OBJECTIVES: Fentanyl has come to dominate the U.S. illicit opioid supply. We aimed to characterize and examine correlates of preferences for fentanyl vs. other opioids among individuals starting OUD treatment. METHODS: We interviewed 250 adults initiating buprenorphine treatment with positive fentanyl toxicology at intake. We characterized opioid preferences and examined bivariate associations between opioid preference (preference for heroin, fentanyl, heroin-fentanyl mix, or other opioid) and sociodemographic characteristics, psychosocial factors, and substance use behaviors. We then used multinomial logistic regression to examine factors independently associated with fentanyl preferences. RESULTS: Over half (52.0 %) of participants preferred fentanyl (21.2 % fentanyl alone, 30.8 % heroin-fentanyl mix). In bivariate comparisons, participants who preferred fentanyl were a higher acuity group with respect to risks and problems in general. In the multinomial logistic regression, people who preferred fentanyl, either alone or mixed with heroin, used non-prescribed buprenorphine less in the 30 days preceding treatment entry compared to people who preferred heroin or other opioids (RRRalone= 0.88 [0.78, 0.99]; P = 0.037 and RRRmixed= 0.91 [0.84, 0.99]; P = 0.046). People who preferred fentanyl alone were also younger (RRR= 0.93 [0.90, 0.97]; P < 0.001) and more likely to have severe mental illness (RRR= 2.5 [1.1, 5.6]; P = 0.027) than people who prefer heroin or other opioids. CONCLUSIONS: Many people with OUD report preferring fentanyl. People who express preference for fentanyl differ substantively from those with other opioid preferences, and may be at elevated risk for poor health outcomes. Understanding preferences surrounding fentanyl could inform treatment and harm reduction interventions.
