Psilocybin may be coming to the Spravato clinic

Nothing is approved yet. But Compass has now said, more or less out loud, that if the FDA says yes, its psilocybin will run on the infrastructure the field built for esketamine. Here is what that launch would probably look like from inside a busy clinic: the hours, the rooms, the people, the paperwork, and the money. With the uncertainties marked.

By Reid Robison, MD. 6,264 words, with a regulatory clock, an hour-by-hour dosing day, a room-and-staff calculator, a discharge checklist and a clinic-readiness list you can keep.

For six years I have watched interventional psychiatry clinics learn to do something that did not exist in 2019: certify a site under a REMS, monitor a dissociative drug in a chair, bill a payer for the chair, and keep the schedule full. In late September, at an investor conference, Compass Pathways laid out its launch plan for COMP360, its synthetic psilocybin for treatment-resistant depression, and the plan, pending approval, is built on exactly that infrastructure.^1,2 Their chief patient officer, Steve Levine, put it in one sentence: the treatment "slots directly into the way these centers operate today."³

He is right, and I would add a "yes, and." Yes, the rooms, the REMS reflexes and the billing rails are there, and the clinics that run them well are the natural first homes for this. And: the day itself is longer, the hands-per-hour are higher, and the parts of the trial that wrapped around the dosing day have no obvious place in the model yet. That second half is where clinics may stumble if they plan only from the first. So this is the piece I would hand a clinic owner, a medical director, or a nurse manager who is being asked, "should we get ready for this?" It is also for the patients and families who may walk into those rooms in 2027, if it is approved, and would like to know what the day is going to look like. I have tried to be dense and to cite everything, and to say plainly where I am guessing.

1The clock, and where it can slip

Here is the sequence Compass described at the TD Cowen summit, with the parts that are inside its control and the parts that are not.^1,2 Hover or tap a step for the detail. The shaded width is my reading of how much each step can stretch; it is a judgment, not a company number.

Four things about that clock are worth knowing. First, the NDA is rolling, and Compass says the initial FDA review is already underway, so the Q4 completion is not the day the agency first sees the data.⁹ Second, the DEA scheduling step is the one people forget. Psilocybin is Schedule I, and an approved product cannot be shipped until the DEA moves it. Compass's chief financial officer said the company has already submitted its eight-factor analysis to the FDA, which is the scientific-and-medical evaluation that feeds the DEA's decision, and that the DEA step "typically takes about 90 days"; he added that an April executive order directing closer coordination between the FDA and the DEA "could potentially shorten" that.^1,10 For esketamine the step was a formality because ketamine was already Schedule III. For psilocybin it is a real decision, and the company's base case is Schedule III, which the CEO, Kabir Nath, said would not differ much in practice from Schedule II.²

Third, the review-speed numbers are a target. The moderator at the summit called the timeline aspirational, and Nath quoted Dr. Maurizio Fava calling it "a target, not a commitment."² Fourth, Nath has said in print that the earliest patients will be handled "with kid gloves," in a "quite contained, carefully managed launch."³ Read that as: even if every date holds, most clinics will not be dosing in the first quarter after approval.

2What the trials showed, briefly

I wrote a long piece on where psilocybin stands, so this is the operational summary. Two phase 3 trials, both double-blind, both in treatment-resistant depression, both with the six-week change in the Montgomery–Åsberg Depression Rating Scale (MADRS) as the primary endpoint.^7,11

• COMP005 (n=258): a single 25 mg dose versus an inert placebo. Difference at week six, −3.6 MADRS points (95% CI −5.7 to −1.5). About a quarter of the 25 mg group had a clinically meaningful response at week six, and that was durable to six months.^9,11
• COMP006 (n=581): two fixed doses three weeks apart, 25 mg versus 10 mg versus a 1 mg active control, with up to four doses allowed over 52 weeks. Difference at week six, 25 mg versus 1 mg, −3.8 (95% CI −5.8 to −1.8). Thirty-nine percent of the 25 mg arm had at least a 25% MADRS reduction at week six, and the benefit was maintained through week 26. Retreatment helped: nearly 30% of the week-six responders went on to remit (MADRS ≤12) after a further dose.^9,11
• Safety: the common adverse events were nausea, headache, anxiety and visual hallucination, mostly on the dosing day, and 83–88% of them resolved within 24 hours. Serious adverse events over 26 weeks were similar across arms (6.3% at 1 mg, 5.7% at 25 mg). Suicidal ideation was under 1%, and there was one suicidal-behavior event, in the 1 mg arm.^9,11

Two things from that matter for a clinic. The effect is real but modest at the group level, and a lot of the value appears to come from identifying the people who respond and retreating them. And the trial support was not thin: participants had preparation sessions before dosing, a supervised dosing day of six to eight hours, and integration sessions afterward, with the support described by Compass as non-directive and manualized rather than psychotherapy.^7,12 Whatever the label ends up saying, the evidence was generated inside that wrapper.

3What carries over from Spravato

If your clinic is a well-run Spravato site, you already have most of the skeleton.

• REMS certification. Compass expects a REMS "comparable to Spravato's," with a licensed prescriber on site as the authorized representative who certifies the setting and attests that staff are trained.^1,2 Esketamine's REMS asks a healthcare setting to certify, enroll each patient, and submit a monitoring form after each session; it is paperwork and process, and if yours is clean, you are ahead.¹³
• Monitoring habits. Spravato sites already watch blood pressure and dissociation for at least two hours, and since the October 2023 label change they monitor with pulse oximetry too, after postmarketing reports of respiratory depression.^14,15 The reflexes transfer: a blood-pressure cuff on a timer, a nurse who knows what sedation looks like, a quiet room with a recliner, a discharge standard, a companion to drive.
• Billing rails. Category III CPT codes for in-office psychedelic administration are already secured, Compass says; Category I codes are roughly 18 months out, and in the meantime sites would bill with the codes they use for Spravato. Compass plans field reimbursement specialists, as Janssen did.²
• Controlled-substance handling. Spravato is Schedule III. Compass's base case is Schedule III.² A site that already has a DEA registration covering Schedule III, a locked cabinet, a log and a diversion policy has the bones; what changes is the orders, the storage of a product with a much longer in-house session, and the state license questions in section 10.
• The market map. Compass has said that about 80% of Spravato's volume comes from about 20% of its sites, under 2,000 centers.² Elsewhere its executives have talked about an interventional psychiatry footprint of something like 8,500 sites; an independent directory count found only 492 published Spravato listings and found that most ketamine clinics have no Spravato line at all.¹⁶ Whatever the true number, the launch will go first to the high-volume certified sites that already know how to do this, which is what "kid gloves" means. For the long tail of independent practices, Compass announced in April a collaboration with Osmind, whose software runs in more than a thousand interventional clinics, to work out what small sites would need.²¹

4The dosing day, hour by hour

This is the part that does not carry over. Below is a dosing day as the phase 3 protocol ran it, next to a Spravato session, next to my guess at a launch-era day if the label settles at six hours. Drag across the bars or tap a block. The trial bar is from the published design and from the FDA's guidance on how psychedelic trials must be run; the launch bar is a forecast, and I have marked it so.^7,8,1

A few things about that day that the diagram cannot show.

The first hour is admin. Vitals, a check that the person has not taken anything that interacts, a pregnancy test where relevant, a review of the plan for the day, and the quiet conversation about what to do if it gets hard. In the trials there were preparation sessions in the days before, so the dosing-day conversation was a reminder, not an introduction.⁷ At launch, whether a preparation visit happens at all will depend on whether anyone pays for it.

Hours one through four are the drug. Psilocybin's effects begin within roughly half an hour and peak over the next couple of hours, and the person is usually lying back with eyeshades and music, not talking much. The job of the people in the room is to be present and to not do very much, which sounds easy and is one of the harder clinical skills I have watched people learn. Vitals are repeated on a schedule. Nausea, headache, anxiety and visual effects are the common events, and most of them resolve on the day.⁹

Hours four through six are the long tail. The person sits up, talks, eats something, and is gradually less altered, and this is where a clinic's discipline is tested, because the room is quiet and the schedule is pulling. The trial protocol did not let anyone leave until a set of criteria were met, and I will come to those next.

Then the ride home. As with Spravato, the person does not drive, and the label will almost certainly say so until the next day.¹⁴ A companion is part of the plan, not an afterthought.

The Spravato row, for comparison, is a few minutes of spraying, two hours of monitored sitting, and a discharge check. The person is often back in the car by lunch.¹⁴

5Readiness to go home

Compass has said six hours of monitoring, and the published trial design says six to eight, and the honest answer to "how long" is: until the person is ready, as judged by a checklist. As the PI and the dosing-day safety physician at my site, I did a readiness-to-discharge assessment on nearly every participant, and the point of the checklist was never to hit a number on the clock. It was to make sure that the people letting someone leave had looked at the same things, in the same order, every time.

Below is a representative version. It is not the trial's form, which belongs to the sponsor, and it is not a label, which does not exist yet. It is the set of questions any clinic will need some version of, drawn from the Spravato standard (the person must be assessed as clinically stable and ready to leave before they go) and from what the FDA has asked psychedelic trials to do, which is to monitor until the acute effects have resolved.^14,8 Tick it as if you were in the room.

Here is why I belabor this. Compass has said that the monitoring period may be shortened as the FDA allows, and I think that is likely over the life of the product, the way esketamine's requirements have been adjusted in both directions since 2019.^1,15 The thing that lets a sponsor make that argument, and lets a clinic act on it safely, is a discharge standard that does not depend on the clock. A site that builds its day around "six hours and out" will have to rebuild it when the label changes; a site that builds its day around "ready, as documented" will not.

6Rooms and chair-hours

Now the arithmetic. A Spravato visit occupies a treatment room for something like two and a half to three hours including the before and after; Compass's own description of the Spravato day is "a three-hour visit."^16,14 A psilocybin dosing day, at launch, occupies the same room for six to eight.^3,1 But a Spravato patient on the labeled schedule comes twice a week for a month, then weekly, then weekly or every other week, which Levine summarized as 25 to 50 treatments a year; a psilocybin patient is expected to come one to four times a year.^3,14 So the question "how much space does this take" has two different answers depending on whether you ask it per day or per patient. Set your own clinic's numbers and see.

With the defaults, one room-day holds one psilocybin patient or three Spravato patients, which is the number everyone quotes. But over a year, the same room serves several times as many psilocybin patients as Spravato patients (about six, with the defaults), because the Spravato patients keep coming back. A clinic that gives a quarter of its rooms to psilocybin loses a quarter of its Spravato slots per day and gains a psilocybin panel larger than the Spravato panel it displaced. Whether that is good business depends entirely on the next three sections: who has to be in the room, what they cost, and what the payer pays for the day.

Two caveats on the calculator. It assumes rooms are the constraint, and in most clinics I have run, they were not; people were. And it assumes a psilocybin room is a Spravato room, which is nearly true (a recliner, a door that closes, a window you can cover, a bathroom close by) but not quite: the psilocybin room has to be quiet for a working day, which rules out the room next to the waiting area, and it needs somewhere for two staff to sit comfortably for eight hours, which rules out the closet-sized rooms that work fine for a two-hour spray.

7People and qualifications

This is where I expect the real friction, because the headcount per patient-hour is unlike anything in a Spravato workflow. In a Spravato clinic one nurse can watch several patients at once. In the phase 3 trials, and in the FDA's guidance for psychedelic trials, there were two trained monitors with one participant for the whole session.⁸ What the label and the REMS will require at launch is not yet public. Here is the likely cast, with what we know about each role and what is still open.

Put plainly: the trials needed two trained people in a room for a working day, per patient. If the label keeps anything close to that, a clinic that wants to dose two patients a day needs four people free all day, plus a prescriber, plus the person who does the vitals and the discharge, plus someone to do the preparation and integration visits. If the label relaxes to "a trained healthcare provider who monitors," it is a nurse per room and a prescriber in the building, and the economics are recognizably Spravato's with a longer tail. The FDA heard both positions on September 14 and has not yet said which way it leans; one witness proposed that credentialing should follow each sponsor's labeled evidence rather than a uniform rule, which would mean COMP360's label describes something like the trial model.^17,18

8The training burden

Spravato's REMS never mandated a training course. A site certifies, a prescriber signs, patients are enrolled, forms are filed; the know-how came from Janssen's field teams, from conferences, and from the hard way.¹³ Compass's plan, as described in September, is different in one word: the authorized representative will attest that personnel are trained.² Compass says it is funding the development of training programs through grants and will supply materials to organizations; what I have heard, secondhand and not from any document, is that the company intends to run much of the training itself. Treat that as rumor.

What does "trained" mean in hours? For the trials, Compass's published therapist training combined online modules, in-person work and supervised practice, and ran to days rather than hours; the FDA's guidance expects monitors to be trained in the drug's effects, in managing psychological distress, and in the ethics of a session (power dynamics, touch, consent), and the hearing heard calls for training that covers the whole episode of care and extends to referring providers.^12,8,18 My rough forecast:

Attestation is a floor, not a standard. The clinics that will have a bad first year are the ones who hear "attest" and read "checkbox." A psychedelic session that goes sideways at hour three, with a frightened person and two staff who have never seen it, is a very different event from a Spravato patient whose blood pressure spikes, and the drug does not wear off because you want it to.

9The money

Compass has described a per-dose price, with an annual cost it benchmarks to Spravato's roughly $50,000 to $65,000, at an expected two to four doses a year, and it plans to seek a J-code so that sites can buy-and-bill.¹ Buy-and-bill means the clinic pays for the drug, administers it, and then bills the payer for both the drug and the administration, and waits. If a single dose is priced in the five figures, a clinic dosing a handful of patients a week will be carrying a six-figure receivable before the first payment clears. The lesson from esketamine, which launched the same way, was that cash flow killed more small Spravato programs than clinical problems did.

The rest of the money picture is familiar to anyone who has run Spravato on insurance:

• Prior authorization will be the gate. Expect payers to require documented treatment resistance (that is the indication), and I would not be surprised to see some require an esketamine course first, the way esketamine itself was gated behind oral antidepressants.
• The codes. Category III CPT codes exist for in-office psychedelic administration; they are tracking codes, and many payers do not pay on them. Compass says the interim plan is the esketamine codes, and that Category I is about 18 months out.² Whether a six-hour day gets paid as prolonged monitoring, as a bundle, or as the two-hour esketamine rate with the rest absorbed, is the number that decides whether the room math in section 6 works.
• Patient-assistance and copay programs will exist, because they existed for Spravato, and the staff who know how to work them are worth their weight. Compass has said it will field reimbursement specialists.²
• Preparation and integration have no line item in any of this. A therapist's hour before and after is billable as psychotherapy, if the therapist is credentialed with the payer and the visit meets the definition; the monitors' day in the room is not psychotherapy and should not be billed as such. This is where I expect creative and sometimes wrong coding, and where I would want a compliance opinion before the first claim.

10Day 31: the scheduling question

Approval is a federal act; scheduling is federal and then fifty times state. Psilocybin is Schedule I under the Controlled Substances Act. After approval the DEA must reschedule the approved product, which Compass says typically takes about 90 days, and the company has pointed to the April 2026 executive order on serious mental illness as a reason that might be faster.^1,10 Then each state has to follow. Compass counts 44 or 45 states that reschedule automatically when the federal government does, most within 30 days, covering about 93% of the population.² The rest need a board action or a bill.

Utah, where I practice, fixed this in advance. S.B. 83, Controlled Substance Scheduling Amendments, passed in the 2026 session and took effect on May 6, 2026. It says that when the federal government reschedules a "pharmaceutical composition," defined as a drug with FDA approval after clinical trials, that composition "is deemed scheduled under this section in the same schedule."^19,20 Raw psilocybin stays Schedule I in Utah. An approved product would not.

For a clinic, the practical checklist is short and dull: does our DEA registration cover the schedule the product lands in; does our state controlled-substance license; does our malpractice carrier know; does the wholesaler we buy esketamine from have a channel for this product. None of that is hard. All of it takes longer than a week, and the first shipments will go to the sites that did it in advance.

11What we don't know

• How often people come back. COMP006 allowed up to four doses in a year, retreatment improved outcomes, and Levine has said about 40% of patients may respond to a second dose, but "the field cannot yet predict who."^9,3,16 Two to four doses a year is a model, not an observation. If the real number is one, the room math gets easier and the revenue per patient gets smaller.
• Whether the label describes the trial. If it specifies two monitors and their qualifications, clinics inherit the trial's cost structure. If it says "a trained healthcare provider," it inherits Spravato's. The September hearing was the field arguing about exactly this, in public, and the FDA took written comments until October 5.^17,18
• Whether therapy is in or out. The FDA has not established that psychotherapy contributes independently to the effect; the trials used non-directive support, not therapy; one psychologist at the hearing cited data that therapist warmth outperformed pharmacology alone.^17,12 I wrote about this argument in how much is the molecule. The commercial model answers the question before the science does.
• Real-world safety. One Oxford researcher warned the hearing that risks may be underestimated once trial safeguards loosen.¹⁷ Esketamine's respiratory-depression label change came from postmarketing reports, four years in.¹⁵ Expect the psilocybin label to move too.
• The dates. Every one of them is a company target. The honest range for "first patient dosed outside a trial" runs from spring 2027 to a year later.

12A readiness list you can keep

If I were a Spravato clinic deciding whether to be in the first wave, this is the list I would work through now. Tick it; it is saved in your browser only.

13My take

The Spravato model is the fastest route to patients, and it is probably the only one that payers will fund in 2027. Compass is right that the rooms, the REMS reflexes and the billing rails exist, and that the clinics that run them well are the obvious first customers. I would be doing the same thing in their position.

And "it slots right in" is the beginning of the story, not the end of it. A six-to-eight-hour psychedelic session with two trained people in the room, with preparation before and integration after, is a different clinical product from a two-hour dissociative session with a nurse down the hall, and the evidence for the first was generated with the whole wrapper on. A clinic that treats COMP360 as another SKU on the Spravato shelf will get Spravato-level economics on a drug that needs more hands per hour, and I suspect worse-than-trial outcomes, and the outcomes will be blamed on the drug.

The clinics that do well will be the ones that treat the first year as a service line and not a product: that build the day around a discharge standard rather than a clock, that staff for the trial model and are pleasantly surprised if the label is looser, that find a way to pay for the conversation before and after, and that run their first twenty sessions slowly, with a debrief after each one. That is not a sales line. It is just how I would want it done for someone I love.

This is education, not medical or business advice, and it is not a prediction of what the FDA, the DEA or any payer will do. Everything attributed to Compass is from its public statements, cited below; I have no access to its regulatory filings. My relationships to the company and to the field are stated at the top.

References