Can Placebos Work—Even When Patients Know They’re Fake?

Researchers showed that a saline spray “treatment” reduced people’s emotional distress, even though the study subjects knew the spray wouldn’t do anything.
Brown pill standing out from crowd of capsules on green background
Photograph: Larry Washburn/Getty Images

A quick spray into the nose, a small sugar pill, or a fake surgery—placebos have taken many forms throughout history, and they have shown that expectation may produce results as powerful as a drug. But there has been one constant throughout: deception. A doctor or therapist has always had to lie to their test subject or patient about which version of a treatment they are getting.

“I used to say deceptive statements,” remembers Luana Colloca, a physician scientist and associate professor at the University of Maryland who studies pain modulation and placebo use and has used placebos in her own studies. “And every time I felt this dilemma inside me. I felt like, ‘I don't want to lie to people.’”

Some ethicists argue that using placebos in clinical trials, or prescribing patients a placebo intervention, violates fundamental principles of medicine. It means withholding standard treatment or an experimental treatment that could help patients get better. Using placebos in laboratory studies or as treatment is directly contrary to informed consent, since participants are not thoroughly informed, says Franklin Miller, a professor of medical ethics at Weill Cornell Medicine. “I'm OK with deception if it’s methodologically necessary to answer a valuable scientific question,” Miller says. But he suggests alerting subjects to the possibility that they will be deceived at some point and obtaining informed consent, called “authorized deception.” Or … researchers could get rid of deception altogether.

Within the last decade, researchers have been studying the benefits of doing just that—giving subjects a placebo while disclosing that they’re getting one. In a breakthrough 2010 study, researchers at Harvard Medical School found that non-deceptive placebos, also called open-label placebos, reduced symptoms in patients with irritable bowel syndrome. In the years since, non-deceptive placebos have been found to be effective in more than 20 conditions, including back pain, test anxiety, and cancer-related fatigue. And now, an intriguing new study from researchers at the University of Michigan, published in Nature Communications in July, shows that when researchers are transparent about placebo use, they’re able to reduce emotional distress in people put in an upsetting situation.

Lead author Darwin Guevarra, who is now a postdoctoral researcher in at Michigan State University's psychology department, has long been fascinated with placebo literature, including a 2013 pilot study finding that non-deceptive placebos mitigated symptoms of major depressive disorder. To him, it suggested that this could also be a way to address other mental health disorders. So Guevarra decided to test placebos in the context of emotional regulation, or the ability to control an emotional response, which is dysregulated in many mental health conditions. “Emotions robustly respond to [expectations]. That's why we thought that this was the best domain to look at,” he says. He wanted to find out if non-deceptive placebos would be able to reduce emotional distress, and if so, to find a corresponding brain mechanism that could explain what was happening.

In the first phase of his study, Guevarra and his colleagues recruited 62 healthy college students: One group of participants read about the placebo effect and then were told they would receive a placebo, while another group read about pain management and were kept in the dark about their treatment. Next, the researchers sprayed a saline solution into each person’s nose. For the non-deception group, the students were told prior to receiving the saline solution that it was a placebo. In the group that was deceived, they were told the saline solution would help obtain better physiological recordings. (It wouldn’t—the spray actually does nothing.)

Next, the students looked at distressing images on a screen, such as a person with an open injury. Immediately after seeing each picture, participants rated how it made them feel, with 1 being not at all negative and 9 being very negative. Overall, those who had been told they received a placebo felt less distress in response to seeing the pictures. The non-deceived group rated their distress lower—a 6 on average—than those that had been deceived, who on average reported feeling around 7.5 on the distress scale.

But this study had a twist. Most previous studies only used self-reported measures like questionnaires to ask subjects to rate how depressed, anxious, or stressed they feel. These are subjective evaluations, which may be influenced by response bias, in which participants do not accurately report what they are feeling. Instead, Guevarra’s team wanted to obtain an objective neural measure about what happens in response to non-deceptive placebos. Again, they recruited 198 healthy college students and gave them the saline spray placebo, with and without deception. This time, while showing the participants the negative pictures, they measured the electrical activity in each person’s brain using electroencephalography. EEG measures the electrical signals emitted from the entire brain, as recorded from electrodes stuck to the participant’s head.

Specifically, they looked at an indicator of emotional response called the late positive potential (LPP), an increase in electrical activity as a response to an emotional stimulus, such as distressing images. Strategies such as positive reappraisal, in which a person reinterprets a negative experience in a positive manner, can suppress this activity, which has led researchers to believe that LPP can be used to track emotional regulation—in this case, it would show the brain responding less intensely after an upsetting or emotional event.

When the researchers looked at the EEG readings from the two groups, they saw that the magnitude of the LPP of participants in the non-deceptive placebo group was smaller than those that had been deceived, meaning that their brains responded less to the distressing images than the other group. Three seconds after they were shown the distressing image, participants who received the non-deceptive placebo had an LPP amplitude of about 0.5 microvolts, while those that did not had an amplitude of about 3 microvolts. An analysis of this difference revealed that non-deceptive placebos had a moderate effect on the LPP over the control conditions, suggesting that they can modulate and dampen early neural reactions to emotional distress. In other words, the spray worked—even though participants knew it was a fake.

For Guevarra, this was evidence that the effect of the placebos was not response bias, but rather a real change in the brain. “I think it's a genuine psychobiological effect,” he says. “The manipulation we have really fine-tunes and leverages people's expectations.”

How might this work translate to the real world of mental health treatment? While the idea is still theoretical, Guevarra feels that non-deceptive placebos might be tried for conditions that consistently respond to expectations, such as anxiety, depression, and pain—and for mild to moderate cases. He envisions them being used by therapists as a first, cost-effective step or as co-interventions, given along with established treatments such as antidepressants and cognitive behavioral therapy, a type of talk therapy that has become an important tool in psychology. “Let's give them placebo pills first and see how it goes,” he says.

If the placebo doesn’t work, then they could move on to other alternatives. “The beauty of this is that it’s relatively low-cost and arguably side effect-free,” says University of Michigan psychology professor Ethan Kross, the principal investigator on the study.

Other scientists who study placebos called the study fascinating and said they are hopeful that it will stimulate further research, but cautioned that much more needs to be considered before this work could be extrapolated to clinical settings. For example, the study only observed a very short-term response among the subjects. Although EEG is a reliable approach to measure early brain responses, Miller says, seeing an effect for the first few seconds of an emotional response in healthy people makes it hard to interpret what that could mean for treating a long-term psychiatric condition. “I see no way you can make any inferences from that kind of short-term outcome,” he says.

For that reason, Tor Wager, a professor of neuroscience at Dartmouth College and a co-author of the study, says it will be important to track changes in the long run. “We need to know something about which placebo effects last a long time, which ones fundamentally change something in your brain,” he says. He suspects that different “ingredients”—like reinforcing belief in the effectiveness of the placebo at a particular time—could make placebos more or less durable. Ultimately, he says, the way to discern this would be to track whether non-deceptive placebos implemented through different strategies really affect a person’s long-term behavior and decision-making in a way that contributes to their life.

It will also be important to observe the response in different brains areas using functional MRI, says Fabrizio Benedetti, a professor of neuroscience at the University of Turin. EEG gives readings of the whole brain, but looking more granularly at the brain using fMRI “allows us to identify specific regions involved in a given effect,” Benedetti wrote in an email to WIRED.

Additionally, Guevarra’s subjects did not include people with diagnosed mental health conditions. Would the effect he observed look the same in clinical populations as it does in a healthy one? It’s hard to say. In her own work, Colloca has found that the placebo effect is identical in healthy participants and chronic pain patients. But mental health conditions may be more complex to address, which makes her suspect that these findings may be harder to replicate. The placebo effect may be influenced by “the way we perceive context around us and [how] learning experiences shape our own vision in conditions like major depression, anxiety, generalized anxiety or schizophrenia,” says Colloca. “I think that is an area that we didn't thoroughly explore yet and may be fascinating.”

As with any other possible intervention, non-deceptive placebos would have a lot of obstacles to overcome should they make it from lab to clinic. The effects would have to be shown to work in a larger, more diverse population, especially since the placebo effect itself has been shown to be different across people of different races, ages, and genders. And as Colloca points out, practitioners in fields such as psychology and pain management are not at the stage of prescribing non-deceptive placebos. “You can’t go to the pharmacy and say to your physician, ‘I want a placebo’. We are not there yet,” she says.

In fact, some scientists raise ethical caveats about the use of placebos in clinical research. Benedetti worries it may fuel pseudoscience, and could increase what he calls “pseudotreatments.” In a commentary published last year, he wrote that because scientists have shown how powerful expectation can be, people may be led to believe that anything—be it talismans, bizarre rituals, or even water—can be used to boost expectations and trigger the brain mechanisms that control the placebo effect.

Benedetti argues that there must be careful considerations about whether or not to bring non-deceptive placebos into the clinic. If they are, he says, the science cannot be oversold and has to be presented carefully. Not all conditions respond to the placebo effect, and people should not avoid necessary treatment, Benedetti points out. “The psychological component of some illnesses can indeed be modulated by placebos, but placebos cannot stop cancer growth nor can they kill the bacteria of pneumonia,” he wrote by email.

Still, even if non-deceptive placebos are not able to overcome enough obstacles to be of use in treatment, Kross says understanding them could provide insight into how expectations influence the brain. “Just learning about this has potential information and value. Knowing, for example, how powerful our expectations can be for influencing how we think, feel and behave, not just subjectively, but also physiologically,” he says.

“It's an experimental window into studying how those things work,” adds Wager, “how information, suggestions, and cognitive context ends up controlling physiology, neurochemistry and outcomes that we care about, like chronic pain.”

Non-deceptive placebos will need more robust studies to prove their efficacy before anyone will be willing to use them as a clinical treatment. Guevarra and Kross are now looking to test whether non-deceptive placebos could be used to help alleviate the increased stress that the Covid-19 pandemic has caused. They’re also looking to see how non-deceptive placebos compare to deceptive ones, as well as other emotional regulation strategies, in terms of tempering distress.

And not only will researchers have to prove their value to medical professionals, but patients will have to buy in, too. After all, belief is a crucial component of the placebo effect. “That’s the next part: crafting really persuadable, digestible, easily accessible manipulations,” says Guevarra. “To convince people that this could possibly work.”

Update 10-20-2020 3:37 PM: This story was updated to correct the name of Darwin Guevarra's current university.

Update 2-26-2021 8:34 PM: This story was updated to correct the description of Franklin Miller's stance on placebo use.


More Great WIRED Stories