Zhang Y, Fatemi P, Medress Z, et al. - Researchers sought to develop models for predicting long-term opioid use in patients after elective spine surgery by using preoperative risk factors and 30-day postoperative opioid prescribing patterns. Inpatient, outpatient, and pharmaceutical data were used from MarketScan databases (Truven Health) for this retrospective cohort study. Among 19,317 assessed patients who were newly diagnosed with low back or lower extremity pain and underwent thoracic or lumbar surgery within 1 year after diagnosis, they identified 903 (4.6%) of patients who met criteria for long-term opioid use. Seven models were generated on 80% of the dataset using demographic variables, medical and psychiatric comorbidities, preoperative opioid use, and 30-day postoperative opioid use, and the models were tested on the remaining 20%. Three regression-based models (full logistic regression, stepwise logistic regression, least absolute shrinkage and selection operator), support vector machine, two tree-based models (random forest, stochastic gradient boosting), and time-varying convolutional neural network were used. Per outcomes, the regression-based models had high sensitivity and AUC; this model is provided online to screen patients for high risk of long-term opioid use based on preoperative risk factors and opioid prescription patterns in the first 30 days after surgery.  Following were the three strongest positive predictors of long-term opioid use: high preoperative opioid use, number of days with active opioid prescription between postoperative days 15 to 30, and number of dosage increases between postoperative day 15 to 30. Number of dosage decreases in the 30-day postoperative period were the strongest negative predictors.