The problem at the heart of modern psychedelic clinical research
How much of a problem is it if participants in psychedelic clinical trials can easily tell they have been given a placebo? New Zealand researcher Suresh Muthukumaraswamy suggests it is a big problem and many findings from recent psychedelic trials may be over-estimated.
For around half a century the randomized controlled trial (RCT) has been considered the gold standard in evaluating the effectiveness of any new medical intervention. The process involves recruiting a group of participants with a certain condition and randomly separating them into at least two groups.
One group receives the experimental intervention while the other group, known as the control, receives a placebo. To eliminate any bias the RCT would ideally also be blinded, meaning not only do the participants not know what group they are in but neither do the caregivers or researchers assessing the trial. All of this ultimately should lead to a robust trial that delivers clear empirical evidence as to whether a novel treatment fundamentally works.
One of the key reasons for blinding a RCT is to minimize the influence of the placebo effect, that phenomenon whereby inactive drugs result in positive therapeutic outcomes. While the impact of the placebo effect can vary depending on the condition, it has been consistently shown that patient outcomes can be significantly influenced when they are primed with a certain expectancy. If 10 people suffering from a mild headache were given a sugar pill and told it was a cutting-edge new drug it could be expected at least one or two would have a positive response.
Placebo effects can be positive or negative. Often excessive warnings about a drug's side effects can generate negative placebo responses. This so called nocebo effect can be seen in clinical trials where subjects in the placebo group still report side effects from an inactive drug.
For the most part, our modern RCT framework is a robust way to determine whether new drugs actually work, but what happens when you can’t effectively blind your trial? What happens when it is obvious to both participants and assessors whether someone has been given a placebo or not?
This is the problem at the center of the promising emerging field of modern psychedelic science. And according to Suresh Muthukumaraswamy, a psychopharmacologist from the University of Auckland, this is such a major issue that most modern psychedelic clinical trials are probably reporting over-estimated effect sizes.
Suresh recently co-authored a comprehensive article on the subject with colleagues Anna Forsythe and Thomas Lumley. While recognizing the promising research findings from psychedelic studies over the past 10 to 20 years the article presents a case for why many modern psychedelic clinical trials are flawed and suggests what can be done to deliver more rigorous results. Speaking to New Atlas about his latest article, Suresh, who is also running one of the world's first placebo-controlled LSD microdosing trials, is skeptical there is enough good evidence to show any of these current psychedelic medicines are ready for full public approval.
“On the basis of the evidence that’s been collected, at this point, I don’t see any reason why you would think that these drugs and interventions are approvable with the current issues they have in terms of design,” Suresh says.
The psychotherapy problem
The blinding and expectancy problem in many psychedelic trials goes deeper than participants being able to simply tell they have been administered the placebo, according to Suresh. Many modern psychedelic drug trials involve comprehensive accompanying psychotherapy protocols. These interventions are not simply claiming a single psilocybin dose will cure a person’s depression, but instead they often incorporate two or three day-long active drug sessions into a larger multi-month psychotherapy framework.
Therapy sessions both precede and follow the drug sessions in order to help a patient effectively integrate the psychedelic experience into a positive therapeutic outcome. And the entire treatment protocol can span anywhere from 6 to 12 weeks.
Knowing one is part of a placebo group is more than just a momentary disappointment in this context. It often means both therapist and patient still have to engage in the entire treatment protocol with full awareness they are not part of the active drug cohort.
“So you imagine someone getting a therapist who's potentially slightly motivated to treat this patient," explains Suresh. "And so they're sitting in a psychedelic therapy session with the patient. They don't know what they've been given, but the patient quickly realizes that they've been given placebo, and the therapist is sitting there going like, this person's been given a placebo. So is that therapist going to deliver as good a psychiatric intervention for the patient? Because one of the things we know about psychotherapy is that the build up of that relationship between the patient and the therapist is one of the strongest predictors of good outcome that we can have.”
Add to this the popular wave of media coverage heralding the revolutionary potential of psychedelic medicine and you have a perfect storm of bias and expectancy both enhancing the positive effects for those in the active group and amplifying the negative outcomes for those getting the placebo.
But how much of a problem is this really? According to Suresh there are hints in the literature quantifying how much this blinding problem could be influencing research findings. One of the clues comes from the growing body of research investigating the antidepressant qualities of ketamine.
Many of these ketamine trials use what is called an active placebo to try and help maintain blinding. A sedative known as midazolam is administered to the placebo group in the hopes that its mild psychoactive properties somewhat confuse some subjects into thinking they have been given ketamine. Metastudies comparing the results from ketamine trials using inactive placebos to those using midazolam suggests significant differences in efficacy when a subject is clearly aware they have been given an inactive placebo.
“When they compare ketamine to a saline placebo, the effect sizes that are reported are around D = 1.8. That's a massive effect size. It's just enormous. When they use midazolam as the active placebo, that effect size drops from 1.8 to 0.7," says Suresh. "That is an absolutely enormous drop.”
Suresh adds that midazolam is not even a strongly convincing active placebo, with many trial participants still effectively guessing which group they ultimately were randomized into. And while active placebos certainly may be a factor in helping deliver better results moving forward it could be challenging, if not impossible, to find an active placebo that is comparable to drugs such as LSD or psilocybin.
Despite these challenges Suresh does not suggest we need to abandon RCTs for psychedelic research … at least not yet. High dose/low dose groupings could be a way around the active placebo problem, offering extra insights into treatment outcomes for different doses. But in general he calls for more rigor and transparency in the ways psychedelic trials are designed and reported.
More detailed reporting of how effective the blinding was in any given trial is one suggestion. Another is more pre-publishing of trial protocols before the trials are actually conducted. Suresh says it is shocking how few psychedelic clinical trials are clearly pre-publishing study protocols and it is a massive problem leading to unsuccessful trials being buried, or worse, being spun with positive outcomes by cherry-picking effect measures.
Alongside this he calls for more clarity into exactly how trial designs are communicated to prospective subjects. Just describing likely outcomes to trial participants has been shown to influence treatment effects. In the past these factors have not generally been acknowledged in great detail, after all, if your blinding is strong it shouldn’t really make much of a difference.
But in the world of psychedelic science things seem to play out a little differently. A compelling study published last year from a team of Canadian researchers set out to try and experimentally quantify the psychedelic placebo effect by inventing a fictional psychedelic drug and staging a fake party.
Tripping on nothing
The study recruited a cohort of subjects under the pretense of testing the effect of psychedelic drugs on creativity in a natural environment. The subjects were introduced to a drug called iprocin, described as a fast-acting psychedelic similar to psilocybin.
The entire experience was staged to give participants the impression they were taking part in a serious psychedelic drug experiment, from amplified security procedures upon entering the research facility, to undercover research assistants wandering around the simulated party environment pretending to be other trial participants feeling the effects of the fictional drug.
The results were striking, with over 60 percent of subjects reporting feeling some kind of psychedelic effect. Only 35 percent of the cohort correctly guessed the drug was a placebo, and 50 percent of those who reported feeling a psychedelic effect actually had previous experience with psychedelic drugs.
One conclusion from the study was a call for psychedelic researchers to better describe the experimental conditions of any given trial. If the behavior of those overseeing psychedelic experiments can influence outcomes to such a significant degree then we certainly need more transparency in detailing these factors. It may be unconventional to describe these factors in a study reporting data from a RCT, but there is nothing conventional about psychedelic science.
Moving forward, Suresh is frank about the problems psychedelic researchers are facing. He is quick to point out the myriad of problems in his own prior research, often citing his mistakes as examples instead of referencing others. But he says the most important takeaway is now that we know what the problems in the research are, from here on in we should start tightening up practices and improving protocols.
“It’s a new field of science,” he says. “The first wave of stuff is likely to be flawed. So the question is how can we make it better? We may not be able to achieve perfection, but we can at least try harder.”
The new article was published in the Expert Review of Clinical Pharmacology.