A study of people who use syringe service programs in New York City has produced an unexpected finding about one of the most persistent obstacles in opioid addiction care: stigma. Researchers expected people who reported greater discrimination and social rejection because of drug use to be less likely to remain connected to buprenorphine treatment. Instead, overall stigma scores did not predict whether participants had an active buprenorphine prescription one month after enrolling. The participants who were most likely to be engaged in treatment were those who said they had previously been denied medical care because they used drugs. The result does not suggest that stigma is harmless. On the contrary, the study found that stigma was widespread, with most participants reporting experiences of discrimination, alienation or being viewed as dangerous and unreliable. The findings instead point to a more complicated relationship between past mistreatment, healthcare-seeking behavior and access to low-threshold addiction treatment.
Buprenorphine is a medication used to treat opioid use disorder, a chronic condition involving compulsive opioid use despite harmful consequences. As a partial opioid agonist, buprenorphine activates the brain’s mu-opioid receptors but produces a more limited effect than drugs such as heroin or fentanyl. This helps suppress withdrawal symptoms and cravings while reducing the risk of respiratory depression associated with full agonists. When taken as prescribed, it can lower illicit opioid use, overdose risk and opioid-related illness. Yet many people who could benefit from the medication never begin treatment, or stop before establishing long-term care. Barriers include cost, insurance, a shortage of trained prescribers and strict clinic policies. Stigma adds another layer: patients may fear being judged, while clinicians may mistakenly regard medication treatment as replacing one addiction with another. Such attitudes can discourage people from asking for help and can shape the way they are treated once they enter the healthcare system.
The new analysis, published in Addiction Science & Clinical Practice, drew on data from a randomized controlled trial conducted at three syringe service programs between 2020 and 2024. Syringe service programs provide sterile equipment, overdose-prevention resources, testing, referrals and other services to people who inject drugs. They are designed around harm reduction, an approach that aims to reduce injury and disease without requiring abstinence as a condition of care. The sites served predominantly low-income Black and Latinx communities in neighborhoods with some of New York City’s highest opioid overdose mortality rates. The researchers analyzed 97 adults who met the criteria for opioid use disorder under the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, expressed interest in buprenorphine and agreed to participate in the parent trial. Participants had a mean age of 42; 78.4 percent were men, 53.6 percent were Hispanic, 29.9 percent were non-Hispanic Black and 16.5 percent were non-Hispanic White.
The parent trial compared two ways of starting medication. Some participants received an expedited, two-week “bridging” intervention at the syringe service program. A trained clinician evaluated them, prescribed buprenorphine and met with them twice during the initial period before referring them to a federally qualified health center for continuing treatment. Others were referred directly to that health center. The secondary analysis did not find a significant difference in 30-day engagement between the two randomized groups, so assignment was not included as a statistical adjustment. Instead, the investigators examined whether participants’ reported experiences of stigma were associated with treatment engagement regardless of how they began care. Engagement had a precise operational definition: a participant needed to have an active prescription expected to cover at least one day during days 27 through 33 after enrollment. Because the onsite intervention supplied medication for only two weeks, the 30-day measure was intended to capture whether participants successfully linked to longer-term treatment.
At enrollment, participants completed the Ahern substance-use stigma scale, a 15-item instrument measuring several dimensions of stigma. The researchers focused on discrimination, alienation and perceived devaluation when calculating a total perceived-stigma score, while also examining coping responses separately. Discrimination included experiences such as being treated unfairly because of drug use. Alienation measured social exclusion and feelings of being cut off from others. Perceived devaluation captured beliefs that people who use drugs are considered dangerous, untrustworthy or inferior. Each item was answered yes or no, allowing the researchers to sum the number of experiences endorsed. A total of 10 items contributed to the main summary score, with higher scores indicating more stigma. The analysis also isolated one especially relevant question: whether participants had ever been prevented from obtaining medical care because they used drugs. Researchers adjusted their logistic regression models for recent injection drug use, unstable housing, mental-health comorbidity and previous buprenorphine treatment.
The prevalence of stigma was striking. Of the 97 participants, 78.3 percent endorsed at least one item in the discrimination domain. Among the 96 participants with complete responses for the other domains, 85.4 percent reported at least one alienation experience and 84.4 percent reported at least one perceived-devaluation experience. The median total score was seven endorsed items, with an interquartile range of 4.5 to 9. In other words, the typical participant reported stigma across much of the scale rather than describing a single isolated incident. Many said they felt they had to prove themselves because of drug use, were looked down on or believed that society viewed people who use drugs as dangerous and unreliable. Stigma was also embedded in personal relationships: rejection by friends or family was common. Reports of being denied housing or medical care were less frequent than other forms of stigma, but that may reflect the narrower circumstances those questions measured rather than an absence of institutional discrimination.
Contrary to the researchers’ hypothesis, the total stigma score was not significantly associated with having an active buprenorphine prescription at 30 days. Neither were the individual discrimination, alienation, perceived-devaluation or coping domains. The statistical signal emerged only for the question about medical-care denial. After adjustment for the prespecified covariates, participants who reported having been prevented from obtaining medical care because of drug use had 2.64 times the odds of being engaged in buprenorphine treatment compared with those who did not report that experience. The 95 percent confidence interval ranged from 1.05 to 6.66, indicating a statistically significant association but also considerable uncertainty around its magnitude. Because the study was cross-sectional in its stigma assessment and observational in this secondary analysis, the result cannot show that discrimination caused people to pursue buprenorphine, or that the treatment model directly reversed the effects of discrimination.
One explanation is that the question identified people who had more contact with healthcare in the first place. Individuals with frequent or urgent medical needs have more opportunities to encounter discrimination, but they may also be more practiced at navigating clinics, referrals and prescriptions. Another possibility is that the participants represented a particularly resilient group: people who had been turned away yet continued seeking care until they found a service willing to help. The researchers also note that the stigma scale did not distinguish clearly between experienced discrimination, anticipated mistreatment and internalized stigma. Someone who expects poor treatment may avoid care, whereas someone who has actually experienced it may become more determined to locate an accessible provider. The low-threshold model, based at a familiar harm-reduction setting and built around flexible, patient-centered policies, might have been especially attractive to those who had previously been rejected elsewhere. But the study did not measure these pathways directly, and qualitative interviews from the parent trial will be needed to explore why participants did or did not continue treatment.
The findings carry an important warning against interpreting the null association as evidence that stigma does not matter. Nearly every participant entered the study carrying a substantial burden of negative social experiences, yet a short-term treatment measure may not capture how stigma affects long-term retention, disclosure, quality of care or willingness to return after a relapse. The sample was small and drawn from one urban area, limiting its generalizability. The study did not examine how race, gender or other identities interacted with substance-use stigma, even though participants were predominantly people of color and discrimination can operate through several overlapping systems. Follow-up for this analysis ended at 30 days, and the researchers did not collect detailed information about prior treatment attempts or the reasons participants discontinued care. Fentanyl was also widespread in New York City’s illicit opioid supply during the study period, but its potential influence on treatment engagement and stigma perceptions could not be assessed. Even so, the central message is clear: healthcare systems should not wait for stronger evidence that stigma blocks every stage of treatment before addressing it. Training clinicians in harm reduction, correcting misconceptions about medications and using respectful, person-first language could make addiction care more welcoming. Low-threshold buprenorphine programs may help people connect with treatment after conventional services have failed them, but affirming care should be a universal standard rather than a specialized exception.
