Sixty Patients, No Invoice

Every American conversation about psychedelic rollout starts from a private-clinic premise and then argues about how to subsidize access to it. Read the commentary on scheduling, on reimbursement, on the coming approval, and the delivery setting is never the thing being debated: it is the floor everyone is standing on.

A trial in England skipped that premise. Not by arguing against it. By running the thing somewhere else and reporting what happened.

Rucker, Mantingh, Kerr-Gaffney and colleagues ran a psilocybin trial for treatment-resistant depression inside England's public health service, at a single site associated with King's College London and the South London and Maudsley NHS Foundation Trust. Sixty adults were randomized one-to-one, double-blind, against placebo. Each received a single 25 mg dose with preparation, dosing support, and integration. Follow-up ran for six weeks, and 59 of the 60 completed the depression rating at every follow-up point.

Now the part that determines how the rest of this post is allowed to read. This was a feasibility trial. Its title says randomized controlled trial and its abstract says feasibility trial, and both are accurate, but only one of them tells you what the study was powered to answer. Its primary outcomes were recruitment, retention, and estimating the variance of the depression measure. It was not designed to establish whether psilocybin works for treatment-resistant depression, and the authors do not claim it did. They conclude that the findings support a future confirmatory trial.

It did report a between-group difference on the Montgomery-Åsberg Depression Rating Scale at week three, adjusted, of −10.41 points, with a 95% confidence interval from −14.86 to −5.95, favoring psilocybin and sustained at week six. Non-serious adverse events in the randomized phase numbered 164 in the psilocybin arm and 123 in the placebo arm. Four serious adverse events occurred across the whole trial, one on placebo and three on psilocybin, and where psilocybin had been received, each was judged unrelated to treatment.

One more thing the paper reports about itself, and it matters more than any of the above. The blind did not hold. Every participant who received psilocybin correctly guessed their allocation, as did seven in ten of those on placebo, among the twenty-three who were asked. The authors draw the conclusion themselves: expectancy effects in both groups likely account for some of the between-group difference. A trial describing its own main limitation in plain language is the reason the effect size cannot carry the weight people will put on it.

If you have read anything else on this blog, you know where this goes, and I am going to be consistent about it even though the numbers here point the way I would like them to.

A feasibility trial is powered to answer feasibility questions. When you calculate an effect size from one, you are calculating it from a sample chosen for a different purpose, at a single site, with a confidence interval wide enough to tell you the estimate is unstable. The effect size in this paper is large. It is also a by-product, and treating it as the finding is the specific error I spend a lot of time asking clinicians not to make. The methods framework for reading psychedelic trials is the same one that applies here. The honest reading is that the trial recruited and retained a treatment-resistant sample at one NHS site, that almost nobody dropped out of the outcome measurement, and that a properly powered trial is now justified. What such a trial has to do is recruit enough patients across enough sites to stabilize the estimate. Until one does, the honest description of this evidence does not change.

The adverse-event counts deserve the same discipline. The pattern is roughly what you would expect from an active drug with acute perceptual effects. What it cannot tell you is anything about rare harms: sixty people cannot characterize an event that occurs in one patient in five hundred, and the people most likely to be screened out of a feasibility protocol are often the people a clinician worries about first. Read the safety line as absence of evidence rather than evidence of absence.

One more precision, because this is the overclaim the post is most likely to be accused of. A trial conducted at an NHS trust site is not an NHS treatment service. Nobody was handed psilocybin therapy as routine care. What ran inside the public system was research, with research funding and a research protocol. The distinction matters, and collapsing it would make the political point louder and false.

So here is the narrower claim, which survives all of that.

The dominant assumption in the US is that this work arrives as a high-cost private product, and that the policy problem is how to get poor people access to a private product. That assumption is doing an enormous amount of work and it is almost never argued for. It is inherited from where the capital is.

Something delivered inside a public health system, even as research, is evidence that the assumption is a choice rather than a property of the treatment. Preparation, supported dosing, and integration are labor-intensive. They are not intrinsically boutique. The reason they arrive as a boutique product in the US is that the entities building the delivery infrastructure are the entities that need to recoup investment from it.

This post's companion on the coming FDA decision makes the point that approval creates a drug, not a treatment system, and that is the gap this trial sits in. The molecule is being pushed through a regulatory pathway. The delivery model is being decided by default, by whoever is paying to build it, while the conversation is busy with scheduling.

I want to be precise about what I am not saying. Not that a public system delivers this better, cheaper, or more ethically. England's public health service is itself rationed and under-resourced, and a feasibility trial at one site is not a commissioning decision. Not anything, either, about the price of anyone's training or anyone's course, including mine. The claim is only this: the private-clinic premise is a premise, and there is now a published description of this work happening somewhere that does not share it. That is worth knowing when you next read a piece treating the delivery model as settled.

Practically, for the clinicians who read this blog:

• Cite it as feasibility. If you bring this paper to a case consultation or a journal club, the sentence is a feasibility trial supporting a confirmatory trial, not a trial showing psilocybin works. Your colleague who reads it afterward will check.
• Watch for the effect size traveling alone. This number will be quoted without its design. When you see it, ask what the trial was powered for.
• Do not use it to answer a safety question. A client asking whether psilocybin is safe for someone like them is asking about rare harms in a specific body. A 60-person feasibility trial is the wrong instrument for that, and saying so is a better answer than a number.
• Separate the molecule question from the delivery question when a client or a colleague asks what approval will mean. They are different questions with different answers, and conflating them is how the delivery model gets decided without discussion.
• Notice what your own answer assumes. When a client asks where they could access this, listen to whether your answer is organized around what is clinically indicated or around what they can pay for. That is not a judgment about you. It is information about the system you are practicing inside, and it is the same information the access-equity question raised by the right-to-try order turns on.

Your clients are not waiting for a delivery model to be settled. A structured place to start when someone tells you what they did over the weekend.

No, and its authors do not claim it did. It was a feasibility trial whose primary outcomes were recruitment, retention, and estimating the variance of the depression measure. It reported a large between-group difference at week three, but a feasibility trial is not powered to establish efficacy. The authors conclude the findings support a future confirmatory trial.

Sixty adults with treatment-resistant depression, randomized one-to-one against placebo at a single site, double-blind, with a single 25 mg dose plus preparation, dosing support, and integration. Follow-up ran for six weeks, and 59 of 60 completed the depression rating at all follow-up points.

No. This was research conducted at an NHS trust site, with research funding and a research protocol. It is not a commissioned treatment service, and no patient received psilocybin therapy as routine care. The distinction between research delivered inside a public system and treatment offered by one is the whole difference.

Because preparation, supported dosing, and integration are labor-intensive but not intrinsically boutique, and the assumption that they arrive as a high-cost private product shapes what clinicians and clients treat as possible. A published account of this work running inside a public system shows that the delivery model is a choice rather than a property of the treatment.

The trial recorded 164 non-serious adverse events in the psilocybin arm and 123 in the placebo arm during the randomized phase. Note that this is a count of events rather than of patients, and that a six-week window does not describe longer-term effects. Psychosis history is an exclusion criterion in psilocybin trials generally.

Peter H. Addy, PhD, LPC, LMHC is a Portland-based licensed therapist and the founder of Psychedelic Affirming Education, an NBCC-approved continuing education provider for licensed mental health professionals and Oregon Psilocybin Services facilitators. His research background includes work at Yale School of Medicine on psychedelic substances. PAE's accuracy and editorial standards describe how these posts are sourced.