When the Treatment Works and We Can't Fully Say Why
Here is the awkward secret sitting at the center of psychedelic medicine: some of the most impressive results in modern psychiatry may be partly, substantially, or even mostly placebo effects — and nobody can currently say which parts, how much, or what to do about it. This isn't a takedown. It's the field's own most-discussed methodological crisis, aired openly in its top journals, and understanding it is essential for anyone who wants to hold the psychedelic promise honestly — neither as hype nor as debunk, but as science in the act of becoming rigorous.
The Structural Problem
Clinical trials depend on blinding: participants and raters shouldn't know who got the active drug and who got the placebo, so that expectation doesn't contaminate the measurement. Psychedelics shatter blinding structurally, in ways no trial design has fully solved:
Participants know. If you took 25mg of psilocybin, you know — the walls moved. If you took the niacin placebo, you know you didn't. Modern trials' own data confirm it: at session's end, the overwhelming majority of both groups correctly guess their assignment. The blind is functionally open.
Raters know. The same obviousness applies to the clinicians administering depression scales. A participant who spent yesterday in tears describing cosmic unity gives different signals than one who had a quiet nap.
Therapists believe. This is the deepest layer. In psychedelic trials, the therapists want the drug to work — many built careers on the hypothesis — and their warmth, attention, and expectancy flow into the very therapy that constitutes half the treatment. The FDA's 2024 rejection of MDMA-assisted therapy cited exactly this: trial conduct in which expectancy and unblinding couldn't be disentangled from the drug effect. It wasn't hostility to psychedelics; it was the regulator insisting that the field's beautiful results survive the same evidentiary standards as less beautiful ones.
The Evidence That Expectation Is Large
Several findings keep the problem vivid. The 2021 Imperial College self-blinding microdosing study — where participants didn't know which of their own doses were placebo — found reported benefits tracking belief about dosing, not actual dosing, with objective measures showing no drug-placebo separation. Sham-controlled studies of psilocybin, using active placebos designed to produce some noticeable bodily effect, have shrunk the apparent drug advantage considerably versus open-label designs. And the broader antidepressant literature offers the sobering benchmark: effect sizes in psychedelic trials, while often large, overlap with the placebo-response range that has been shrinking conventional drug approvals for two decades — in a treatment context where hope, ritual, attention, and meaning are all deliberately maximized.
None of this means the effects are fake. Placebo responses in depression are real improvements — neurobiologically real, clinically real, life-changingly real for the responder. The problem is attribution: if the trial can't separate drug from context, it can't tell regulators, insurers, or patients what caused the improvement, which doses are needed, for whom it works, or whether it works at all beyond the most elaborately supportive possible care.
Why "It's Just Placebo" Is Also Wrong
The lazy debunk fails for the same reason the lazy hype does: psychedelic effects demonstrably exceed expectation in key domains. Dose-response relationships are real — higher doses produce stronger effects and, in several trials, better outcomes, which pure expectancy doesn't generate. Neuroimaging shows characteristic, drug-predictable changes — default mode network modulation, altered connectivity — that don't appear with placebo. And some populations and outcomes show signals that would strain any expectancy account, like the rapid reduction in suicidal ideation observed with ketamine (a different but related problem child of the same methodology). The honest synthesis: expectancy and pharmacology are thoroughly entangled, each amplifying the other, and current designs can't cleanly apportion credit.
The Field's Response
The methodological arms race is genuinely underway, and it's one of the more encouraging stories in modern trial design:
Active placebos. Compounds producing noticeable but non-psychedelic effects (high-dose niacin, low-dose stimulants) so that some signal exists to blind against. They help; they don't solve it — a person who knows what a trip feels like still knows.
expectancy-manipulation designs. Some trials now vary the information participants receive — telling some to expect a profound experience and others to expect mild effects — to measure the expectancy contribution directly. Clever, and early results suggest expectancy contributes meaningfully but doesn't account for everything.
Decentralized raters. Independent clinicians blind to assignment conducting the outcome assessments, so at least the measurement end of the pipeline is protected even if the participant end isn't.
Comparative effectiveness trials. Head-to-head against established treatments (SSRIs, in one landmark psilocybin study) rather than against nothing. The bar moves from "better than placebo" to "better than the best current option" — harder, slower, and the only standard that ultimately matters.
What It Means for Everyone Else
For the person considering psychedelic therapy, the takeaway isn't cynicism — it's calibration. These compounds plausibly work, and also derive enormous power from context, belief, ritual, and relationship, and you can't fully separate the two — which means the container matters as much as the compound, and choosing a serious, well-supported setting isn't woo-woo optimization but a substantial fraction of the treatment effect. For the policymaker and the journalist, the takeaway is epistemic humility: this field's effect sizes are among the most dramatic in psychiatry and among the most methodologically fragile, and the correct posture toward both facts is patience while the better trials run.
The Bottom Line
The placebo problem isn't a scandal to be suppressed or a debunk to be wielded — it's the defining scientific challenge of the psychedelic era, and the field's willingness to face it head-on is, honestly, a reason for confidence rather than doubt. Medicine that survives its own scrutiny deserves the name. The trials now running — active placebos, expectancy manipulations, comparative designs — will tell us, within the decade, how much of the miracle is the molecule and how much is the meaning. Either answer will be worth knowing. The worst outcome would be refusing to ask.
