Post written by Hassam Ali, MD, from the Department of Gastroenterology, Brody School of Medicine, East Carolina University, Greenville, North Carolina, USA.

This study aimed to develop a predictive model for 30-day reintervention rates following endoscopic sleeve gastroplasty (ESG) procedures. Using supervised machine learning techniques, the model incorporates factors such as patient demographics, clinical variables, and laboratory data to predict the likelihood of reintervention within a 30-day period post-ESG.

Reintervention after ESG can be costly and risky for patients. By developing a predictive model, healthcare providers can better assess the risks associated with ESG procedures, allowing for more informed decision-making. This could lead to improved patient outcomes and more efficient use of healthcare resources.

Our model successfully predicted 30-day reintervention rates with significant accuracy, as evidenced by various performance metrics. We included 3583 patients in our analysis, and the 30-day endoscopic reintervention rate for ESG cases was 3.3%. This study adds a quantitative, data-driven approach to assessing the risks associated with ESG procedures. Future work should focus on validating the model in different clinical settings and possibly extending it to predict other postoperative outcomes.

I would like to thank my team and the healthcare providers who contributed valuable data for this study. The goal is to continually refine this model to make it a standard tool for risk assessment in ESG procedures.

ROC curves with mean cross-validated area under the curve (cvAUC) after 10-fold-cross-validation. AUC, .74 (bootstrap bias-corrected 95% confidence interval, .70-.78). ROC, Receiver-operating characteristic; SD, standard deviation.

Read the full article online.

The information presented in Endoscopedia reflects the opinions of the authors and does not represent the position of the American Society for Gastrointestinal Endoscopy (ASGE). ASGE expressly disclaims any warranties or guarantees, expressed or implied, and is not liable for damages of any kind in connection with the material, information, or procedures set forth.