Psychedelic Therapies Are Almost Here. The Infrastructure Isn’t.
An average of 17.6 veterans die by suicide every day. That number is why the federal psychedelic policy shift is a clinical operations problem not a regulatory abstraction.
On April 18, an executive order directed the FDA to fast-track review of psychedelic compounds, allocated $50 million for federal-state research collaboration, and instructed the DEA to establish new patient access pathways. Harvard Law’s I. Glenn Cohen noted the order leaves “considerable discretion to the FDA commissioner.” The door is open. The room is not yet furnished.
The clinical signal is real
Psilocybin’s trajectory is the clearest evidence. In June 2025, Compass Pathways announced its Phase 3 trial of COMP360 met its primary endpoint, a statistically significant reduction in depression severity at six weeks. A second pivotal trial confirmed the result in February 2026. It is the first classic psychedelic to produce two positive Phase 3 results in the United States.
The MDMA outcome clarifies what’s at stake. In August 2024, the FDA rejected Lykos Therapeutics’ MDMA-assisted therapy application for PTSD, citing data reliability concerns, trial oversight failures, and site misconduct that led to a journal retraction. Lykos subsequently cut its workforce by 75%. That outcome is a design specification, not a reason to wait.
The operational problem is the actual problem
Psychedelic-assisted therapy doesn’t fit the infrastructure most behavioral health programs are built to run. A single psilocybin treatment arc involves preparation sessions, a monitored dosing session lasting several hours, integration therapy, and longitudinal follow-up. EHR architectures built for episodic encounters aren’t designed for it. The most clinically significant events delay adverse reactions, integration crises may occur days after dosing, outside any existing monitoring window. When oversight fails between clinical touchpoints, the consequences reach the FDA’s desk.
Three workstreams need to begin now.
Care model design. Every touchpoint needs clinical ownership before the first patient is scheduled. Who owns the preparation session? Who is responsible for a delayed adverse event on a Sunday? Oregon and Colorado have legalized psilocybin therapy. Health systems there are encountering these questions without answers.
Workforce and credentialing. Psychedelic-assisted therapy requires therapist-physician teams in a model most institutions haven’t defined and the credentialed pipeline doesn’t yet exist at scale. Training requirements aren’t standardized. Certification bodies are nascent. These frameworks take 12 to 18 months to build. Workforce is not a downstream problem. It is the first constraint.
Between-session monitoring. The clinical window doesn’t close when the patient leaves the building. Monitoring infrastructure that ends at discharge is inadequate for a treatment arc where the most consequential patient experiences may occur days later, at home, without clinical contact.
Two constraints worth naming
Ibogaine carries known cardiotoxicity and has been linked to more than 30 deaths in the medical literature. Frederick Barrett of the Johns Hopkins Center for Psychedelic and Consciousness Research has been explicit that rigorous research is still needed. Ibogaine and psilocybin require separate planning tracks.
Reimbursement for a multi-week, therapist-physician care arc remains unresolved. No major payer has established coverage frameworks. That is not an argument against building. It is an argument for clear-eyed cost accounting and no assumptions about how quickly payer frameworks will follow FDA action.
What this requires in practice
The care coordination model psychedelic-assisted therapy demands longitudinal, multi-disciplinary, continuous between sessions doesn’t exist at scale anywhere in behavioral health today. It requires integrating data across physicians, psychiatrists, psychologists, and therapists around a single patient, across a treatment arc that spans weeks, not appointments.
That is precisely the problem Wonder Sciences has spent four years building toward. Wondermed, its B2B telemedicine platform, treated 1,919 patients across 24 U.S. states, with a UCLA IRB-reviewed study following 431 patients over 46 months generating the longitudinal, real-world clinical data most behavioral health platforms never accumulate. WondermedAI, the company’s AI-native clinical operating system, was built from that operational experience: a system designed from the ground up for coordinated, continuous care rather than episodic encounters. Wonder Sciences is currently partnering with health systems and behavioral health networks on the WondermedAI pilot program and the psychedelic treatment arc is exactly the use case it was architected for.
What to do this quarter
The VA is currently participating in at least five psychedelic drug trials across New York, California, and Oregon. Health systems serving veteran populations are not preparing for a future patient. They are looking at a current one.
Three actions, this quarter: Map clinical ownership gaps, run a full psychedelic treatment arc on paper and identify every handoff with no named owner. Audit your credentialing framework if your institution has no pathway to credential therapist-physician teams for this modality, 12 to 18 months of build time starts now. Define your monitoring perimeter to identify where current clinical oversight ends and what would need to change to extend it through the full post-dosing window.
17.6 veterans a day. The systems that move now will have infrastructure when a treatment option arrives. The systems that wait will be building it after the fact.
Ryan Magnussen is Founder and CEO of Wonder Sciences, which is reengineering mental healthcare through advanced AI and FDA clinical research. Learn more at wondersciences.com