INTRODUCTION: We aimed to compare the effectiveness of methadone and sublingual buprenorphine/naloxone and their association with treatment discontinuation using real-world data from Ontario, Canada. METHODS: We conducted a longitudinal retrospective cohort study utilizing linked, population-level administrative databases in Ontario. We included data on all Ontario residents with an indication of opioid use disorder (OUD) between January 1, 2014, and a maximum follow-up of December 31, 2022 (n = 45,230). We defined exposure groups as methadone and buprenorphine, and the primary outcome was treatment discontinuation. We applied propensity score matching was applied to compare treatment discontinuation outcomes between patients initiating methadone or buprenorphine/naloxone, controlling for demographics and clinical measures. We incorporated various databases, including the Narcotics Monitoring System, Discharge Abstract Database, and the National Ambulatory Care Reporting System, and analyzed repeated treatment windows to evaluate healthcare utilization and treatment patterns across Ontario's diverse geographic regions. RESULTS: From 2014 to 2022, 45,230 people with OUD contributed case or control windows. During the first 30 days of treatment, buprenorphine was associated with lower discontinuation compared with matched methadone controls (61% vs. 57.7% retained at day 30). However, beyond 60 days, methadone demonstrated lower discontinuation, with Cox proportional hazards model indicating a lower risk of treatment discontinuation. We also observed lower discontinuation time among individuals in rural and remote Northern areas. Cox proportional hazards models confirm time-varying effects, with unadjusted HR 0.95 0.94-0.96 and fully adjusted HRs ranging from 1.03 1.02-1.04 to 1.09 1.05-1.12, and rural/remote Northern areas show longer time to discontinuation. CONCLUSION: Our study highlights the importance of considering both medication type and geographic location when developing strategies to improve treatment retention for individuals with opioid use disorder.
Morin et al. (Mon,) studied this question.