Randomized controlled trial
researchA study in which participants are assigned by chance to an intervention or a comparison condition, so that differences in outcome are less likely to reflect pre-existing differences between the groups. It is the design that supports causal claims most strongly, which is why the corpus so often records its absence: observational and naturalistic reports describe what happened to people who chose their own exposure, not what a substance does.
A randomized controlled trial (RCT) is a study design in which participants are assigned by chance to receive either the intervention under study or a comparison condition — most often a placebo or an existing treatment.
Random assignment is what distinguishes an RCT from observational research. When people choose their own exposure, that choice introduces confounding: someone who seeks out a substance differs in many ways from someone who avoids it. Randomization distributes those pre-existing differences across both groups before the study begins, so that any difference in outcome at the end can be attributed to the intervention rather than to who chose it.
In a double-blind design, neither participants nor the researchers measuring outcomes know which condition anyone is in, removing the placebo effect and observer bias from the result at once. It is this layering of controls that makes an RCT the primary tool for establishing causal claims — and why the corpus so often notes when no such trial exists for a given substance or population.
What this design can establish
A well-conducted RCT can establish that an intervention caused a difference in a measured outcome, not merely that the two occurred together. The claim it supports is specific: this substance, administered in this way to this population, produced this change over this period, compared with a group that differed only in exposure.
Within that scope, RCTs can also support dose-response claims when multiple doses are tested across separate arms. A rigorous negative result — in a trial large enough to have been capable of detecting a meaningful effect — is genuine evidence of absence for the studied outcome in the studied population. That is a different thing from a null result in an underpowered trial, which may simply mean the study was too small to see what was there.
What it cannot
An RCT establishes what happened in its enrolled population under its specific conditions; those results do not automatically extend beyond them. Clinical trials commonly exclude people with psychiatric comorbidities, concurrent drug use, or pregnancy — precisely the populations most relevant to harm reduction. A result from a narrow trial group is evidence about that group, not about everyone.
Most trials run for weeks or months and capture outcomes within that window. They say nothing reliable about effects that develop over years of use, or about what happens after the trial ends.
Very rare adverse events are also largely invisible to standard-sized trials. A study enrolling several hundred participants will not reliably detect an outcome that occurs in one in several thousand. Those signals typically emerge from post-market surveillance, registries, and case series — designs that answer different questions.
The wrong conclusion a reader is most likely to reach is that the absence of an RCT means the absence of an effect. Much of the corpus on psychoactive substances has never been subjected to a controlled trial. This reflects research conditions, legal scheduling, and funding constraints more than it reflects pharmacology. A missing study is a stated gap in the record, not a finding about the substance itself.
AI-generated · not yet verified by a human reviewer
Harm-reduction reference — not medical advice.