Skip to Main Content

A doctor’s subtle facial cues — an encouraging smile, or perhaps a wince — can influence a patient’s own beliefs about whether a treatment will work and can even help induce a placebo effect, a new study suggests.

The study, published last week in Nature Human Behaviour, suggests that patients will report that a treatment works better when their doctor’s expression conveys their belief in the treatment, even when the medicine is a placebo. It offers a new potential explanation for how patients’ pick up on placebo effects — and raises questions about the ways in which factors beyond basic biology can influence people’s health outcomes.

“Our study focused on … how one person’s expectations can influence another’s,” said Dr. Pin-Hao “Andy” Chen, a postdoctoral fellow at Dartmouth College and one of the authors of the study. 


The study relied on 194 undergraduate students cast in the role of either patient or doctor, not on actual clinicians or their patients.

First, the researchers worked to condition the “doctors” to believe that one treatment worked better than the other. They showed those students two identical petroleum-jelly-based creams but told them one cream was “Thermedol,” a medicine designed and proven to help reduce the pain from extreme heat. The other, the researchers told the students, was just petroleum jelly. 


The researchers then applied heat to the doctors in a pain simulator with each of the creams on their skin and — without telling them — they dialed back the heat when the “Thermedol” was applied so that the doctors believed the cream “worked.” Sneaky, right? 

The researchers had the doctors apply the two creams to the remaining students, acting as patients, with strict instructions not to talk about the two medicines. 

This time, when the researchers used the pain simulator on the “patients,” they applied the same level of heat for both creams — but the patients still reported feeling less pain when they were being treated with the cream the doctors knew to be “Thermedol.”  

To figure out how the “doctors” transmitted their belief in the “Thermedol” cream to the “patients,” the researchers studied their faces closely with video cameras and used machine learning to suss out different cues, like lowering their brows, curling their lips, and wrinkling their noses. They concluded that those changes influenced the patients’ reports of pain. 

“Through their expressions, they influenced the patients to believe as they did,” Chen said.

Chen admitted that as exciting as the results from this experiment are, there’s still a lot more work to be done: “We’ve only just cracked the window open.” The study explains that there were many outside factors, like a patients’ prior experiences or body language cues, that also could have influenced the results. 

Chen hopes to continue this research to find more evidence that can eventually be used to train physicians about better ways to interact with their patients. Chen believes that further studies could eventually be used to encourage physicians to make positive expressions when they prescribe medicines.

Chen isn’t the only researcher studying placebo effects who seemed encouraged by the study results. 

“These are very convincing controlled results,” said Dr. Luana Colloca, a professor at the University of Maryland who was not involved in the Dartmouth study. “The experiment is very creative and fun, and the results are interesting.”

Colloca said the study is a big win for the study of socially transmitted placebo effects, and not as big of a win for any hopes to change physician training. 

“There are still a lot of variables to consider before we can say we have tangible proof that this something that needs to change in physician training,” Colloca said. 

One of the variables Colloca pointed out was how much the patients could have influenced one another nonverbally, beyond facial cues, like hand gestures or general body language. “There’s no comparative research about the impact of patient-to-patient social impact and how that could’ve affected the study,” Colloca said. She wondered how much that could’ve played as a factor in the study results.  

Colloca elaborated with a personal story of how when she just went to go get her flu shot and the woman in front of her didn’t seem to express any kind of discomfort, across her whole body not just her face, upon receiving it which gave her a standard of how she would feel. When it was her turn, Colloca said she didn’t feel pain. “Now, was that because I saw the woman in front of me not experiencing pain? Or because I didn’t feel it?” Colloca said. “I don’t know.” 

“There’s still so much to explore and social learning is a topic of high interest right now,” Colloca said. “And I’m excited to see what Dr. Chen and his colleagues do next.”