The Power of Placebo in Psychedelic Trials

It’s a story that launched a thousand research grants. A patient suffering from treatment-resistant depression – a stubborn, chronic loss of joy that has seen off multiple traditional interventions – takes one dose of a drug and their world opens up. Boris Heifets, a professor of anesthesiology, perioperative and pain medicine at Stanford University, recently ran a trial of the rapidly acting antidepressant ketamine. Heifets, who is also a by courtesy professor in Stanford’s department of psychiatry and brain sciences, is telling me about a patient on the trial. Their response to treatment was remarkable, but familiar to those seen in response to other psychedelic and psychedelic-adjacent drugs: “She says she’s never felt like this before; she's talking to old friends. I can only describe it as a psychological transformation.” To Heifets, this was a clear-cut example of ketamine’s power.

There was just one problem: the patient had been injected with plain old saline.

Achieving the double blind

Heifets’s trial, published in Nature Mental Health*, had a unique design. Virtually all clinical trials of novel antidepressants, like psilocybin, DMT and ketamine, dance around a nonsensical conceit. All split their patients into two groups – one receiving the active, game-changing treatment of the day, and the other a blank pill or inert injection. This study design, the randomized controlled trial (RCT) approach – is intended to rule out any contributions from the placebo effect by shielding participants from the knowledge of which treatment they received. Of course, the curtain drops away after the treatments are given. Heifets calls this “the moment of affirmation” for those given a drug, and the “moment of betrayal” for those in the control group, who applied to go on a trial with a revolutionary psychedelic and instead find themselves sitting on a couch, “trapped in their own head”.  Heifets’s approach promised the first fully “double-blinded” psychedelic study. There would be no affirmation or betrayal; patients would genuinely have no idea which treatment they received. That’s because they would be fully unconscious for the duration of the intervention.

The team hoped that this would answer a long-standing question in the field: is a conscious, subjective experience required to mediate these drugs’ effects?

Heifets’s work as an anesthesiologist gave him access to what he calls a “population of convenience” – people booked in for surgery that would be conducted under general anesthesia.

Working with such a vulnerable group meant patient selection was important – they recruited 40 participants, none of whom were undergoing brain or cardiac surgery, and all of whom had moderate-to-severe, treatment-resistant depression. Privately, Heifets had confident predictions about the experiment’s outcome. The placebo group, he had forecasted, “would do what placebo groups have done in anesthesia trials before, which is get worse.” This is due, he says, to the stress of an operation and the subsequent recovery. He was equally confident that patients given ketamine would get better without even knowing that they received it, as the body responded to neurobiological effects that the drug induces.

Looking at the data post-experiment, Heifets noted that the ketamine group did get better – much better. Their scores on the widely used Montgomery-Åsberg Depression Rating Scale (MADRS) depression scale halved on average – an improvement on par with that achieved in other ketamine studies with awake patients. The problem was that patients receiving placebo got better at the same rate.

Placebos: Everything but a blank pill

The placebo effect has fascinated and frustrated clinicians for decades. Henry Beecher’s influential 1955 paper, The Powerful Placebo, reflected that these inert substances have “doubtless been used for centuries by wise physicians as well as by quacks.” In the 20^th century, our ability to effectively treat disease raced forward, and the use of placebos to heal fell away. Instead, Beecher championed them as a tool that could tease out the pharmacological effects of a new drug. Chloé Pronovost-Morgan, a researcher at Maastricht University’s Department of Neuropsychology and Psychopharmacology, argues that in his enthusiasm to promote placebos, Beecher made a jump in logic that has sowed much confusion about what they actually are. “You still have people that are saying the placebo is an improvement caused by an inert pill,” she says.

Pronovost-Morgan and Heifets both explain the same thing: the real placebo effect, also called a non-specific effect, comes from everything but the pill. When a clinician sits down with a patient, the placebo effect takes in the room they are in, the tone of voice the doctor uses and the words they choose to describe the intervention. In a clinical trial, there are dozens of treatment-related factors that have nothing to do with what is in the pill being trialed – all of these come under the banner of placebo. Those effects add up and are found in many more trials than just the ones testing psychedelics. The placebo effect is “the single most reproducible effect in clinical medicine, period,” says Heifets. “When we talk about the placebo effect, that's really how we should conceptualize it –  something that is powerful and long-lasting,”

Gerard Sanacora, George D. and Esther S. Gross Professor of Psychiatry at Yale University, draws on a neat analogy to summarize placebo. “If for 40 years of your life you take an ibuprofen when you have a headache, and the headache goes away, at a certain point, you can take any pill and it's going to generate that same set of physiologic responses, that are going to cause your headache to go away.”

In many areas of modern medicine, however, placebo has become a dirty word. “In biomedical sciences, I think that there's quite a negative, pejorative connotation to the concept of placebo,” says Pronovost-Morgan. In a new review paper, co-authored with Maastricht Professor Jan Ramaekers, she argues that psychedelic science, and the way it has embraced non-specific effects, can provide a route to reassessing the placebo.

Set and setting

For thousands of years, psychedelics drugs were used as part of rituals by Indigenous peoples around the globe. The writers, researchers and free-love hippies that spearheaded the wave of Western psychedelic enthusiasm in the 1950s and 60s did so with little understanding of the unique culture in which these rituals originally occurred. What they did realize was the need to create a safe external setting and internal mindset of openness and relaxation for those embarking on psychedelic experiences.

These factors don’t change the drug itself – they are part of the placebo effect. Importantly, these effects are embraced, not rejected, by the field. “In psychedelics, set and setting are not taboo. It's just common knowledge, even within the lay public, that it is important to pay attention to these variables,” says Pronovost-Morgan.

These details are even more important in the context of the drugs’ effects on the brain. They're thought of as consciousness amplifiers, says Pronovost-Morgan: “They can put people in a more suggestible state where they're more responsive to the cues around them.”

The biomedical approach to placebo has been one of identification and elimination. As early as 1955, Beecher was totting up “total treatment effect” as a sum of specific and placebo effects. The difference between the two totals remains the key factor in assessments of RCT performance. At the same time, the field has been aware for decades that placebo and treatment effect interact. “It's possible that the placebo effect is enhanced if you realize that you are taking the active pill,” Pronovost-Morgan points out. “This means that the placebo effect would not be the same in the placebo group as in the treatment group. You can't simply pull them apart – because they're interacting.” When dealing with psychedelic drugs and their blind-breaking potential, that possibility is almost guaranteed.

Ramaekers and Pronovost-Morgan suggest that a different type of trial design that reflects this interaction might have a benefit for psychedelic research and beyond. This alternative model is called a “balanced placebo” design. Instead of two groups, there are four:

This type of study design is unusual: while not telling participants what drug they are getting is standard in clinical trials, outright deception is more controversial. Additionally, four-armed trials require more participants. When each patient goes through hours of aftercare, and each dose of the drug is heavily regulated and extortionately priced, it is little surprise that this approach is so rare.

“Missing the forest for the trees”

Heifets’s study provides what is perhaps a window onto how powerful the psychedelic placebo effect can be. What Heifets regrets is not measuring expectancy, another facet of the placebo effect with an outsize effect in psychedelics. His view is that these expectations of getting better are what drove his patients’ recoveries. His data hints at this – people whose scores recovered thought they had received ketamine, while non-responders assumed they had received a placebo. “Why would they think that, unless they had some prior belief about the value of ketamine?” Heifets points out. But without having asked patients prior to the trial how much they expected to benefit from participation, this theory cannot be causally proved.

In the absence of this neat explanation, alternative theories have sprung up from around the field. Perhaps ketamine’s inherent effects were blocked by general anesthetic, goes one idea, leaving only the expectation-driven benefit. Heifets says that this “doesn’t really fit the data.” If some of ketamine’s effect had been dulled, his patients’ recoveries might have been impaired in comparison to other trials of the drug. But the effects seen were just as strong. Additionally, if general anesthetic impaired ketamine’s action more generally, Heifets points out, the generations of anesthetists who used it every day probably would have noticed. “It has been studied for decades in other contexts, and somehow no one has ever complained about ketamine no longer being an opioid-sparing analgesic or being good for chronic pain when it's given under anesthesia.” It is possible that only the antidepressant qualities of the drug are blocked by general anesthetic, says Heifets, but it “isn’t a parsimonious explanation of the data.”

Other responses have, in Heifets’s words, “missed the forest for the trees.” One argument goes that Heifets’s paper has merely identified that general anesthetics themselves are antidepressants. There’s no evidence of such an effect being previously recorded with the doses and drugs Heifets used, but he wasn’t entirely surprised to hear these arguments. “Part of it is, I think, the cognitive bias created by our mental health system,” he says.

“It’s like there has to be a light switch somewhere; you give a drug, and you turn the depression on and off. If it wasn’t the ketamine, it must have been the propofol.” The medical model of mental health puts all the power in the pill.

What Heifets is most eager to say is that his data doesn’t suggest ketamine is ineffective. He points to a recent review authored by his postdoctoral fellow, Dr. Tuuli Hietamies. This looked at thousands of case reports of people visiting ketamine clinics and reporting recovery. What he suggests instead is that ketamine, like many other medical treatments, works by tapping into that placebo. Our medical model has put all the power in the pill, and that bleeds into patients’ expectations. As Sanacora sums up, “People hear that nonspecific effects are so important, but the corollary of that is, well, the drug doesn't really work. It's just all fake. That’s not the case. It's just that a large part of this effect is nonspecific, but you don't get any of it if you don't actually take the drug.”

 Cutting out placebo

That mindset has now extended into psychedelic science as well. A commentary paper was recently coauthored by some of the leadership team at Compass Pathways. This drug developer is leading a Phase 3 clinical trial of psilocybin-derived drug COMP360. The paper proposes that psychedelic drugs should be evaluated alone, without the trappings of psychotherapy. To Ramaekers, this argument is driven by a need to comply with the United States Food and Drug Administration (FDA) regulations: “They're changing their perspective because the FDA may make them change their perspective. The FDA is trained to develop drugs. They're not trained to understand how psychotherapy works. They have no means of evaluating psychotherapy.”

Heifets predicts that this approach will reduce the placebo effect, and with it, dull these drugs’ efficacies, risking a mighty drop-off in performance post-trial. That’s an effect seen across psychiatric drugs and is not unique to psychedelics. But given how central the powerful placebo has been to psychedelic clinical trials up until now, what remains might not be the revolution in mental health foreseen by so many. 

“As soon as you strip away everything else, all of the human attachment and support and the validation and the being seen, and all that stuff that we intuitively know is important but is very difficult to get intellectual property on and to sell and to reimburse, what you'll be left with is very transactional,” says Heifets. “Go to a clinic, take your trip, then leave and do the [therapy] app. The field is at a crossroads will it embrace placebo, using different trial designs and formal definitions of set and setting or remove these factors altogether.” Regardless of the route that the field decides to take, it will have to acknowledge the power of placebo. “I wish I could be more hopeful about decriminalization and commercialization of psychedelics,” he concludes. “I don't think they're going to reimburse the right things.”

*This article was updated October 20, 2023 to reflect that this study had now been published in Nature Mental Health.