Fewer than one in three adults with treatment-resistant depression ever receives a trial of a rapid-acting intervention, and most of those who do have already exhausted years of failed antidepressants. That number has sat in the back of my mind across more than 30,000 ketamine treatments at Mindbloom, and it sharpens every time the regulatory landscape shifts. This week it shifted again: the September 10 issue of the New England Journal of Medicine published the FDA’s new framework for psychedelic drugs, a short document with long consequences for how trials are designed, how patients will eventually access these therapies, and where the competitive landscape lands after years of regulatory turbulence.
The framework arrives in a field still absorbing the aftershocks of August 2024, when the FDA issued a Complete Response Letter to Lykos Therapeutics rejecting MDMA-assisted therapy for PTSD after a 10-to-1 advisory committee vote against approval. That vote was not primarily a verdict on MDMA’s therapeutic signal. It was a verdict on study design: inadequate functional unblinding controls, concerns about trial integrity, and the agency’s discomfort with the functional unblinding that accompanies an experiential therapy. The NEJM framework reads, in part, like the FDA’s answer to that mess. It signals that psychedelics will require purpose-built evidentiary standards, not retrofitted antidepressant protocols.
The Framework’s Structural Bet
What the FDA appears to be constructing is a tiered oversight architecture that separates the pharmacological effect of a compound from the psychotherapeutic context in which it is delivered. That distinction sounds procedural. Clinically, it is foundational. When I think about what produces durable remission in the patients who respond to ketamine at Mindbloom, the pharmacology is necessary but not sufficient. The preparatory session, the integration work afterward, the therapeutic relationship during the medicine session: these are not decoration. They are mechanism. Any regulatory framework that treats the drug and the therapy as separable components will generate trial designs that underperform in real-world populations, because it will optimize for the compound in isolation.
The framework also signals a REMS-forward posture for future psychedelic approvals, consistent with the January 2025 expansion of Spravato's monotherapy indication under a mandatory risk management program. For MDMA and psilocybin, a REMS almost certainly means certified treatment centers, trained therapist dyads, and supervised administration windows. Those requirements will concentrate access in urban academic settings for years before any broader rollout, which means the patients I see in underserved communities, many of them Spanish-speaking, many carrying dual diagnoses, will be last in line unless trial sponsors build enrollment infrastructure that explicitly compensates for that gravitational pull toward the well-resourced and the well-connected.
COMPASS Pathways is among the furthest along in the psilocybin space. Their COMP360 Phase 3 program has produced 52-week open-label data from the COMP005 trial, with 258 participants enrolled in the United States across 25 mg and placebo arms. What that program will now have to navigate is the FDA's expectation that psychedelic trial designs account for functional unblinding in their statistical models, not dismiss it. That is a harder problem than it sounds when the functional unblinding problem is acute in trials of this kind.
What This Means for Dual-Diagnosis Populations
Here is the counterintuitive read on the new framework that the field tends to resist: tighter regulatory structure could actually accelerate access for the most complex patients if it forces sponsors to stop excluding them. Every psychedelic Phase 2 trial I have reviewed in the past four years has carried extensive comorbidity exclusion criteria. Active suicidal ideation out. Current substance use disorder out. Bipolar spectrum disorder out. These are, almost without exception, the patients sitting across from me in clinic. The FDA's framework, by demanding more rigorous safety monitoring architecture, gives sponsors the evidentiary scaffolding they need to run carefully monitored studies in patients who have previously been treated as too risky to enroll rather than too important to exclude.
In my practice, the patients who have shown the most striking responses to ketamine are often the ones with the most complicated histories: treatment-resistant depression layered on alcohol use disorder, or PTSD driving the substance use that every prior clinician tried to treat in isolation. The neuropharmacology here is not mysterious. Ketamine's glutamatergic mechanism, MDMA's oxytocin and serotonin surge, psilocybin's 5-HT2A agonism: each of these acts on circuits that addiction and chronic stress have reorganized. Treating the reorganization requires accessing those circuits directly. Excluding dual-diagnosis patients from psychedelic trials is not a safety decision. It is a scientific one that makes the resulting data less generalizable precisely where the disease burden is highest.
The AtaiBeckley acquisition by Eli Lilly, approved by stockholders on September 8, 2026 and expected to close September 11, brings BPL-003, VLS-01, and EMP-01 under Lilly's development umbrella three days after this framework publishes. That timing is not coincidental. Lilly now inherits the benefit of FDA's new clarity and the obligation to meet it across a portfolio of novel psychedelic compounds at various trial phases.
The Design Failure Nobody Wants to Name
The FDA's framework makes explicit what trial designers have been quietly absorbing for two years: the standard placebo-controlled model does not translate cleanly to experiential therapies. Functional unblinding rates in psilocybin trials routinely exceed 80%. Designing around that requires active placebo conditions, expectancy measurement at baseline, and statistical models that treat the therapeutic alliance as a moderating variable rather than noise. Most current protocols treat it as noise.
Sponsors who read the NEJM framework as primarily a compliance document will build REMS programs and stop. Sponsors who read it as a trial design signal will restructure their endpoints, their population definitions, and their therapist training protocols before the next IND hits the agency's desk. The difference will appear in FDA review letters two to four years from now, but the choices get made in the next six months.
I am watching the DEA's parallel psilocybin rescheduling petition closely. The DEA has advanced Dr. Sunil Aggarwal's petition to move psilocybin from Schedule I to Schedule II, now under Department of Justice review. If that rescheduling moves forward while the FDA's new framework is still being operationalized, the regulatory and scheduling timelines will intersect in ways the field is not yet prepared to manage. That intersection is what I will be tracking in the next Psych Pulse.
References
- New England Journal of Medicine, "FDA's New Framework for Psychedelic Drugs," Vol. 395, Issue 10, September 10, 2026
- First Word Pharma, FDA Complete Response Letter to Lykos Therapeutics, MDMA-Assisted Therapy for PTSD, August 9, 2024
- Headlight Health, FDA Approves Spravato as First Monotherapy for Treatment-Resistant Depression, January 21, 2025
- Psychiatric Times, 52-Week Topline Open-Label Data from COMPASS Pathways COMP005 Phase 3 Trial
- Investing.com, AtaiBeckley 2026 Mental Health Treatment Pipeline Milestones; Eli Lilly Merger
- Reason.org, DEA Considers Rescheduling Psilocybin, Psychedelics Policy Newsletter
Dr. Leonardo Vando, MD is Principal Investigator of Psychiatry at Vando Medical Services.
