Most psychedelic trials don't check whether blinding worked — and when they do, it often fails

The question

In randomized trials of psychedelics for psychiatric conditions, how often do researchers check whether participants and raters can guess who got the real drug — and how often does that blinding fail?

What they did

The authors searched four research databases for randomized controlled trials testing psychedelics as psychiatric treatments, plus manual searches of three earlier reviews for older studies. They included 112 trials: 78 of ketamine, 17 of LSD, 11 of psilocybin, 11 of MDMA, 2 of ayahuasca, 2 of DMT and 1 of noribogaine. For each trial they recorded whether blinding integrity was measured — that is, whether anyone asked participants or raters to guess which treatment had been given — and what those guesses showed.

What they found

Only 29.5% of trials (33 of 112) assessed blinding integrity, while 57.1% (64) listed blinding as a limitation in their write-up. Where it was measured, failure was common: psilocybin, LSD and ayahuasca trials often reported more than 90% of participants and raters correctly guessing allocation, and MDMA trials using an inert placebo exceeded 85%. Ketamine trials rarely checked blinding at all (17.9%), though blinding held up better when midazolam, a sedative, was used as the comparison instead of saline. No control strategy consistently kept blinding intact.

The fine print

This review counts and describes what trials reported; it does not measure how much unblinding actually changed the results, so it cannot say efficacy findings are wrong. Because only a minority of trials assessed blinding, the failure rates come from a selected subset and may not represent all trials. Ketamine trials dominate the sample, and definitions of blinding failure varied between studies. This summary is based on the abstract only.

What it means

Participants and assessors in psychedelic trials can usually tell who received the active drug, and most trials never check. That makes it hard to separate the drug’s pharmacological effect from expectation, and argues for standardized blinding checks and better study designs.

Evidence grade: Moderate. A systematic review of 112 randomized trials, but it examines study methods rather than pooling treatment effects, and reporting of blinding was inconsistent across the included trials.

This summary is based on the study’s published abstract.