The unblinding problem: why psychedelic trials may overstate their own results

How much of psychedelic therapy's impressive track record is pharmacology, and how much is expectation? A study published March 18 in JAMA Psychiatry suggests the answer tilts more toward expectation than many in the field had assumed.

The analysis, led by Balazs Szigeti, PhD, at UC San Francisco, tackled one of the most persistent methodological headaches in psychiatry: the impossibility of blinding psychedelic trials. When a participant takes psilocybin and enters a profoundly altered state of consciousness, they know immediately that they did not receive a placebo. That knowledge changes everything about how they respond - and how the trial's results should be interpreted.

A drug you cannot hide

The gold standard in clinical research is the double-blind, placebo-controlled trial. Neither the patient nor the clinician knows who received the active drug. This design controls for the powerful role that belief and expectation play in treatment outcomes, particularly in psychiatric conditions where subjective experience is the primary measure of success.

Traditional antidepressants fit this model well enough. A sertraline pill looks and feels much like a placebo pill. Side effects occur in some patients but not all, and they are rarely dramatic enough to break the blind reliably. Enough uncertainty remains to keep the comparison meaningful.

Psychedelics shatter this framework. The subjective effects of psilocybin and LSD are intense, distinctive, and unmistakable. Studies have confirmed what common sense suggests: nearly all participants in psychedelic trials correctly identify whether they received the drug or the placebo. The trial is effectively unblinded from the moment the drug takes effect.

This creates a systematic bias. Participants who know they received the psychedelic approach their therapy sessions with heightened expectation and openness. Those who know they got the placebo may feel deflated, skeptical, or disappointed - emotions that can worsen depressive symptoms. The resulting gap between groups is real, but it reflects psychology as much as pharmacology.

Equalizing the expectation effect

Szigeti and his colleagues - including co-first author Zachary J. Williams, MD, PhD, of UCLA and Hannah Barnett, MSc, of Imperial College London - devised a creative workaround. Rather than comparing psychedelic trials to traditional blinded antidepressant trials, where the playing field is uneven, they compared them to open-label antidepressant trials. In these studies, every participant knows they are receiving an active medication. No one is left wondering whether they got a sugar pill.

This approach equalizes the expectation variable. In both sets of trials - psychedelic and open-label antidepressant - participants know they are getting a real treatment. Both groups carry the psychological benefit of that knowledge into their treatment experience. If psychedelics are genuinely more effective than antidepressants at the pharmacological level, that superiority should still show up even when expectation effects are matched.

It did not. The analysis found that psychedelic-assisted therapy and open-label traditional antidepressants produced virtually identical reductions in depression symptoms. Both brought scores down by approximately 12 points on standardized depression rating scales - a clinically significant improvement, but indistinguishable between the two treatment types.

Blinding changes outcomes for antidepressants but not for psychedelics

The study revealed another telling finding. For traditional antidepressants, blinding made a measurable difference in outcomes. Patients in blinded trials (where some uncertainty existed about their treatment assignment) showed smaller improvements than patients in open-label trials (where they knew they were getting the drug). This is the expected pattern: knowing you are receiving treatment boosts your response.

For psychedelic-assisted therapy, however, blinding made no difference at all. The outcomes were the same regardless of whether the trial was nominally blinded or not. This confirms what researchers have long suspected: psychedelic trials are always effectively open-label, regardless of how they are designed on paper. The drug's own effects serve as an unmistakable signal that no placebo control can mask.

Recalibrating the hype

The implications extend beyond academic methodology. Over the past decade, psychedelic therapy has attracted enormous investment and public enthusiasm, fueled largely by clinical trial results showing large effect sizes - the magnitude of improvement compared to placebo. These effect sizes were consistently larger than those seen in traditional antidepressant trials, leading to widespread claims that psychedelics were fundamentally more effective.

This analysis reframes that narrative. The large effect sizes in psychedelic trials appear to be driven not by superior pharmacology but by the amplified expectation effects that come with unblinded treatment. When those effects are equalized, the advantage disappears.

That does not make psychedelic therapy useless. A treatment that works as well as existing antidepressants still has value, particularly for patients who have not responded to conventional options. Roughly one-third of patients with major depression do not achieve adequate relief from standard medications, and for those individuals, psychedelic therapy's different mechanism of action could offer a meaningful alternative path.

But it does mean the field needs to adjust its claims. Psychedelic therapy may be a useful addition to the treatment landscape. It does not appear to be the categorical advance that early enthusiasm suggested.

What remains uncertain

This was a meta-analytic comparison across different trials, not a single head-to-head study. The psychedelic and antidepressant data were drawn from separate patient populations treated in different settings under different protocols. While the analytical approach was thoughtful, a properly designed direct-comparison trial would provide more definitive evidence.

The study also examined only major depression. Other conditions under investigation - PTSD, substance use disorders, end-of-life distress - involve different brain circuits, different treatment contexts, and different outcome measures. The findings here should not be generalized beyond depression without additional research.

The durability of treatment effects is another open question. Some proponents argue that psychedelic therapy's advantage lies not in the immediate magnitude of symptom reduction but in the longevity of that reduction - that a single or small number of psychedelic sessions might produce lasting benefits where daily antidepressant use requires indefinite continuation. This study did not directly address that comparison, and the evidence on long-term durability remains mixed.

The study received no external funding. Williams disclosed consulting fees from Roche. The remaining authors reported no conflicts of interest.

The measurement matters as much as the molecule

Perhaps the most important takeaway is not about psychedelics specifically but about the science of measuring psychiatric treatments. When the tool you use to assess a drug's effectiveness is contaminated by expectation effects - and when your trial design cannot control for those effects - the numbers you generate will mislead. They will look like evidence of pharmacological superiority when they are actually evidence of psychological influence.

The psychedelic field is hardly alone in facing this challenge. Surgical trials, device trials, and behavioral intervention trials all grapple with similar blinding problems. But the psychedelic case is particularly stark because the drug's defining characteristic - its ability to produce a powerful altered state - is precisely what makes blinding impossible.

Resolving this will require creative trial designs, honest reporting, and a willingness to let the data temper the enthusiasm. Researchers will need to develop new methods for evaluating treatments whose effects are inherently unblindable - perhaps through active placebo controls that produce some perceptual changes without the full psychedelic experience, or through outcome measures less susceptible to expectation bias. On that front, this study represents a useful step forward - not by answering every question, but by making the right question impossible to ignore.