In a new article in Lancet Psychiatry, researcher Matt Burke at the University of Toronto draws attention to powerful results in the placebo group in psychiatric trials. He notes that a recent brain stimulation study for depression found that more people improved in the sham (placebo) treatment group.
“We often ignore placebo effects as a nuisance… but the more we learn about the brain, the clearer it becomes that we have missed the boat,” Burke wrote on Twitter.
The brain stimulation study was also published in Lancet Psychiatry and was led by Richard Morriss at the University of Nottingham, UK. In that study, all participants had moderate to severe depression and had already taken antidepressants without success, and about three-quarters had clinical anxiety as well. Strikingly, 42% of the participants achieved full remission from depression in the sham (placebo) group, compared with only 30% in the group that received actual brain stimulation.
The difference was statistically significant—favoring the placebo group—in several analyses, including “in self-rated depression at 4 and 16 weeks, in generalized anxiety at four weeks, and in overall health at 16 weeks.” The results favored the placebo group in all of the analyses, even those that did not reach statistical significance.
Thus, the study showed that even after a failed trial of antidepressants, 42% of moderately to severely depressed people improved without receiving any actual treatment—and those who did receive treatment did worse.
According to Burke, the researchers in the brain stimulation trial “do not consider more deeply how the placebo response might have contributed to the beneficial effect of sham stimulation.”
However, although the brain stimulation article doesn’t use the word “placebo,” it includes a detailed discussion of possible reasons for the improvement in the sham treatment group, including regression to the mean, increased hope of improvement, the support of seeing researchers and clinicians, and that even the sham group involved having a relaxing and structured day, which may have similar effects to behavioral activation, a type of psychotherapy.
“The act of planning the day and devoting 1 h per day for personal relaxation for up to 8 weeks might have been of some clinical benefit. People with depression typically have low motivation to do activities that might be rewarding to them, and they put the needs of others before themselves. Structuring the day to do a relaxation task, such as Alpha-Stim AID, that increases a sense of control or pleasure might be construed as compatible with behavioural activation, an effective psychological treatment for depression. However, unlike behavioural activation, with Alpha-Stim AID there was no progressive increase in activities over time. Other reasons for the substantial decrease in depression over time in both groups might have been regression to the mean, benefits of support from the research team, or hope engendered by participating in a novel intervention trial.”
Despite these psychosocial explanations, Burke focuses on the notion that the placebo effect may rewire the very “brain circuits” theorized to be involved in psychiatric problems. He cites the example of Parkinson’s disease but does not cite any examples from the psychiatric diagnoses at play here—depression and anxiety—presumably because while Parkinson’s disease has known biological correlates, depression and anxiety do not.
“A crucial point is that placebo responsiveness does not make a symptom or disorder any less real or biological. Instead, placebo responsiveness means that pathophysiology likely overlaps with brain or brain-body circuits that can be modulated by placebo effects. Placebo-induced ventral striatum dopamine release resulting in symptomatic improvement of dopamine-deficient patients with Parkinson’s disease is a prime example.”
Another explanation, though, is that hope, support, and relaxing, structured activities have a psychological effect on psychological distress—an effect that perhaps does not need to be over-complicated by unsubstantiated theories about how these affect “brain or brain-body circuits.”
Additionally, as the original authors point out—and Burke does not mention—regression to the mean could be a key factor here. When you take a snapshot of people at their worst, it’s likely that another snapshot (say, at follow-up 16 weeks later) will show them doing much better. This is particularly true for psychiatric diagnoses like depression and anxiety, which are transient when untreated.
One of the issues that neither the article by Burke nor the original brain stimulation piece addressed is why the brain stimulation group did worse than the placebo on every measure. This implies that the treatment is doing more harm than good.
This is interesting in the context of antidepressant drugs as well, which have repeatedly been shown to be no better than the placebo in clinical trials, and which, in the long-term, have demonstrated worse outcomes in those who take them than in those who receive no treatment, even after controlling for baseline severity.
Burke goes as far as suggesting that failed trials—those that show a drug or device failing to beat the placebo—may not be “failures” at all. Instead, he suggests that perhaps the placebo effect and the drug/device have the same biological effect on the same brain systems!
“Psychiatrists should realise that treatment trials like Alpha Stim Anxiety Insomnia and Depression are not failing (in a literal sense) as both groups showed impressive and clinically meaningful improvements in their symptoms. The failure lies in the use of conventional measurements of efficacy that assume a categorical divide between the value of brain changes generated by placebo effects and brain changes generated by specific treatments.”
It would be fascinating to see a study investigating such an interesting speculation.
Ultimately, according to Burke, researchers need to investigate the placebo effect more thoroughly.
“A structured framework for appraisal and systematic collection of data relevant to placebo responses (e.g., patient expectancy and blinding integrity measures) should be established,” he writes.
Burke, M. J. (2023). A fundamental change is needed for appraising placebo responses in psychiatry. Lancet Psychiatry, 10(5), 316-317. DOI: https://doi.org/10.1016/S2215-0366(23)00068-8 (Link)
Morriss, R., Patel, S., Boutry, C., Patel, P., Guo, B., Briley, P. M., . . . & Kai, J. (2023). Clinical effectiveness of active Alpha-Stim AID versus sham Alpha-Stim AID in major depression in primary care in England (Alpha-Stim-D): A multicentre, parallel group, double-blind, randomised controlled trial. Lancet Psychiatry, 10(3), 172-183. DOI: https://doi.org/10.1016/S2215-0366(23)00007-X (Link)