(A-321) Risk of adverse opioid-related outcomes among adult Medicaid enrollees with new long-term prescription opioid use

Background: Both overall and high-risk opioid prescribing have decreased in the U.S. in recent years, potentially contributing to population-level declines in incident opioid use disorder (OUD) and prescription opioid-involved overdose deaths. However, opioid prescribing remains common. Therefore, increasing our understanding of patient and prescribing factors associated with adverse opioid outcomes can help target risk mitigation efforts.

Objectives: To estimate risk of incident adverse opioid outcomes in adult Medicaid enrollees with new long-term prescription opioid episodes.

Methods: Using national Medicaid data from 2015-2019, we identified new long-term prescription opioid episodes, defined as ≥6 months with no fills followed by ≥3 months (90 days) of prescription opioid receipt with no gaps >30 days in supply. Episodes were restricted to those among adults 18-64 years old with no Medicare eligibility or cancer diagnosis at any point in this 9-month period. We analyzed the first qualifying episode per person.
The composite outcome was the first of any incident adverse opioid event including: 1) OUD diagnosis (excluding OUD in remission), 2) medication for OUD (MOUD), or 3) opioid overdose within one year of qualifying for new long-term prescription opioid use (day 91 to 456 follow-up). Enrollees were excluded if any adverse opioid outcome was observed in the 6 months before or first 90 days of the episode. We also measured patient sociodemographic and clinical characteristics, opioid treatment characteristics, community characteristics, and state policies. We calculated crude event rates overall and by covariate levels. Fully adjusted Cox proportional hazards models estimated risk of any event.

Results: Of 476,547 enrollees with new long-term prescription opioid use, 37,730 (7.9%) had an incident adverse opioid outcome in follow-up. Of these, 83.1% were OUD diagnoses, 14.2% were MOUD receipt, and 7.4% were opioid overdoses. Five percent of events (n=1817) included both an OUD diagnosis and another measure, mostly MOUD receipt (n=1429, 78.6%). The overall crude event rate was 9.9 per 100 person-years. The strongest risk factor was prescription opioid dosage, with a gradient from low (HR=1.12) to very high (HR=2.35) mean dosage in the first 90 days. Male sex, younger ages, substance use disorders, hepatitis C, and mental health diagnoses and treatment were also significant risk factors.

Conclusions: Incidence of adverse outcomes after transitioning to long-term prescription opioid use was high and concentrated in patients with known risk factors and high-risk treatment patterns. These findings may help guide clinical practices toward more selective opioid initiation, cautious dose escalation, and careful monitoring, especially early in treatment.