Psilocybin with psychological support in industry trials

1. Introduction

Version followed: the Compass Psychological Support Model (CPSM) as published for the COMP360 phase 3 programme (Kirlic et al., 2025, American Journal of Psychiatry), with phase 2b operational details from Goodwin et al. (2022, NEJM) and the phase 2b training paper (Tai et al., 2021); the Usona Institute "Set and Setting" facilitator protocol (PSIL201 protocol v10.0, 2021, and the phase 3 uAspire registration) is described as a second industry/non-profit sponsor version where it differs.

What it is. This is not a stand-alone psychotherapy. It is a manualised, sponsor-defined frame of psychological preparation, in-session safeguarding and integration that surrounds administration of an investigational psilocybin product (COMP360, a synthetic psilocybin formulation from Compass Pathways; Usona's investigational psilocybin) in regulated clinical trials for treatment-resistant depression (TRD), major depressive disorder (MDD) and, for Compass, PTSD. The CPSM authors state the model is designed to accompany the drug and is not intended as a treatment for depression on its own [source: Kirlic 2025 via Blossom summary and psychiatryonline]. By 2026 Compass describes it as "monitoring and support" and its CEO has said it is "incorrect to describe the treatment" as psilocybin-assisted psychotherapy [source: Psychiatric Times 2026; Compass Q2 2026 release].

How change is thought to happen (sponsors' account). The drug is held to open a time-limited window of altered perception, mood and cognition; the support frame keeps the person safe, keeps distress within a tolerable "bandwidth", and helps them stay with whatever arises rather than avoid it, then draw their own meaning from it afterwards. The phase 2b model grounded this in Perceptual Control Theory: distress reflects conflicting goals the person is controlling for, and awareness of those goals (via self-directed enquiry and experiential processing) allows reorganisation [source: Tai 2021]. The phase 3 CPSM emphasises trust/psychological safety, present-moment focus and reducing experiential avoidance, and participant autonomy [source: Kirlic 2025].

Who can deliver it, and legal status (as of the sources' dates). Only trained, licensed mental-health professionals working inside an approved trial or authorised scheme [source: Tai 2021; Kirlic 2025].

USA: psilocybin is Schedule I. Compass's COMP360 had two positive phase 3 TRD trials (COMP005, COMP006), FDA Breakthrough Therapy designation, and a rolling NDA expected to be complete in Q4 2026; launch is anticipated in the first half of 2027 subject to FDA approval and prior DEA rescheduling, with REMS certification of sites expected [source: Compass Q2 2026 release, 5 Aug 2026]. Usona's phase 3 uAspire MDD trial (about 240 participants, 29 US sites) was active, not recruiting, with listed end January 2027 [source: Blossom trial record]. State-level programmes (e.g. Oregon, Colorado) are separate from this model and were not reviewed [uncertain].
UK: COMP360 holds an Innovative Licensing and Access Pathway designation; no marketing authorisation [source: Compass Q2 2026 release]. Psilocybin remains a controlled drug in the UK and EU; detailed EU status [uncertain].
Australia: since 1 July 2023 psilocybin is Schedule 8 for TRD only, prescribable only by psychiatrists approved under the TGA Authorised Prescriber scheme with ethics-committee approval; the psychiatrist must supervise and remain present throughout the session, other qualified professionals may assist, and treatment needs a supervised setting able to manage behavioural and medical emergencies [source: NSW Health, current at 2 Sep 2025].

Outside these settings, a psychotherapist's role is assessment, harm reduction, referral to trials and integration support, never facilitation (see section 5) [synthesis].

2. Framework

Three stages, each with defined goals; safety underpins all of them [source: Tai 2021; Kirlic 2025].

Stage 0: eligibility, screening and consent (medical/research team; therapist contributes).

Phase 2b Compass TRD: adults with DSM-5 MDD without psychotic features, confirmed by clinical assessment, records and the MINI; non-response in the current episode to two to four adequate antidepressant trials (MGH-ATRQ); those at clinically significant suicide risk were excluded; further criteria sit in the protocol [source: Goodwin 2022].
Usona PSIL201 published exclusions include: psychotic disorders, bipolar I or II, antisocial personality disorder, moderate-to-severe alcohol or drug use disorder in the past year, any condition judged incompatible with establishing rapport with facilitators or with safe exposure; a first-degree relative with a psychotic disorder or bipolar I; recent suicidal ideation or behaviour on the C-SSRS or MADRS item 10; significant cardiovascular disease, uncontrolled blood pressure, stroke/TIA, epilepsy, significant hepatic impairment, insulin-dependent diabetes; pregnancy or nursing; recent extensive psychedelic use; positive drug screen. Participants needed an identified support person, and ongoing psychotherapy had to be paused during the trial [source: Usona PSIL201 v10.0]. uAspire adds active PTSD in the prior six months and certain personality disorders [source: Blossom uAspire record].
Screening includes psychiatric interview, physical examination, bloods, ECG, urine drug screen and pregnancy testing; prior history confirmed with the treating doctor [source: Goodwin 2022; Carhart-Harris 2018 for the academic precursor].
Interacting psychiatric medicines are tapered and stopped under psychiatric supervision before baseline; this is a medical-team decision and the therapist does not advise on it [source: Goodwin 2022; Usona PSIL201].

Stage 1: preparation. Goals: build the alliance and trust; psychoeducation and realistic expectations; set a personal intention; co-create and rehearse a support plan, including touch boundaries; practise experiential skills so the person can meet the experience "safely and autonomously" [source: Kirlic 2025]. Format: individual, with the lead therapist who will be present on the day. Count: at least three preparation meetings in Compass phase 2b, held during a 3- to 6-week run-in [source: Goodwin 2022]; Usona phase 2 gave 6 to 8 hours of preparation with both facilitators [source: The Hospitalist on Raison 2023]. Phase 3 CPSM session counts were not given in the article [uncertain]. Participants also review sponsor psychoeducation material on a dedicated online platform before the first session [source: Kirlic 2025].

Stage 2: administration session. Goals: safety, keep distress tolerable so the experience stays meaningful, therapist as safeguard with minimal intervention [source: Tai 2021; Kirlic 2025]. Format: a full-day session (about 6 to 8 hours in Compass phase 2b; 7 to 10 hours listed for uAspire) in a non-clinical, calming room; eyeshades and a specially designed music playlist (phase 2b); two trained therapists/facilitators present throughout; a trial psychiatrist on site; participant leaves only when effects have fully resolved and, for Usona, with their support person [source: Goodwin 2022; Usona PSIL201; Blossom uAspire]. Medical monitoring (vital signs etc.) is a medical-staff responsibility.

Stage 3: integration. Goals: participant-led reflection, linking what happened to their own life story, consolidating insights, and monitoring for risk states [source: Tai 2021; Kirlic 2025]. Count/format: Compass phase 2b had two sessions, the day after (both therapists) and one week later (lead therapist) [source: Goodwin 2022]; Usona PSIL201 had three integration sessions, the first the next day [source: Usona PSIL201]; uAspire lists integration sessions of about 4 hours [source: Blossom uAspire]. Follow-up safety and symptom contacts continue for weeks to a year, rated remotely and by blinded raters [source: Goodwin 2022; Blossom uAspire].

Criteria for moving between stages. Not formalised as clinical milestones in the sources; progression is protocol-driven. Inferred readiness markers for preparation-to-session: eligibility reconfirmed at baseline, no emergent suicidality (in Usona, suicidal ideation in the screening/preparation period leads to screen failure or withdrawal), intention set, support plan and touch boundaries agreed and practised, participant can describe what to expect and how to signal distress [source: Usona PSIL201; Kirlic 2025; ordering of readiness markers is synthesis].

3. Guiding principles

Safety first [source: Tai 2021; Kirlic 2025]: every decision is filtered through physical and psychological safety; you stop, escalate or call the psychiatrist rather than "push through".
Trust and psychological safety [source: Kirlic 2025]: get to know history, trauma, identity and culture with unconditional positive regard, so the person can let go on the day.
Non-directive, participant-centred stance [source: Tai 2021; Kirlic 2025]: no teaching, advice, interpretation or steering towards answers; your questions help them find their own.
Self-directed enquiry [source: Tai 2021]: ask "sensitive but searching questions" that turn attention to thoughts, emotions, sensations, images and memories as they arise.
Experiential processing / reducing avoidance [source: Tai 2021; Kirlic 2025]: help the person stay with experiences "in and through" until they pass or change, since avoidance is linked with worse sessions.
Present-moment focus [source: Kirlic 2025]: enquire about what is happening now, not about theories of why.
Autonomy [source: Kirlic 2025]: the participant sets and revisits their own intention, chooses the support strategies and can decline any of them, including touch, at any time.
Minimal intervention on the day [source: Kirlic 2025; Psychiatric Times 2026]: presence is the main intervention; in the Compass PTSD study about 78% of session time was silence.
Agreed in advance, nothing new on the day [source: Kirlic 2025]: only strategies rehearsed and consented in preparation are used, one at a time.
Let distress unfold if it is processable [source: Kirlic 2025; Tai 2021]: distinguish workable emotional processing from intolerable distress; intervene only for the latter, from least to most intensive.
Therapist self-awareness [source: Tai 2021]: notice your own boredom, anxiety or fear and keep modelling openness and interest.
Ethics over conviction; heightened suggestibility [source: Tai 2021]: you sign up to put participant welfare above personal, ideological, religious or spiritual views and avoid anything that could manipulate a suggestible person.
Research integrity and blinding [source: Goodwin 2022; Kirlic 2025]: therapists do not collect efficacy data, are blind to allocation and should not speculate about dose; fidelity is rated by independent raters.
Vigilance for suicidality after dosing [source: Goodwin 2022; Kirlic 2025]: suicidality "demands clinical vigilance"; watch especially non-responders with prior self-harm.

4. Intervention inventory

The sponsor manuals (Compass therapist manual; Usona Facilitator Training Manual and Manual of Procedures) are not public. The items below are those named in the published descriptions; step detail beyond what the sources state is marked [synthesis]. All example language is illustrative unless quoted and cited.

A. Preparation stage

4.1 Psychoeducation and expectation management

Purpose / when: first preparation session onward; give an accurate picture of the treatment and correct misconceptions [source: Kirlic 2025].
Steps: (1) Before session 1 the participant reviews the sponsor's online material (putative mechanism, subjective effects, safety, the process of the day, previous participants' accounts) [source: Kirlic 2025]. (2) Ask what they took from it and what they expect or fear. (3) Clarify the sequence of the day (room, eyeshades and music if used, two therapists, psychiatrist on site, staying until effects resolve, going home with a support person) [source: Goodwin 2022; Usona PSIL201]. (4) Name common transient effects reported in trials (brief anxiety, headache, nausea, fatigue; occasional transient paranoia or confusion) without promising outcomes [source: Goodwin 2022; Carhart-Harris 2018]. (5) Explain that improvement is not guaranteed, that low mood can return or worsen afterwards, and how to reach the team [source for risk: Goodwin 2022; framing is synthesis]. (6) Check understanding by asking them to describe it back [source: Kirlic 2025 notes therapists check understanding].
Example language (illustrative): "What did you notice in the material, and what are you hoping or worried will happen?" / "Some people find parts of the day frightening. That doesn't mean something has gone wrong, and we'll be with you throughout."
Watch for: fixed expectations of cure (risk of post-session despair), expectations of being guided or analysed, misunderstanding of blinding (they may receive placebo or a comparator). Working when the participant can explain the day in their own words.

4.2 Trust-building and life-history exploration

Purpose / when: preparation; build the alliance and identify themes that may surface ("discussing key life events" is a named preparation component) [source: Kirlic 2025; Usona PSIL201 states preparation identifies personal themes].
Steps: (1) Invite the person's story: key life events, relationships, trauma history, identity and cultural factors, current mental state [source: Kirlic 2025]. (2) Hold unconditional positive regard; reflect rather than interpret. (3) Note themes likely to arise and agree how they want you to respond if they do [synthesis]. (4) Keep pacing so disclosure does not destabilise before the day [synthesis].
Example language (illustrative): "Which events in your life do you think might show up for you on the day?"
Watch for: undisclosed trauma, suicidal thoughts or substance use emerging during disclosure (report to the medical team; Usona treats suicidality in this phase as grounds for screen failure or withdrawal) [source: Usona PSIL201]. Working when the person can talk about difficult material with you without shutting down.

4.3 Intention setting

Purpose / when: preparation, then revisited at the start of the administration session to reinforce "focus, purpose, and autonomy" [source: Kirlic 2025].
Steps: (1) Ask the participant to formulate their own intention for the session and for taking part. (2) Do not supply or edit content; help them make it brief and personal. (3) Record it in their words. (4) On the day, invite them to restate it before effects begin [source: Kirlic 2025; steps 2–3 synthesis].
Example language (illustrative): "If you were to hold one question or wish going into the day, what would it be, in your own words?"
Watch for: intentions that are really demands for a specific outcome; gently ask how they will be with whatever comes instead [synthesis].

4.4 Collaborative support plan

Purpose / when: preparation; pre-agree exactly how you will support distress so nothing new is introduced on the day [source: Kirlic 2025].
Steps: (1) Review the menu: experiential practices (breath, body awareness), physical proximity without touch, brief verbal cues such as "let go" or "be open", and consented touch [source: Kirlic 2025]. (2) Choose with the participant according to history, presentation and preferences. (3) Practise each one in session. (4) Agree a simple signal for wanting more or less contact [synthesis]. (5) At the start of the administration session, review and re-consent the plan; do not add new strategies then [source: Kirlic 2025].
Example language (illustrative): "If things feel too much, what would help most: me reminding you to breathe, me moving closer, or a hand on your shoulder?"
Watch for: agreement given to please you; re-check consent in a later session [synthesis]. Working when the plan is specific and the participant recalls it unprompted.

4.5 Establishing touch boundaries

Purpose / when: preparation; CPSM lists "establishing touch boundaries" as a named preparation task [source: Kirlic 2025].
Steps: (1) Explain policy: touch is optional, never sexual, limited to the arm, the shoulder or non-interlocking hand-holding [source: Kirlic 2025]. (2) Ask whether they want any touch at all; no is a full answer. (3) Practise the agreed touch in preparation so it is familiar. (4) Explain they can withdraw consent at any moment on the day, including non-verbally [source: Kirlic 2025]. (5) Document the agreement; two therapists are present on the day [source: Goodwin 2022; documentation synthesis].
Example language (illustrative): "Touch is entirely your choice. If you want it, we'd only ever use a hand on your arm or shoulder, or you holding my hand. You can say no now or at any point."
Watch for: trauma histories involving touch; ambivalence. Default to proximity without touch if unsure [synthesis].

4.6 Experiential exercises (breath awareness, body awareness, visualisation, self-compassion)

Purpose / when: preparation; build capacity to stay with intense states and to self-regulate; reduce experiential avoidance [source: Kirlic 2025; Tai 2021 drew on mindfulness, ACT and Focusing].
Steps (generic; exact scripts are in the unpublished manual): (1) Breath awareness: invite attention to the natural breath, noticing without changing it, then returning attention when it wanders. (2) Body awareness: slowly scan for sensations, naming them neutrally. (3) Visualisation: an image the participant chooses that represents safety or openness. (4) Self-compassion: a kind inner stance toward pain. (5) Debrief each: what was hard, what helped. (6) Pick one or two to include in the support plan [source: Kirlic 2025 names the four; steps are synthesis].
Example language (illustrative): "Just notice the breath as it is. If your mind goes somewhere else, that's fine; come back when you notice."
Watch for: exercises triggering dissociation or panic in trauma survivors; offer eyes-open or grounding variants [synthesis]. Working when the participant can use one exercise to settle in a mildly stressful moment.

4.7 Self-directed enquiry (open-ended, present-moment questioning)

Purpose / when: throughout preparation and integration; used sparingly and only as agreed on the day. Models "curious, internal exploration" [source: Kirlic 2025; Tai 2021].
Steps: (1) When the participant talks about an experience, ask what they notice now: thought, feeling, sensation, image, memory [source: Tai 2021]. (2) Follow their lead with "sensitive but searching questions" [source: Tai 2021]. (3) Invite them to look at it from another angle or notice what is in the background of their mind (Method of Levels influence) [source: Tai 2021 cites MOL; specific move synthesis]. (4) Do not offer explanations, advice or solutions [source: Tai 2021; Kirlic 2025].
Example language (illustrative): "As you say that, what are you aware of right now?" / "What's it like to notice that?" / "Is there anything in the background as you describe it?"
Watch for: sliding into interpretation or teaching; the participant looking to you for answers. Working when they generate their own observations and links.

4.8 Experiential processing ("in and through")

Purpose / when: preparation (practice) and on the day; help the person stay with a difficult state until it passes or changes [source: Tai 2021].
Steps: (1) When something difficult arises, encourage approach rather than avoidance. (2) Invite them to locate it in the body and stay with it. (3) If agreed, use a brief cue ("let go", "be open") [source: Kirlic 2025]. (4) Allow it to unfold naturally if it is processable; switch to a support strategy only if distress becomes intolerable [source: Kirlic 2025].
Example language (illustrative): "You don't have to make it go away. Can you let it be there and see what it does?"
Watch for: flooding or dissociation beyond the "bandwidth of tolerance" [source: Tai 2021]; signs that "staying with" has become re-traumatising rather than processing [synthesis].

B. Administration session

4.9 Setting and opening the day

Purpose / when: start of the administration session; create a calm, predictable, non-clinical setting [source: Goodwin 2022; Usona PSIL201].
Steps: (1) Room arranged as non-clinical and calming (Compass) or "aesthetically pleasing" (Usona), identical procedures for active and placebo arms [source: Goodwin 2022; Usona PSIL201]. (2) Both therapists present; confirm the psychiatrist is on site and how to reach them [source: Goodwin 2022; Tai 2021]. (3) Participant restates their intention [source: Kirlic 2025]. (4) Review and re-consent the support plan and touch boundaries [source: Kirlic 2025]. (5) Eyeshades and the sponsor's music playlist are offered to direct attention inward (Compass phase 2b) [source: Goodwin 2022; the phase 3 article does not mention them, uncertain whether unchanged]. (6) Drug administration and medical checks are done by the medical/research team under protocol; the therapist does not handle them.
Example language (illustrative): "Before we start, can you tell us your intention again? And let's run through what we agreed: you'd like reminders about breathing and for one of us to sit closer if it gets hard."
Watch for: last-minute ambivalence; the participant may decline to proceed at any time [synthesis based on research consent].

4.10 Presence and minimal intervention (safeguarding)

Purpose / when: the whole session; the core in-session "intervention" is attentive, quiet presence [source: Kirlic 2025; Psychiatric Times 2026].
Steps: (1) Sit nearby, attentive, mostly silent; in the Compass PTSD study about 78% of session time was silence [source: Psychiatric Times 2026]. (2) Do not lead, redirect or reinterpret the experience [source: Psychiatric Times 2026]. (3) Encourage a "courageous, trusting and open attitude" when invited or when the participant surfaces [source: Tai 2021]. (4) Two therapists remain in the room throughout, with only brief breaks (Usona) [source: Usona PSIL201]. (5) Lead therapist (fully trained) and assisting therapist (may be a trainee) share observation; specific role division is not published [source: Tai 2021; uncertain]. (6) Pass any medical concern to the psychiatrist/medical staff immediately.
Example language (illustrative): "We're right here." (only if the participant checks in)
Watch for: your own urge to fill silence or "help"; boredom, anxiety or fear in yourself [source: Tai 2021]. Participants in Compass's PTSD study described providers as a reliable resource "largely out of their awareness" [source: Psychiatric Times 2026], which is the intended effect.

4.11 Brief verbal cues

Purpose / when: when distress is rising and the participant has consented to cues in preparation [source: Kirlic 2025].
Steps: (1) Use only the cues agreed (e.g. "let go", "be open") [source: Kirlic 2025]. (2) Short, calm, one at a time. (3) Then return to silence and observe the effect [source for "one at a time": Kirlic 2025].
Example language: "let go" / "be open" [source: Kirlic 2025, verbatim cue words].
Watch for: cues heard as commands or criticism; stop if they increase agitation [synthesis].

4.12 Breath and body-awareness prompts on the day

Purpose / when: for rising anxiety, using the exercise rehearsed in preparation [source: Kirlic 2025; Psychiatric Times 2026].
Steps: (1) Remind them of the practice by its agreed name. (2) Offer a single prompt to notice the breath or a body area. (3) Allow time; do not stack prompts [synthesis].
Example language (illustrative): "Remember the breathing we practised. Just notice the next breath."
Watch for: breath focus increasing panic in some people; switch to proximity [synthesis].

4.13 Proximity adjustment

Purpose / when: when the participant needs more sense of safety and has agreed to it [source: Kirlic 2025].
Steps: (1) Move closer, without physical contact. (2) Let them know you have moved if eyeshades are on [synthesis]. (3) Move back once settled or if they signal [synthesis].
Example language (illustrative): "I'm sitting a bit closer now."
Watch for: startle or feeling crowded.

4.14 Pre-consented supportive touch

Purpose / when: only if agreed and practised in preparation, and only for reassurance during distress [source: Kirlic 2025].
Steps: (1) Check the plan says yes. (2) Ask or signal first where possible [synthesis]. (3) Use only the agreed form: hand on arm or shoulder, or non-interlocking hand-holding; never sexual [source: Kirlic 2025]. (4) Withdraw at once if the participant declines or pulls away [source: Kirlic 2025]. (5) Second therapist present [source: Goodwin 2022].
Example language (illustrative): "Would you like my hand on your shoulder, as we agreed?"
Watch for: increased suggestibility and altered perception can distort the meaning of touch [source: Tai 2021 on suggestibility; link to touch is synthesis]; record any touch used [synthesis].

4.15 Graded response to distress and escalation

Purpose / when: whenever distress appears [source: Kirlic 2025].
Steps: (1) Judge whether the distress is processable (let it "unfold naturally") or intolerable [source: Kirlic 2025]. (2) If intolerable, apply agreed strategies one at a time, least intensive first: cue, breath/body prompt, proximity, touch [source: Kirlic 2025 states sequencing from less intensive; exact order is synthesis]. (3) For acute unmanageable anxiety or psychotic symptoms, call the psychiatrist for evaluation; any rescue medication is a medical decision, reported as rare (under about 1.5% of participants in prior trials) [source: Kirlic 2025]. (4) Document and report as an adverse event per protocol [source: Goodwin 2022; Usona PSIL201].
Example language (illustrative): "This is very intense. You're safe, we're here. Do you want me closer?"
Watch for: the participant becoming uncommunicative (seen in Imperial's academic precursor; it resolved as effects waned) [source: Carhart-Harris 2018]; transient paranoia; confusion about the reality of recovered memories (see 5.6).

4.16 Ending the session and safe release

Purpose / when: end of day.
Steps: (1) Stay until effects have fully resolved [source: Goodwin 2022]. (2) Medical staff complete release checks [source: Usona PSIL201 refers to a release procedure; details not published]. (3) Participant leaves with their identified support person to a safe destination (Usona) [source: Usona PSIL201]. (4) Confirm the next-day integration appointment and how to reach the team overnight [synthesis].
Example language (illustrative): "We'll see you tomorrow. Tonight, just rest; you don't need to make sense of it yet."
Watch for: lingering perceptual changes or distress; report to medical staff.

C. Integration stage

4.17 Participant-led integration session

Purpose / when: day after the session and later sessions; reflect, make meaning, consolidate [source: Goodwin 2022; Kirlic 2025; Tai 2021].
Steps: (1) Invite an open account of the experience. (2) Ask the participant to select the most salient themes [source: Kirlic 2025]. (3) Explore them with non-directive enquiry, "without active guiding" [source: Goodwin 2022]. (4) Help them connect emotional, cognitive and bodily experiences to their personal narrative [source: Tai 2021]. (5) Do not interpret, advise, lead to answers or impose your values [source: Kirlic 2025]. (6) Help them identify activities to continue integration outside sessions [source: Kirlic 2025].
Example language (illustrative): "Of everything that happened, what stands out most for you today?" / "What do you make of that?" / "Is there anything you'd like to do differently in the coming weeks because of it?"
Watch for: pressure to have "the right" experience; if they ask what it meant, return the question gently. Working when the participant articulates their own insights and concrete next steps.

4.18 Post-session risk monitoring

Purpose / when: every integration contact and follow-up [source: Kirlic 2025; Goodwin 2022].
Steps: (1) Check in regularly for despair, disappointment, worsening depression and "persisting ideas" that work against the participant's goals [source: Kirlic 2025]. (2) Stay vigilant to persistent unusual beliefs or ideas of harm to self or others [source: Kirlic 2025]. (3) Suicidal ideation with intent or any suicidal behaviour was reported as a serious adverse event in Compass phase 2b; follow the protocol's C-SSRS-based process and escalate to the trial physician [source: Goodwin 2022]. (4) Pay particular attention to non-responders with prior suicidal behaviour or self-injury, the group in which later suicidal behaviour occurred [source: Goodwin 2022].
Example language (illustrative): "Some people feel let down or lower in the days after. How has your mood been, and have you had any thoughts of harming yourself?"
Watch for: disappointment in placebo or non-responding participants; the antidepressant-free period after dosing in the Compass protocol [source: Goodwin 2022].

D. Therapist-side procedures

4.19 Training, mentoring and fidelity self-rating

Purpose / when: before and throughout trial work [source: Tai 2021; Kirlic 2025].
Steps: Phase 2b: about 20+ hours online learning, a five-day in-person role-play training, at least four supervised sessions as assisting therapist, then at least monthly mentoring and webinars [source: Tai 2021]. Phase 3: online self-study with knowledge tests, a three-day remote live workshop, review of case recordings, direct observation, competency benchmarks, and group mentoring with review of recordings and transcripts [source: Kirlic 2025]. Sessions are video-recorded; therapists self-rate each session on a fidelity scale (phase 2b: 25 preparation, 18 session and 12 integration items, rated yes / yes but below desired level / no) and compare with independent ratings in mentoring [source: Tai 2021].
Watch for: drift into directive or interpretive work; personal convictions influencing participants [source: Tai 2021].

E. Outside a trial or authorised scheme

4.20 Referral, harm reduction and integration support for clients using psilocybin elsewhere [synthesis]

Purpose / when: a client asks about these treatments, wants to join a trial, or reports use outside a regulated setting. This is not part of the sponsor models; it is general practice derived from them.
Steps: (1) Explain the treatment is investigational in the US and UK and lawful in Australia only via authorised psychiatrists [source: Compass Q2 2026; NSW Health]. (2) Refer to registered trials where suitable; trials usually ask that ongoing psychotherapy be paused, so plan a handover and re-entry [source: Usona PSIL201]. (3) If the client intends to use outside a regulated setting, do not facilitate, supply or advise on use; screen for the risk factors trials exclude (psychosis or bipolar history in self or first-degree relatives, recent suicidality, cardiovascular disease, epilepsy, pregnancy, interacting psychiatric medication) and discuss these risks in general terms [source for criteria: Usona PSIL201; harm-reduction stance synthesis]. (4) Advise against stopping prescribed medication without their prescriber. (5) Afterwards, offer integration using 4.17 and risk monitoring using 4.18.
Example language (illustrative): "I can't help you obtain or use it, but I can help you think through the risks for you specifically and talk through whatever comes up afterwards."
Watch for: post-experience destabilisation, persisting perceptual changes, new fixed beliefs, suicidality; refer to psychiatry promptly.

5. Handling problems

The sponsor publications say little about problem-handling beyond the graded distress response. Items without a source tag are [synthesis].

5.1 Overwhelming anxiety or panic in session. Use the agreed plan, one strategy at a time, least intensive first; let processable distress unfold; escalate to the psychiatrist for unmanageable anxiety or psychotic symptoms; rescue medication is a medical decision and rare [source: Kirlic 2025; Goodwin 2022 reported one participant needing rescue medication for acute anxiety]. Afterwards, review with the participant what helped [synthesis].

5.2 Paranoia, confusion or uncommunicativeness in session. Reported transiently in academic precursor work and resolved as effects waned [source: Carhart-Harris 2018]. Respond with calm presence, orientation if agreed, and medical review if safety is at risk [synthesis]. Do not argue with content [synthesis].

5.3 Resistance or avoidance (fighting the experience). Sources link experiential avoidance to worse sessions and train approach in preparation [source: Kirlic 2025]. In session, a single agreed cue ("let go", "be open") [source: Kirlic 2025]; never force. In integration, explore what the resistance protected without labelling it failure [synthesis].

5.4 "Nothing happened" or disappointment (including possible placebo). Therapists are blind and must not speculate about allocation or dose [source: Goodwin 2022; Tai 2021]. Validate disappointment, explore what was experienced, and monitor closely for despair and worsening depression, which CPSM names explicitly [source: Kirlic 2025].

5.5 Post-session suicidality or worsening depression. In Compass phase 2b, suicidal ideation, self-injury and suicidal behaviour occurred after dosing in active arms, mainly in people with prior self-harm who had not responded [source: Goodwin 2022]. Ask directly at every contact, use the protocol's C-SSRS process, escalate to the trial physician, and do safety planning per standard practice, including lethal-means counselling on secure storage or removal [source for vigilance: Goodwin 2022; safety-planning steps synthesis].

5.6 Emergent memories of abuse whose reality is unclear. In the Imperial precursor study, a patient became unsure whether a recalled memory was real; the team chose not to judge its accuracy and kept listening openly, and the patient later improved [source: Carhart-Harris 2018]. Do not confirm or refute; avoid suggestion given heightened suggestibility [source: Tai 2021]; involve the medical team if distress persists; handle any safeguarding duties per local law [synthesis].

5.7 Persisting unusual beliefs or ideas after the session. CPSM directs vigilance for persistent unusual beliefs and ideas of harm to self or others [source: Kirlic 2025]. Explore curiously, check function and risk, refer for psychiatric assessment if beliefs are fixed, distressing or impairing [synthesis].

5.8 Requests for interpretation or advice. The model forbids interpretation and advice [source: Kirlic 2025]. Return the question ("What do you make of it?" (illustrative)), and offer only reflections of their own words [synthesis].

5.9 Ruptures (e.g. participant felt abandoned by silence, or touch felt wrong). Raise it early in integration, apologise for the impact, re-establish consent and boundaries, and report boundary concerns to the sponsor's mentor or supervisor [synthesis]. Touch can be declined at any moment [source: Kirlic 2025].

5.10 Non-adherence to protocol (missed sessions, non-disclosed substance use, restarting medication). Report to the investigator rather than managing it yourself; medication decisions belong to physicians [source: Goodwin 2022 notes a physician could restart antidepressants if clinically necessary; reporting synthesis].

5.11 Therapist drift and boundary risk. Fidelity rating, video review and mentoring exist to catch drift into directive work and attitudes that could expose participants to manipulation; therapists sign an ethics code placing participant welfare above personal convictions [source: Tai 2021]. Bring any boundary concern to mentoring immediately [synthesis].

5.12 Gaps the sources do not address. Not covered in the sources: the published session-by-session manual content; the phase 3 division of roles between lead and assisting therapist; current Compass policy on eyeshades and music in phase 3; any guidance for cultural adaptation; and how the model will operate post-approval under REMS [uncertain]. Integrating a client back into prior psychotherapy after a trial is [synthesis]: obtain consent to liaise with the trial team, review what changed, and continue risk monitoring.

Sources

Sources:

Further reading (not drawn on): Raison CL et al. (2023) phase 2 randomized clinical trial of psilocybin in major depressive disorder, JAMA [uncertain beyond the news summary]; Dougherty RF, Modlin NL, McGowan NM et al. (2026) Silence is golden, J Psychopharmacol [via Psychiatric Times only]; COMPASS PTSD protocol NCT05312151 (text unreadable) [uncertain].

← Previous · VII-055Psilocybin-assisted therapy (Hopkins, NYU, Imperial)Next · VII-057 →MDMA-assisted therapy for PTSD