High-dose (≥24 mg) buprenorphine daily doses (BDD) may be important in treating patients with opioid use disorder (OUD) to improve retention and prevent overdose, particularly in the context of increased illicit fentanyl use. This study sought to: (1) identify trajectories for average BDD among patients initiating buprenorphine treatment for OUD and (2) assess patient characteristics associated with these identified trajectories. Buprenorphine treatment episodes among patients in the US Veterans Healthcare Administration (VHA) from federal fiscal years 2006 to 2020 were identified. Group-based trajectory modeling (GBTM) was used to identify BDD trajectories based on weekly averages of BDD over the 180 days after buprenorphine episode initiation. A total of 79 303 buprenorphine treatment episodes among 44 583 patients were included in the analytic sample. GBTM identified 9 latent trajectories for BDD: (1) moderate dose, early discontinuation (10.1%), (2) moderate dose, delayed discontinuation (4.5%), (3) moderate dose, moderate-paced discontinuation (5.2%), (4) low-moderate dose, delayed discontinuation (7.0%), and (5) low-moderate dose, early discontinuation (21.1%), (6) low dose retention (9.6%), (7) low-moderate dose retention (16.7%), (8) moderate dose retention (18.6%), and (9) high dose retention (7.4%). Patient BDD can broadly be characterized as low [2-4 mg/day], low-moderate (6-8 mg/day), moderate (12-18 mg/day), and high dose (≥ 24 mg/day). Patients with episodes in the high BDD trajectory have the lowest social risk (eg, lowest rate of past-year history of homelessness) and the lowest diagnosed rate of physical and mental health-related comorbidities compared to those following other trajectories. BDD ranges widely and patient characteristics are significantly different between those episodes following differing BDD trajectories. Future research on the association between BDD and subsequent patient outcomes (eg, overdose) needs to carefully consider these differences in baseline characteristics.
Read full abstract