Abstract
IMPORTANCE: Fentanyl has spread through the illicit opioid supply in Canada, driving increasing overdose deaths. However, the effectiveness of methadone and buprenorphine-naloxone in treating opioid use disorder during the fentanyl era is unknown. OBJECTIVE: To evaluate methadone and buprenorphine-naloxone treatment duration, a core effectiveness outcome in the treatment of opioid use disorder, in Ontario, Canada, between 2014 and 2022. DESIGN, SETTING, AND PARTICIPANTS: This population-based, retrospective cohort study included individuals who initiated methadone or buprenorphine-naloxone between January 2014 and December 2022 in Ontario, Canada. Data were analyzed from July 18, 2023, to June 11, 2025. EXPOSURE: Period of medication initiation (2014-2016, 2017-2019, or 2020-2022). MAIN OUTCOMES AND MEASURES: The main outcome was treatment duration, measured as time to medication discontinuation (5 consecutive days without dispensation of the initial opioid agonist treatment or availability of take-home doses). RESULTS: The cohort included 72 717 new recipients of opioid agonist treatments (45 256 [62.2%] male; median [IQR] age, 35 [28-46] years), with 34 538 individuals (47.5%) receiving methadone and 38 179 individuals (52.5%) receiving buprenorphine-naloxone. Among individuals starting methadone, median treatment duration decreased from 193 (95% CI, 185-202) days in 2014 to 2016 to 139 (95% CI, 130-149) days in 2017 to 2019 and 86 (95% CI, 78-95) days in 2020 to 2022. Among individuals starting buprenorphine-naloxone, median treatment duration decreased from 51 (95% CI, 49-54) days in 2014 to 2016 and 50 (95% CI, 48-53) days in 2017 to 2019 to 38 (95% CI, 36-40) days in 2020 to 2022. In adjusted Cox regression models including time-varying effects and using 2014 to 2016 as the reference period, hazards of discontinuation measured at treatment initiation were higher during later periods of methadone initiation (2017-2019: adjusted hazard ratio [aHR], 1.18 [95% CI, 1.15-1.22]; P < .001; 2020-2022: aHR, 1.45 [95% CI, 1.39-1.51]; P < .001) and for buprenorphine-naloxone initiation in 2020 to 2022 (aHR, 1.11 [95% CI, 1.08-1.15]; P < .001). Age categories, neighborhood income quintile, rurality, sex, and number of comorbidities were also associated with time to discontinuation in adjusted models. CONCLUSIONS AND RELEVANCE: This cohort study found that treatment duration among individuals starting methadone and buprenorphine-naloxone during 2020 to 2022 was lower than during 2014 to 2016. This study highlights the importance of ongoing evaluation of treatment effectiveness, given the dynamic nature of the opioid crisis. Further research is needed to improve treatment retention and improve the effectiveness of opioid use disorder treatment.