Below is a summary of “Improving Surgical Site Infection Prediction After Colorectal Surgery Using Machine Learning” published in the March 2023 issue. Hematology by Chen,other.
The main purpose of this investigation was to examine the feasibility of using machine learning to improve existing predictive models for C. diff. Surgical site infection. A cohort of colorectal surgery patients from 2012 to 2019 was extracted from the American College of Surgeons National Quality Improvement Program database and stratified into training, validation, and test sets. Random forests, gradient boosting, and artificial neural networks are all examples of machine learning approaches. Additionally, a logistic regression model was built. Finally, the area under the receiver operating characteristic curve was used to assess model quality.
Surgical site infections of all depths and locations within the organ were included in the primary outcome. Due to the implementation of inclusion and exclusion criteria, the dataset contained information on 275,152 patients. A total of 10.7% of patients developed infections near the surgical incision. The area under the receiver operating characteristic curve for the artificial neural network was 0.769 (95% CI, 0.762-0.777), whereas the area for gradient boosting was 0.766 (95% CI, 0.759-0.774) and the random forest was 0.764 ( 95% CI, 0.759-0.774). CI, 0.756-0.772), logistic regression was 0.677 (95% CI, 0.669-0.685).
According to an artificial neural network model, the strongest predictors of infection were the presence of surgical site infection in the organ space at the time of surgery, length of surgery time, bowel preparation with oral antibiotics, and surgical approach. Colorectal surgical site infections are easier to predict using machine learning techniques than logistic regression. By using these methods to identify people at greatest risk, we can identify those most in need of precautions against surgical site infections.
sauce; journals.lww.com/dcrjournal/Abstract/2023/03000/Improved_Prediction_of_Surgical_Site_Infection.19.aspx
