Related Experiment Video
Updated: Jan 8, 2026
14:56
An Experimental Paradigm for the Prediction of Post-Operative Pain PPOP
Published on: January 27, 2010
21.9K
Explainable machine learning to predict prolonged post-operative opioid use in rotator cuff patients
Jason Guattery1,2, Liane M Miller3, James J Irrgang1,2
1School of Health and Rehabilitation Sciences, University of Pittsburgh, Pittsburgh, PA, United States.
BMC Musculoskeletal Disorders
|December 12, 2025
Summary
Machine learning models accurately predict prolonged post-operative opioid use in rotator cuff surgery patients. These transparent tools aid surgeons in preventing opioid prescriptions for at-risk individuals, addressing a significant healthcare challenge.
Area of Science:
- Orthopedic Surgery
- Data Science
- Machine Learning
Background:
- Opioid overuse is a significant problem in the US, with surgery contributing to prescriptions.
- Lack of clear prescribing guidelines exacerbates the issue.
- Machine learning can help identify patients at risk for prolonged post-operative opioid use.
Purpose of the Study:
- Develop accurate and transparent machine learning models.
- Predict prolonged opioid use after rotator cuff surgery.
Main Methods:
- Trained six machine learning models on 852 rotator cuff surgery patients.
- Evaluated model predictive accuracy.
- Used SHAP and LIME for model transparency and interpretation.
Main Results:
- Four models exceeded 0.71 predictive accuracy.
- Top models achieved 0.98 (XGBoost), 0.94 (Random Forest), and 0.74 (Decision Tree) accuracy.
- SHAP and LIME provided insights into model predictions.
Conclusions:
- Successfully developed predictive machine learning models for prolonged post-operative opioid use.
- Achieved high predictive accuracy.
- Enhanced model explainability and transparency.
