developed predictive model offers healthcare professionals a dynamic tool to assess the risk of nosocomial infections (NIs) in patients following colon cancer surgery, potentially improving early intervention strategies. Published February 27 in Frontiers in Oncology, the study introduces a nomogram—a statistical model that visualizes key risk factors—to enhance infection prevention efforts.

Researchers analyzed data from 1,146 patients who underwent colon cancer surgery at a tertiary hospital affiliated with Shandong University between 2020 and 2022. Patients were divided into training and validation sets to build and test the model. The study used LASSO regression to identify significant predictors, followed by logistic regression to construct the final risk prediction tool. Model accuracy was assessed using receiver operating characteristic (ROC) curves, calibration curves, decision curve analysis (DCA), and clinical impact curves (CIC).

Among the study cohort, 110 patients (9.60%) developed NIs, with the most common types being:

• Lower respiratory tract infections (34.55%)
• Surgical site infections (30.91%)
• Multiple-site infections (24.55%)
• Abdominal, urinary tract, ascites, and bloodstream infections (10%)

The model identified six independent predictors of nosocomial infections:

• Peak temperature on the second postoperative day
• Braden score on the first postoperative day (lower scores correlated with higher infection risk)
• Retention of abdominal drains for ≥10 days
• American Society of Anesthesiologists (ASA) class ≥ III
• Surgical type (open vs. laparoscopic)
• Postoperative complications (e.g., anastomotic leakage, respiratory failure)

Results demonstrated strong predictive performance: the model achieved an area under the curve (AUC) of 0.881 in the training set and 0.813 in the validation set, indicating good discrimination. Calibration and decision curve analyses further supported its clinical utility. The study also introduced a dynamic online version of the nomogram, allowing clinicians to input patient-specific data to calculate real-time infection risk.

The findings reinforce the need for vigilant postoperative monitoring and early intervention in high-risk patients. The study highlights the role of surgical technique, early temperature monitoring, and minimizing prolonged abdominal drain use in reducing infection rates. While the model requires external validation, it represents a promising tool for improving perioperative infection control and patient safety.