Nomogram Development and Feature Selection Strategy Comparison for Predicting Surgical Site Infection After Lower...

Area of Science:

• Orthopedic Surgery
• Infectious Disease Epidemiology
• Clinical Prediction Modeling

Background:

• Surgical site infections (SSIs) are a significant complication following lower extremity fracture surgery.
• Current tools for predicting individual SSI risk are limited, necessitating improved risk assessment strategies.

Purpose of the Study:

• To develop and internally validate a nomogram for predicting individualized SSI risk.
• To identify key perioperative clinical parameters associated with SSI development in this patient population.

Main Methods:

• Retrospective cohort study of 638 adult patients undergoing lower extremity fracture surgery.
• Logistic regression and bootstrap inclusion frequency were used for feature selection and model development.
• Model performance was evaluated using AUROC, calibration slope, Brier score, sensitivity, and specificity.

Main Results:

• The final model identified seven predictors: red blood cell count, preoperative C-reactive protein, chronic kidney disease, operative time, chronic obstructive pulmonary disease, body mass index, and blood transfusion.
• The nomogram achieved a high predictive performance with an AUROC of 0.924.
• Chronic kidney disease and blood transfusion emerged as the strongest predictors of SSI.

Conclusions:

• The developed nomogram demonstrates robust predictive capabilities for SSI risk.
• This tool has the potential to enhance personalized risk assessment and inform preventative strategies for patients undergoing lower extremity fracture surgery.