Bringing health services directly to people who inject drugs may help reduce deaths, according to results from a National Institutes of Health-supported randomized clinical trial published in JAMA Psychiatry. The study found that participants assigned to integrated care delivered through mobile medical units, together with peer navigation, had lower all-cause mortality than participants offered peer navigation to community-based services.
The finding is important, but it needs careful interpretation. Mortality was a secondary outcome, the study involved 409 adults in five U.S. cities, and the researchers said the result should be replicated. The trial does not show that every mobile-clinic program will produce the same effect, nor does it establish which individual component of the intervention was responsible.
What the trial tested
The HPTN 094 trial enrolled adults who did not have HIV, injected drugs and had opioid use disorder. Participants were randomly assigned to one of two approaches. One group received integrated services through mobile medical units, including clinical care and support from peer navigators. The comparison group received peer navigation that connected them with services available in the community.
The integrated model was designed to reduce practical barriers by bringing several forms of care to locations that were easier for participants to reach. For people facing unstable housing, transportation problems, stigma or fragmented services, the ability to receive care in one setting may make continued engagement more feasible.
A substantial mortality difference, with limits
NIH reported that all-cause mortality was 71% lower in the mobile-unit group than in the comparison group. Because mortality was not the trial’s primary outcome and the total number of participants was modest, this relative difference should not be read as a precise prediction for other populations. Absolute event counts, local service conditions and participant characteristics all matter when judging how transferable the result may be.
The trial did not find a significant difference between groups in the use of daily oral pre-exposure prophylaxis (PrEP) for HIV prevention or medications for opioid use disorder. At 26 and 52 weeks, only 14% of participants were using medication treatment for opioid use disorder. That low uptake underscores how difficult it remains to connect people with evidence-based treatment and sustain care over time.
Why service design may matter
People who inject drugs often face overlapping risks, including overdose, infection, untreated chronic illness and barriers to routine medical care. A mobile service can combine screening, prevention, treatment referral and follow-up in a setting that may feel more accessible than a traditional clinic. Peer navigators can add practical knowledge and trust, helping participants move through complex health and social-service systems.
However, an integrated model is not automatically effective simply because it is mobile. Programs still need qualified staff, reliable medication access, laboratory capacity, privacy protections, links to hospitals and community services, and plans for continuity of care. Local implementation is likely to influence outcomes.
What should happen next
The researchers called for replication of the mortality result. Future studies should examine which parts of the intervention drive benefit, whether similar outcomes occur in other regions and health systems, and how programs can improve uptake of PrEP and medication treatment for opioid use disorder.
For public-health planners, the trial offers evidence that care delivery itself can be an important part of the intervention. Making services easier to reach may influence more than appointment attendance, but the strongest claims will require larger studies and confirmation in different settings.
Evidence note: This article summarizes an NIH-supported randomized clinical trial involving a specific population in five U.S. cities. The mortality result was a secondary outcome and should be interpreted as promising rather than definitive.
Source: National Institutes of Health, November 4, 2026; HPTN 094 trial, JAMA Psychiatry, DOI: 10.1001/jamapsychiatry.2026.3644.
