How Psychedelic Clinical Trials Are Designed

The legal starting point for psychedelic clinical trials

Forty-three peer-reviewed psilocybin trials were listed on ClinicalTrials.gov as of early 2024. That number keeps growing. But before any of these psychedelic clinical trials enrolled a single participant, every research team had to clear a legal hurdle that most science communicators gloss over. I'd rather not gloss over it, because the law here is specific and strict.

In Australia, psilocybin is classified as a Schedule 9 Prohibited Substance under the Therapeutic Goods Administration's Poisons Standard. Possession, supply, or use outside a very narrowly authorised framework is a criminal offence. Full stop. The 2023 rescheduling created a limited pathway: from 1 July 2023, psilocybin may be prescribed under Schedule 8 (Controlled Drug) status, but only by a TGA-approved Authorised Prescriber operating within an approved clinical context. Researchers working outside that framework, or anyone sourcing or using these substances without that authority, remain in serious legal jeopardy.

That's not a caveat buried in fine print. It shapes every methodological decision a trial team makes, from who can administer a session to how the study material is stored. If you want the plain-language background first, our explainers on psilocybin and its legal status in Australia and on how the Poisons Standard (SUSMP) works cover the scheduling system this article assumes.

With that grounding in place, let's look at how the science actually gets done.

Why psychedelic trials can't be designed like standard drug trials

The randomised, double-blind, placebo-controlled trial is the gold standard in pharmacology. It works beautifully when neither the participant nor the researcher knows who got the active drug. Psychedelics break that model almost completely. If the mechanics of that standard design are unfamiliar, randomised controlled trials explained is a useful companion, and there is a general overview of the randomised controlled trial on Wikipedia.

Psilocybin produces profound alterations in perception, time sense, and emotional processing, effects so distinct that most participants know within minutes whether they received an active dose or a placebo. Researchers call this the "unblinding problem," and it's one of the most discussed methodological challenges in the field. Why does it matter? Because if participants know they received the active compound, their expectations can colour what they report, and the questionnaires used as outcome measures are exactly the kind of data expectation can shift. A 2021 paper by Szigeti and colleagues in eLife proposed a "self-blinding" microdose design specifically to try to work around this, though it attracted significant debate about whether it truly controlled for expectancy effects.

Active placebos, substances that produce some perceptible effect without the primary mechanism of interest, are one partial solution. Niacin (vitamin B3) at doses that cause flushing has been used in some trials. Low doses of the same psychedelic compound have been used in others. Neither is perfect. A flushing sensation is not remotely like an altered state of consciousness, and a low dose of the real compound can blur the contrast between study arms in a different way.

The problem is genuinely unsolved, and I'd argue that some published trials have been too confident in their blinding claims given the evidence available. A reader should look for whether the authors measured blinding success at all, for instance by asking participants and facilitators to guess their allocation, rather than simply asserting that the trial was blinded. For a fuller treatment, see placebo challenges in psychedelic research.

Set, setting, and the intervention beyond the molecule

Here is where psychedelic trial design diverges most sharply from conventional pharmacology: the non-pharmacological context is formally treated as part of the experimental intervention.

"Set" refers to the participant's mindset, their expectations, psychological history, preparation, and intentions. "Setting" refers to the physical and interpersonal environment during the session. In trials, both are standardised and documented as rigorously as the drug itself. Preparation sessions (typically two to four meetings with trained facilitators), a structured session room design (usually a comfortable couch, eye shades, curated music), and post-session integration meetings are all specified in the protocol and described in published methods sections.

Standardisation matters because variation in the room, the facilitators or the preparation would otherwise be an uncontrolled variable. A trial that let each site improvise would struggle to say anything reliable about the results. The same logic explains why facilitator training and session monitoring are written down in advance rather than left to individual judgement. The supervised model is described in more detail in psychedelic-assisted therapy and the supervised clinical model.

This matters enormously for interpreting results. You can't attribute outcomes purely to the molecule when the protocol also delivered structured psychological support and human contact over several weeks. Some researchers are direct about this, they describe the overall package as a "compound intervention." Others present the pharmacology as primary and the support as context. That distinction has real implications for how regulatory bodies like the TGA evaluate the evidence.

Understanding the pharmacokinetics of psilocybin, meaning how the compound is absorbed, distributed, and eliminated, is also foundational to session design. Psilocybin is a prodrug rapidly converted to psilocin after oral administration; the active phase typically spans four to six hours based on published pharmacokinetic data, which directly informs how long monitored sessions run.

Dose reporting, safety monitoring and ethics review

Doses used in published clinical trials, and I'm describing what appears in peer-reviewed literature, not providing instructions, typically range from low perceptual doses to higher doses in trials seeking more pronounced subjective effects. These are reported in milligrams per kilogram of body weight or as flat milligram doses depending on the study. In Australia, any administration of psilocybin outside the Authorised Prescriber framework, regardless of dose, remains prohibited under Schedule 9.

Data safety monitoring boards (DSMBs) are standard in registered trials. They are independent of the research team, review safety data at pre-specified intervals, and have authority to halt a study. That independence is the point: the people running the trial have an understandable interest in seeing it finish, so someone without that interest watches the safety signals.

Trials must also pass ethics committee review. In Australia, that means a Human Research Ethics Committee (HREC) registered under the National Health and Medical Research Council framework, which operates alongside the TGA's Special Access Scheme and Authorised Prescriber pathways for any clinical use. An ethics committee looks at consent, risk, participant welfare and whether the design can answer its question at all. Readers wanting the regulator's side can start with the Therapeutic Goods Administration explained.

Participant screening and what the safety data can tell you

Participant screening is unusually extensive in psychedelic trials. Personal or family history of psychosis, certain cardiac conditions, and current use of serotonergic medications (because psilocin acts primarily at 5-HT2A receptors) are standard exclusion criteria. The reasoning is straightforward: a trial should not enrol people for whom the risk profile is least understood. A deeper look at these exclusions is in contraindications discussed in clinical psychedelic research.

The published safety data from Phase 2 trials, including work from Imperial College London and Johns Hopkins, report adverse events, predominantly transient anxiety and psychological distress during sessions, but serious adverse events have been rare in screened populations. Those two qualifiers, screened and monitored, carry the weight of the sentence. They describe a carefully selected group in a controlled room with trained staff present.

That safety profile in controlled research settings should not be extrapolated to unsupervised use, where neither screening nor monitoring is present. It is also worth remembering that early-phase trials are small by design. They are built to flag problems and generate questions, which is why the phases of a clinical trial exist as a sequence rather than a single step.

Outcome measures and what they actually capture

The primary endpoints in psychedelic trials are almost always patient-reported: questionnaires like the QIDS-SR for depression, the CAPS-5 for PTSD symptom severity, or scales measuring psychological flexibility and well-being. This is partly practical (subjective experience is what the intervention most plausibly affects) and partly because objective biomarkers for many of the conditions studied remain elusive.

Patient-reported measures have a known weakness, and it loops straight back to blinding. When participants can tell which arm they are in, self-rated scores are more open to expectation. Some trials add clinician-rated scales to balance this, though the clinician may also know the allocation.

Secondary outcomes have started to include neuroimaging data, and fMRI studies showing changes in default mode network connectivity have attracted considerable media attention, as have biomarkers like BDNF. But the field is still sorting out which of these correlate meaningfully with clinical outcomes and which are artefacts of the acute drug state. The short version is: outcome measurement in this area is still maturing, and effect sizes from small Phase 2 trials should be interpreted with appropriate caution before Phase 3 replication data exists.

The Imperial College 2022 psilocybin-versus-escitalopram trial report is a good example of how to write up this kind of work. The authors flagged the limitations around blinding, sample size, and generalisation carefully. That kind of transparency is exactly what you want to see, and not all published work in this space has been equally candid.

The Australian regulatory and research landscape

Australia's 2023 Authorised Prescriber rescheduling made it one of the first countries to create a formal clinical pathway for psilocybin. But this is a prescribing pathway, not a blanket research deregulation. Academic trials still require TGA approval, HREC clearance, Authorised Prescriber status for the administering clinician, and compliance with Schedule 8 controlled drug handling requirements, including secure storage, dispensing records, and reporting obligations consistent with how other Schedule 8 controlled drugs are managed. The prescribing side is covered in the Authorised Prescriber pathway for psilocybin.

Institutions undertaking this research must also meet Good Manufacturing Practice standards for any compounded or supplied psilocybin, and the substance must meet quality specifications consistent with clinical-grade materials. This is a high compliance bar, and it's appropriate. The legal, ethical, and safety infrastructure around these trials exists precisely because the substances involved carry real risk outside controlled settings.

For anyone curious about the research, and there's genuinely interesting science here, the right place to follow it is through published trial registries, peer-reviewed journals, and TGA public communications. The Australian Drug Foundation also publishes accessible, law-grounded overviews of psychedelic regulation that are worth reading alongside the primary literature.

The science of how these trials are built is fascinating in its own right. The field is wrestling with real methodological problems, among them blinding, the compound-intervention question, and small sample sizes, that don't have easy answers. That's not a reason to dismiss the research. It's a reason to read it carefully.

Sources

• TGA approves psilocybin and MDMA-assisted psychotherapy, Therapeutic Goods Administration
• Psilocybin drug facts; Australian Drug Foundation
• Trial of Psilocybin versus Escitalopram for Depression (Carhart-Harris et al., 2021), NCBI/PubMed
• Self-blinding citizen science to study microdosing psychedelics (Szigeti et al., 2021), eLife

Daniel Okoro, Psychedelic science & law writer