How do you study mind-altering drugs when every clinical-trial participant knows they’re tripping?
Listen to this article
00:00
10:47
Produced by ElevenLabs and News Over Audio (NOA) using AI narration.
Tomorrow, a Food and Drug Administration advisory committee will meet to discuss whether the United States should approve its first psychedelic drug. The fate of the treatment—MDMA-assisted therapy for post-traumatic stress disorder—will turn on how the FDA interprets data from two clinical trials that, on their face, are promising. Long-suffering patients who took the drug while undergoing intensive talk therapy were about twice as likely to recover from PTSD as patients who got the placebo with therapy.
If the treatment is approved this summer, it could bring relief to some of the approximately 13 million Americans with PTSD. It could also serve as a model for other psychedelics to meet the FDA’s regulatory bar. But there’s a conundrum at the core of these two clinical trials, one that has plagued virtually all efforts to study psychedelics.
In clinical trials, participants (and the researchers studying them) generally aren’t supposed to know whether they’re getting the actual drug or a placebo, to avoid allowing people’s expectations about a treatment to shape their response to it. Blinding, as this practice is called, is a key component of a randomized controlled clinical trial, or RCT—medicine’s gold standard for demonstrating that a drug actually works. But virtually no one can take a psychedelic drug and not know it.
Some experts believe that unblinding threatens to undermine the entire field of psychedelic research because it means researchers can’t know whether the drugs’ early promise in clinical trials is real or a mirage, driven by the placebo effect and outsize expectations about the power of these drugs. But others argue that RCTs themselves are at fault. To them, psychedelics are exposing long-ignored cracks in our gold standard, especially for testing drugs that act on our minds.
Read: What it’s like to trip on the most potent magic mushroom
When randomized controlled trials are well designed, “there is no substitute,” Boris Heifets, a neuroscientist at Stanford University, told me. In an RCT, participants get randomly sorted into two groups, receiving either the treatment or a placebo. Scientists have prized such trials since the 1960s for their power to rule out all the nondrug reasons people who are given a new medication might get better. Chief among those reasons is the placebo effect, in which a patient’s belief in a treatment, rather than anything about the drug or procedure itself, leads to improvement. If trial participants come in with sky-high expectations (as experts suspect is the case in many psychedelics trials), knowing that they’ve received a drug could fuel positive responses, and learning they’ve been denied it could cause them to react negatively. “We’ve gotten a ton of things wrong by trusting unblinded results,” says David Rind, the chief medical officer of the Institute for Clinical and Economic Review, a nonprofit that evaluates new medical treatments.
For all of RCTs’ advantages, “I think it’s obvious that they’re not well suited for studying psychedelics,” Heifets said. In cancer-drug trials, participants won’t know the difference between a saline IV drip and medicine; to test new surgical procedures, control groups sometimes get cut into and sewed up without the actual treatment. But psychedelics like psilocybin or LSD launch people into hallucinatory states that bend space and time. MDMA, known to many as ecstasy, is less extreme, but still sparks expansive feelings of love and empathy. “Participants will know within half an hour whether they’ve been assigned to the experimental or placebo condition,” Michiel van Elk, a cognitive psychologist at Leiden University, told me. In the MDMA clinical trials, run by the pharmaceutical company Lykos Therapeutics, nearly all participants correctly guessed which group they were in.
Many scientists want to get around this problem by designing better blinds. Some labs have tried to keep patients in the dark by administering drugs under anesthesia or using mind-altering pills like methamphetamines as a placebo. Others are trying to engineer new psychedelics that skip the trip entirely. But to other scientists, clever attempts to stuff psychedelics into the RCT framework ignore the possibility that psychedelics’ benefits aren’t reducible to the biochemical action of the drug itself. Since the 1960s, psychedelic researchers have known that the beliefs and expectations a person brings to a trip can influence whether it’s healing or nightmarish. (That’s why most psychedelic-therapy protocols include several psychotherapy sessions before, during, and after treatment.) By striving to cleave the drug’s effects from the context in which it’s given—to a patient by a therapist, both of whom are hoping for healing—blinded studies may fail to capture the full picture.
Read: Psychedelics open your brain. You might not like what falls in.
From this perspective, high proportions of unblinding in positive psychedelic trials don’t necessarily mean that the results are invalid. “It’s how people engage with those effects and their therapist that’s contributing to the improvement,” Eduardo Schenberg, a neuroscientist at Instituto Phaneros, a nonprofit psychedelic-research center in Brazil, told me. Recent research backs this up. One small study found that among chronic PTSD patients who got MDMA-assisted therapy, the strength of the bond between therapist and patient—something the drug helps forge with its empathy-inducing effects—predicted treatment success. Given the importance of context, even the most perfectly designed RCTs may fail to capture how helpful these drugs are outside trials, Schenberg said.
Such failure, if it exists, might extend beyond psychedelics to several kinds of psychoactive drugs. For instance, a 2022 analysis found that many antidepressant trials fail to effectively blind participants, in part because of side effects. “We know that 80 percent of the treatment response from antidepressants can be attributed to the placebo response,” Amelia Scott, a clinical psychologist at Macquarie University who co-wrote that study, told me. Yet in practice, antidepressants are effective for many people, suggesting that RCTs aren’t quite capturing what these drugs can offer—and that limiting ourselves to treatments that can be perfectly blinded could mean ignoring helpful mental-health interventions. “We shouldn’t be afraid to question the gold standard,” Schenberg told me. “For different kinds of diseases and treatments, we may need slightly different standards.”
RCTs likely won’t lose their perch as the gold standard anytime soon, for evaluating psychedelics or anything else. But they could be supplemented with other kinds of studies that would broaden our understanding of how psychedelics work, Matt Butler, a neuroscientist at King’s College London, told me. Scientists are already trying open-label trials, where participants know which treatment they’re getting, and measuring expectations along with treatment effects. Descriptive studies, which track how treatments are working in actual clinics, could provide a richer picture of what therapeutic contexts work best. “These levels of evidence aren’t as good as RCTs,” Butler said, but they could help deepen our understanding of when therapies that don’t conform to RCTs could be most helpful.
Read: What if psychedelics’ hallucinations are just a side effect?
None of this is to say that Lykos’s flawed RCT data will be enough to convince the FDA’s advisers that Americans with PTSD should be offered MDMA. Several groups, including the American Psychiatric Association, have expressed concern about the trials ahead of the advisory meeting. In addition to the unblinding issue, claims that therapists encouraged participants to report favorable results and hide adverse events prompted the Institute for Clinical and Economic Review to release a report casting doubt on the studies. Lykos CEO Amy Emerson pushed back in a statement, saying, “We stand by the quality and integrity of our research and development program.” Still, some researchers remain worried. “If this sets a precedent that these trials are acceptable data, what does that mean for the future?” Suresh Muthukumaraswamy, a neuropharmacologist at the University of Auckland, told me.
The recent past suggests that blinding may not be a deal-breaker for the FDA. In 2019, the agency approved Spravato esketamine nasal spray—a version of ketamine—for the treatment of depression despite concerns about unblinding in the drug’s clinical trials. And the FDA worked with Lykos to design the MDMA-therapy trials after designating it a breakthrough treatment in 2017. In an email, an FDA spokesperson told me that blinded RCTs provide the most rigorous level of evidence, but “unblinded studies can still be considered adequate and well-controlled as long as there is a valid comparison with a control.” In such cases, the spokesperson said, regulators can take into account things like the size of the treatment effect in deciding whether the treatment performed significantly better than the placebo.
Read: Placebo effect of the heart
Even if the FDA is on board, rolling out psychedelic therapies before scientists fully understand the interplay among expectation, therapy, and drugs could mean missing an opportunity to force companies to provide data that would meaningfully advance the study of these drugs, Muthukumaraswamy said. It also risks undermining these treatments in the long run. If sky-high expectations are ultimately fueling the positive results we see now, the FDA could end up approving a treatment that becomes less effective as its novelty wears off. That’s especially true if we’re missing key components of what makes these treatments work, or what puts people at risk for bad experiences. To better answer those questions—for psychedelics and other psychoactive drugs—we may need studies that go beyond the gold standard.
Psychedelics Are Challenging the Scientific Gold Standard – The Atlantic
