LAUSR

BACKGROUND Opioids play a role in pain management after surgery, but prolonged use contributes to developing opioid use disorder. Identifying patients at risk of prolonged use is critical for deploying interventions that reduce or avoid opioids; however, available predictive models do not incorporate patient-reported data (PRD) and it remains unclear if PRD can predict post-operative use behavior. We used a machine learning (ML) approach leveraging preoperative PRD and electronic health record (EHR) data to predict persistent opioid use after upper extremity (UE) surgery. METHODS Included patients underwent UE surgery, completed preoperative PRD questionnaires, and were prescribed opioids after surgery. We trained models using a 2018 cohort and tested in a 2019 cohort. Opioid use was determined by patient report and filled prescriptions up to 6 months after surgery. We assessed model performance using AUROC, sensitivity, specificity, and Brier score. RESULTS Among 1656 patients, 19% still used opioids at 6 weeks, 11% at 3 months, and 9% at 6 months. The XGBoost model trained on PRD plus EHR data achieved AUROC 0.73 at 6 months. Factors predictive of prolonged opioid use included income, education, tobacco, drug or alcohol abuse, cancer, depression, and race. Protective factors included preoperative PROMIS Global Physical Health and preoperative PROMIS Upper Extremity scores. CONCLUSION This opioid use prediction model using pre-intervention data had good discriminative performance. PRD variables augmented EHR-based ML algorithms in predicting post-surgical use behaviors and were some of the strongest predictors. PRD should be used in future efforts to guide proper opioid stewardship.