A treatment can produce a powerful experience without producing a durable clinical benefit. Conversely, a benefit influenced by expectation is not necessarily imaginary. The central question in psychedelic treatment is how much improvement can reasonably be attributed to the intervention, compared with a credible alternative, and whether that improvement survives beyond the treatment setting.
Did people improve? Did they improve more than a comparison group? What caused the difference? These questions require different evidence. A striking change after treatment answers the first; it does not, on its own, answer the other two.
On this page
- Improvement in a placebo group is not a single mechanism
- Why blinding is unusually difficult
- Expectations can influence the comparison treatment
- What negative and better-masked findings contribute
- Separating the drug from the treatment programme
- Clinical importance goes beyond statistical significance
- Who is represented, and for how long?
Improvement in a placebo group is not a single mechanism
Symptoms can improve during a trial because of attention, support, expectations, concurrent care or the natural course of illness. People often seek treatment when symptoms are particularly severe; some subsequent improvement can reflect movement away from that unusually difficult period. Repeated measurement and the opportunity to discuss distress can also affect what is reported.
The term placebo response usually describes the overall change observed in a placebo group. It does not isolate the psychological effect of believing that treatment will work. Estimating that narrower effect requires an appropriate design. Nor does identifying contextual influences establish that a person’s suffering or improvement was fabricated.1
For clinical decisions, the relevant comparison depends on the question. A waiting-list comparison asks something different from a comparison with a similarly supported active treatment. The latter can help establish whether a demanding new intervention offers an advantage over care the patient could otherwise receive.
Why blinding is unusually difficult
Randomisation helps make groups comparable before treatment. Blinding attempts to prevent knowledge of allocation from changing behaviour, care or assessment afterwards. Classic psychedelics produce conspicuous effects, so a participant or therapist may infer the allocation despite a formally double-blind design. This is often described as functional unblinding.
The consequences can run in both directions. Recognising the active drug may increase hope or engagement; believing that placebo was received may generate disappointment. A therapist’s expectations may affect the interaction, and a participant may disclose the experience to an outcome assessor. Independent raters reduce some opportunities for bias, but cannot eliminate changes in the patient’s expectations or behaviour.2
Correct guesses do not prove that all observed improvement was caused by expectation. Some people may infer allocation because they have genuinely improved. Timing matters: a guess made during the acute experience has a different interpretation from a guess made after several weeks of recovery. Studies should specify when they asked, what the participants believed and how certain they were.
Expectations can influence the comparison treatment
In an analysis of 55 participants from a psilocybin-versus-escitalopram depression trial, baseline expectations were higher for psilocybin. Expectations about escitalopram predicted some outcomes in the escitalopram arm, whereas expectations about psilocybin did not significantly predict response in the psilocybin arm. Trait suggestibility was associated with response to psilocybin.3
A separate secondary analysis of the same trial found that the apparent relative advantage of psilocybin on a clinician-rated depression measure varied with baseline expectations of escitalopram. The comparison was more favourable to psilocybin among people with lower expectations of escitalopram. The analyses were small, exploratory and based on selected participants; the pattern was not statistically significant on every outcome.4
These findings argue for measuring expectations about all treatment arms. They do not show that psilocybin is immune to expectancy effects, or establish that changing someone’s expectations would reproduce the observed association. A baseline questionnaire may also miss expectations that change sharply after a person recognises the drug’s effects.
What negative and better-masked findings contribute
A self-blinding citizen-science study followed 191 people undertaking their own microdosing routines. Psychological outcomes improved over four weeks in the microdosing group, but also in the placebo group, without significant between-group differences on those longer-term measures. Smaller differences on some acute and post-acute measures were related to breaking the blind.5
The study shows why personal before-and-after impressions can be misleading. Its self-selected participants, self-supplied drugs and nonclinical setting also limit generalisation. It does not settle whether a standardised full-dose treatment benefits patients with a diagnosed disorder.
Another approach used surgical anaesthesia to conceal ketamine administration. Among 40 adults with depression, ketamine did not significantly outperform saline on depression severity during the first three days. This is valuable negative evidence under a different masking strategy. However, a small surgical study cannot determine the effect of a course of awake ketamine treatment, and ketamine is pharmacologically distinct from classic psychedelics.6
A result that is not statistically significant may be compatible with several possibilities, including little effect or a worthwhile effect estimated imprecisely. The size and uncertainty of the estimate matter. Equally, a positive secondary measure should not be substituted for a negative prespecified primary outcome.
Separating the drug from the treatment programme
Psychedelic interventions commonly combine medication with preparation, supervised sessions and follow-up support. A trial comparing two such programmes estimates the difference between the packages actually delivered. It does not necessarily identify how much benefit came from the molecule, the therapeutic relationship, the altered experience or subsequent behavioural changes.
Active placebos can mimic some sensations and make allocation less obvious. They have trade-offs: a comparator may have therapeutic effects of its own, different adverse effects, or a noticeably different duration. A low psychedelic dose may also be pharmacologically active. There is no single comparator that solves every design problem.7
Factorial designs can vary both drug and psychotherapy to examine their separate contributions and possible interaction, provided essential safety support is maintained. The FDA’s July 2026 final guidance identifies this as a potentially useful approach. It also recommends explicit descriptions of the support or psychotherapy delivered and consideration of how therapist knowledge of allocation could influence later care.2
Clinical importance goes beyond statistical significance
| Finding | What needs to accompany it |
|---|---|
| A large symptom reduction | The change in a credible comparison group, the starting severity and uncertainty around the difference. |
| A response or remission percentage | The definition, follow-up time, denominator and handling of missing observations. |
| Improvement months later | Information about additional treatment, repeat dosing, relapse and who was lost to follow-up. |
| Few reported adverse events | How events were sought, which patients were excluded, and whether delayed problems were assessed. |
For illustration, if 60% improve with a new treatment and 40% improve with its comparator, the absolute difference is 20 percentage points. Calling the first figure a 60% treatment benefit would omit what happened with the alternative. This is a hypothetical example, not a psychedelic trial result.
Response and remission thresholds make results easier to communicate, but can obscure different experiences near the threshold. A patient just below a remission cut-off and one far below it are counted together. Daily functioning, quality of life and adverse effects help determine whether the numerical change is worthwhile.
Who is represented, and for how long?
A study may answer its question well while leaving other patients unrepresented. Eligibility rules, willingness to receive a psychedelic, previous experience, ability to attend long sessions and access to follow-up can all affect who participates. An outcome in that group should not automatically be assigned to people with substantially different clinical circumstances.
Durability also requires more than locating participants months later. Once comparison groups receive additional treatment or cross over to the active intervention, later observations may no longer preserve the original randomised comparison. Long-term follow-up should record other care, recurrence and retreatment, together with functioning and harm.2
Transparent reporting should distinguish a prespecified result from an exploratory association, and a complete trial report from a sponsor’s preliminary announcement. Financial or professional investment does not automatically invalidate findings; disclosure, independent scrutiny and replication make their reliability easier to assess.
Clinically useful evidence can acknowledge both meaningful improvement and uncertainty about its cause or persistence. Patients need a realistic account of what has been demonstrated, how closely the study resembles the proposed care, and what remains unknown. Confidence should grow with well-controlled, replicated and sustained outcomes.
References
- Butler M, Jelen L, Rucker J. Expectancy in placebo-controlled trials of psychedelics: if so, so what? Psychopharmacology (Berl). 2022. doi:10.1007/s00213-022-06221-6.
- US Food and Drug Administration. Psychedelic Drugs: Considerations for Clinical Investigations. Final guidance for industry, July 2026; supersedes the June 2023 draft.
- Szigeti B, Weiss B, Rosas FE, et al. Assessing expectancy and suggestibility in a trial of escitalopram v. psilocybin for depression. Psychol Med. 2024;54:1717–1724. doi:10.1017/S0033291723003653.
- Dutcher E, Krystal AD. Treatment Expectancies and Psilocybin vs Escitalopram for Depression. JAMA Psychiatry. 2025;82:321–322. doi:10.1001/jamapsychiatry.2024.4387.
- Szigeti B, Kartner L, Blemings A, Rosas F, Feilding A, Nutt DJ, Carhart-Harris RL, Erritzoe D. Self-blinding citizen science to explore psychedelic microdosing. Elife. 2021. doi:10.7554/elife.62878.
- Lii TR, et al. Randomized trial of ketamine masked by surgical anesthesia in patients with depression. Nat Ment Health. 2023. doi:10.1038/s44220-023-00140-x.
- Aday JS, Simonsson O, Schindler EAD, D'Souza DC. Addressing blinding in classic psychedelic studies with innovative active placebos. Int J Neuropsychopharmacol. 2025. doi:10.1093/ijnp/pyaf023.