On July 14, 2026, the Food and Drug Administration published a notice of availability in the Federal Register for the final version of Psychedelic Drugs: Considerations for Clinical Investigations (Federal Register document 2026-14158). The guidance finalizes a draft the agency first issued on June 26, 2023, and it is the agency’s most detailed statement to date on how sponsors should design trials of classic psychedelics — psilocybin, LSD, MDMA, and related compounds — being developed for psychiatric disorders and substance use disorders.
For research administrators, IRB members, and clinical trial teams working on or considering psychedelic-assisted therapy protocols, this is the reference document FDA reviewers will now hold sponsors to. It is nonbinding guidance rather than a regulation — sponsors may propose an alternative approach if it satisfies the underlying statute and regulations — but it represents the agency’s current thinking, and deviating from it without a documented rationale is likely to draw scrutiny at IND review or a pre-IND/Type B meeting.
What the guidance covers
The document addresses the full arc of a psychedelic drug development program: nonclinical and manufacturing considerations, clinical pharmacology, trial design and conduct, abuse potential assessment, and safety monitoring during and after dosing sessions. Two areas stand out as the most consequential for protocol design and IRB review, and are covered in depth below: the problem of functional unblinding, and the agency’s revised, risk-based approach to abuse potential testing.
The functional unblinding problem
Classic psychedelics produce perceptual and subjective effects that are difficult to miss. A participant who receives psilocybin or MDMA in a placebo-controlled trial will, in most cases, know within minutes that they received active drug rather than placebo — and so, often, will the therapist or rater in the room. FDA’s guidance describes this as functional unblinding: even when the randomization code itself stays concealed, the drug’s own effects reveal group assignment, which opens the door to expectation bias in both participants (who may report benefit partly because they expect it) and in raters or clinicians scoring outcomes.
This is not a new concern — it was central to FDA’s 2024 rejection of Lykos Therapeutics’ MDMA-assisted therapy application for PTSD, where reviewers questioned whether functional unblinding had inflated the trial’s efficacy signal. The final guidance is FDA’s first formal, generalizable answer to that problem. It recommends what the agency calls a complementary-trials approach: pairing a conventional randomized, placebo-controlled trial with a separate dose-response trial that compares low, middle, and high doses of the same drug without a placebo arm. Because every participant in the dose-response trial receives some level of active drug, functional unblinding cannot by itself explain a dose-dependent efficacy signal — giving reviewers a second, independent line of evidence that isn’t vulnerable to the same bias as the placebo-controlled arm. The guidance also encourages sponsors to consider active comparators (lower psychedelic doses, or other psychoactive drugs that mimic some subjective effects) as a partial blinding strategy, and to build blinding-integrity and expectancy assessments into the protocol rather than treating unblinding as an unmeasured background risk.
Abuse potential: a risk-based approach
The guidance also revises FDA’s expectations for abuse-potential characterization, a required part of any Controlled Substances Act-scheduled drug’s development program. Rather than mandating the full standard battery of nonclinical studies in every case, the guidance endorses a risk-based approach: sponsors may use computational and pharmacological-class models in place of new self-administration or conditioned-place-preference studies, which the guidance notes typically do not produce a positive abuse signal for classic psychedelics in the first place. Where a drug’s subjective and reinforcing effects are already well characterized from earlier-phase clinical work, a dedicated human abuse potential study may not be scientifically necessary at all. This is a meaningful narrowing of what sponsors have to budget and plan for relative to the 2023 draft and to FDA’s general abuse-potential guidance, though the agency still expects a documented rationale for skipping any specific study rather than a blanket assumption that psychedelics are exempt.
What sponsors, IRBs, and research administrators should do now
- Protocol design: sponsors planning pivotal trials should evaluate whether a complementary dose-response arm strengthens their efficacy case, and should build in prospective measures of blinding integrity and participant/rater expectation rather than addressing unblinding only in the discussion section of the eventual study report.
- IND-enabling work: abuse-liability workplans drafted against the 2023 draft or older FDA abuse-potential guidance should be revisited — the final risk-based framework may reduce the nonclinical studies actually required, and this is worth raising at a pre-IND or Type B meeting before committing budget.
- IRB review: boards reviewing psychedelic-assisted therapy protocols should expect informed consent forms and monitoring plans that explicitly address functional unblinding, acute psychoactive risk during dosing sessions, and the role of trained monitors/facilitators — this guidance gives IRBs a citable reference point when protocols are silent on those issues.
- Safety monitoring: the guidance reinforces that dosing sessions require direct observation and a plan for managing acute psychological distress, consistent with prior FDA thinking on psychedelic trial conduct.
Context: three years from draft to final
FDA’s original draft of this guidance was issued in June 2023, meaning the final version arrives roughly three years later, after an extended stretch of stakeholder comment. In the interim, the field’s highest-profile test case — Lykos Therapeutics’ MDMA-assisted therapy application — was rejected in August 2024, with FDA requesting an additional Phase 3 trial and citing, among other issues, concerns about how functional unblinding may have affected the reliability of the efficacy data. The final guidance’s complementary-trials recommendation reads directly against that backdrop: it is FDA’s attempt to give sponsors a study-design path that produces an efficacy signal reviewers can trust even when blinding cannot be fully preserved.
Coverage of the final guidance from RAPS and Applied Clinical Trials both frame this as a genuine compromise document: it does not resolve the underlying scientific difficulty of testing a drug whose effects are impossible to hide, but it gives sponsors and reviewers a shared, documented framework for handling that difficulty rather than litigating it trial-by-trial.
Related CASRAI resources
- Placebo-Controlled Study Design
- Blinding and Masking in Clinical Trials
- Open-Label vs. Double-Blind Trial Design
- Investigational New Drug (IND)
- Designing a Clinical Trial Protocol: Structure, ICH E6/E8, and Registration
- FDA Expedited Programs: Fast Track, Breakthrough Therapy, Accelerated Approval, and Priority Review Compared
- Randomization Methods in Clinical Trials
- Surrogate Endpoint Validation in Clinical Trials
