Oregon’s psilocybin services program has operated since 2023 with almost no published data on what happens to the people who use it. That changed on August 19, when JAMA Network Open published the first peer-reviewed safety and outcomes study of a state-licensed psilocybin program anywhere in the country. The timing is pointed: the data lands in the middle of an Oregon Health Authority rulemaking process considering fee increases that operators say could shut parts of the program down, the first real evidence of the model’s legitimacy arriving while its financial survival is in question.
What the study measured
Led by Todd Korthuis at Oregon Health and Science University, the cohort study tracked 346 clients who completed a single psilocybin session at 24 of Oregon’s 26 licensed service centers, supervised by 83 licensed facilitators, enrolled between November 2024 and March 2026 with follow-up through June. This is not clinical trial data. Participants were not screened, dosed, or monitored under an FDA Investigational New Drug application; they were ordinary clients of a regulated, non-medical service model operating entirely outside the drug-development pathway this desk has spent the year tracking. The study was federally funded, through SAMHSA, NIH, and NIDA, itself a notable detail given the federal government funded research into a program serving a federally Schedule I substance.
What it found
At one month, 92.2 percent of the 321 participants with follow-up data reported benefiting from their session, and 58.6 percent ranked it among the ten most meaningful experiences of their lives. At three months, participants reported real decreases in moderate to severe symptoms of depression, anxiety, and PTSD, alongside broader improvements in wellbeing, with 90 percent of the original cohort still reporting at that checkpoint. On safety, four participants, roughly one percent of the cohort, experienced serious adverse behavioral reactions, too few to establish reliable predictors of risk; the researchers flagged inexperience with psychedelics, greater psychological vulnerability, and higher doses as possible contributors. Korthuis, who has said he began the project skeptical of dosing people with psilocybin outside a tightly controlled medical setting, described the results as having changed his own view: “The data has swayed me. These results are very reassuring.”
Why this data type matters, independent of the specific numbers
This desk has spent considerable time this year reading psychedelic efficacy data carefully, distinguishing trial-level effect sizes from what a drug does once it leaves the controlled conditions that produced them. A retrospective study of nineteen patients treated under Switzerland’s medical-exemption framework, examined in this desk’s earlier coverage, found real-world psilocybin outcomes landing closer to ordinary antidepressant performance than to the dramatic figures early trials reported. Oregon’s data tells a different story, benefit rates and symptom reductions that hold up closer to trial-level expectations, generated under a model with essentially no clinical screening, no IND oversight, and facilitators who are not physicians. Read together, these two real-world datasets, from two entirely different regulatory frameworks, are the closest thing available right now to a natural experiment in whether medical-model oversight is what determines psilocybin’s real-world performance, or whether something else, session structure, facilitator training, participant self-selection, matters more than the presence of a physician in the room.
The regulatory tension this data lands inside
Marijuana Moment’s reporting on this study’s release surfaced the detail worth sitting with most: Oregon regulators are currently weighing steep increases to the fees licensed service centers and facilitators pay to operate, changes industry participants warn could force closures across the program while it produces its first strong evidence of safety and benefit. More than 20,000 people have gone through Oregon’s program since it launched, per reporting on the study’s release, and this cohort of 346 is the first rigorously tracked slice of that population. A program now generating exactly the kind of evidence base that legislators in Colorado, California, and other states considering their own frameworks have been waiting for is at the same time facing a domestic fee structure that could shrink or eliminate the service centers generating that evidence in the first place.
What Korthuis himself flagged as limits
The study’s own lead author was direct about who this model is not for: psilocybin services under this framework are not appropriate for people with psychosis, bipolar disorder, or those who are pregnant, a screening boundary worth stating plainly given how much public conversation around state-legal psilocybin access tends to treat it as broadly appropriate for anyone seeking the experience. Korthuis was also careful to frame the result honestly rather than as a blanket endorsement: “There are no magic bullets, and the same is true of psilocybin.”
The caveats
This is a single cohort study without a control or comparison group, meaning it cannot establish that psilocybin itself, rather than the broader experience of preparation, facilitated dosing, and integration support, produced the reported benefits. Self-selection is a real limitation: people who seek out and pay for a licensed psilocybin session are not a representative cross-section of the population, and 90 percent three-month retention, while strong, still means one in ten participants was not tracked to the final checkpoint. Four serious adverse events in 346 participants is too small a number to support firm conclusions about which specific risk factors predict harm, a limitation the study’s own authors state directly.
The frame
This is the first time a state-licensed, non-medical psilocybin program has produced peer-reviewed evidence about what happens to the people using it, and the data reads as considerably more reassuring than skeptics of the non-medical service model, this desk included, might have expected going in. That should matter for how Colorado and other states finish designing their own frameworks. Whether it matters enough to change Oregon’s own fee structure before the program that generated this evidence shrinks under its own regulatory weight is a separate, and considerably less certain, question.