Appendix: Impact of the extended-release/long-acting opioid analgesics risk evaluation and mitigation strategy on prescribing practices - Appendix

Objective: To assess the impact of extended-release (ER)/long-acting (LA) opioid prescriber training on prescribing behaviors.Design: Retrospective cohort study.Setting: Prescriber training was evaluated from June 1, 2013 through December 31, 2016. The full study period was 2 years longer, from June 1, 2012 through December 31, 2017, to include data for all prescribers’ 1-year pretraining and post-training periods.Participants: 24,428 prescribers who wrote ER/LA opioid prescriptions for eligible patients, with a record of training from the partner continuing education provider between June 1, 2013 and December 31, 2016.Intervention: ER/LA opioid prescriber training. Main outcome measures: Prescribing behaviors 1-year before (pretraining) and after (post-training) prescribers completed training, specifically the proportion of opioid-nontolerant patients receiving ER/LA opioids indicated for opioid-tolerant patients and for patients receiving ≥100 morphine equivalents dose daily, and the proportion of concomitant users of central nervous system depressant drugs.Results: The differences in the proportion of opioid-nontolerant patients receiving ER/LA opioids indicated for opioid-tolerant patients and for patients receiving ≥100 morphine equivalents dose daily were –0.69 percent (95 percent confidence interval [CI]: –1.78 percent, 0.40 percent) and –0.23 percent (95 percent CI: –1.18 percent, 0.68 percent), respectively. The differences in the proportion of concomitant users of central nervous system depressant drugs were –0.94 percent (95 percent CI: –1.39 percent; –0.48 percent) for benzodiazepines, 0.06 percent (95 percent CI: –0.13 percent; 0.25 percent) for antipsychotics, –0.41 percent (95 percent CI: –0.69 percent; –0.13 percent) for hypnotics/sedatives, and 0.08 percent (95 percent CI: –0.40 percent; 0.57 percent) for muscle relaxants.Conclusions: While prescribers showed some changes in prescribing behavior after completing training, training was not associated with clinically relevant changes in prescribing behaviors.