The first real, physical proof of the placebo effect came in 1978. Running a study on patients who had recently had their wisdom teeth removed, researchers gave some of them a placebo. (The other group got naloxone, a drug that blocks the brain’s natural painkillers.) When some patients from the placebo group were later given naloxone as well, the drug counteracted the pain relief they initially got from the dummy drug.

But the effect wasn’t universal—the naloxone only made a difference among people who had reported pain relief in the first place, a group the authors called “placebo responders.” Among the non-responders, the naloxone didn’t cause any additional pain, leading the researchers to conclude that placebos worked in responders by activating the brain’s natural painkillers.

In partially answering one question—now, finally, scientists had some idea about the underlying mechanism of the placebo effect—the study, published in The Lancet, also raised a new one: Why does the placebo effect work in some people and not others? Or, put another way, what did the responders have that the non-responders didn’t?

The answer is complicated by the fact that “the placebo effect” is more accurately a tangle of interrelated effects, all of which are still closer to theory than to scientific fact. Some studies have shown that personality traits, like thrill-seeking and altruism, may have something to do with it. Other scientists say that the ritual surrounding the medicine—the explanation, reassurance, and simple human connection that a doctor can provide—may be as powerful as actually taking the sugar pill, or the saline solution. (In some cases, the placebo has done its job even when the patients knew full well that it wasn’t the real thing.)

And now, scientists are arguing that genetics may play a role, too. In a paper published today in the journal Trends in Molecular Medicine, researchers from the Program in Placebo Studies (a joint venture of Harvard Medical School and the Beth Israel Deaconess Medical Center) outlined the case for genes’ influence on the placebo effect.

“Because there’s so much big data out there now, we can really start to see patterns associated with people who respond to placebo,” said Kathryn Hall, a researcher in the placebo-studies program and the paper’s lead author. Some of those patterns are linked to levels of molecules like dopamine and serotonin, known as neurotransmitters, which “send messages to the brain to say ‘This feels good’ or ‘Look out for this’ or ‘I’m in love’ or ‘I’ve got to fight,’” she explained. The ways in which these molecules move through the brain (what the researchers call “neurotransmitter pathways”) can vary depending on a person’s genes—which means, the researchers believe, that genetic variation may also affect how the brain handles placebos. “We’re finding that the placebo response seems to be modified by several of these pathways, leading people to think that there’s more than one type of placebo response,” Hall said.

They got their first piece of evidence in 2012. Hall and her colleagues knew that a gene called COMT, which regulates the amount of dopamine in the brain, is linked to pain and to the good feelings that come with reward—people with the high-dopamine mutation are more sensitive to both. They also knew from past brain-imaging studies that when a placebo is working, “the parts of the brain that light up are related to reward, and to pain, and to the alleviation of pain,” she said. They wondered about a connection: Could the COMT gene be a factor in placebo responsiveness?

To test their theory, the researchers ran a study on patients with irritable-bowel syndrome (IBS), published in the journal PLOS One. Participants were divided into three groups: In one, volunteers were told they were on a waiting list for treatment; in another, they were given a fake acupuncture treatment (with needles that didn’t actually pierce the skin) by a cold, seemingly distant practitioner; and in a third, they received the fake acupuncture from a practitioner who appeared invested in their progress, showed concern, and spoke confidently about the treatment. A few weeks after treatment, the researchers asked participants to rate how much their IBS symptoms had improved, then tested them for the COMT gene. Those with the high-dopamine version of the gene, they discovered, were more likely to report that the placebo treatment had relieved their symptoms, an effect that was even stronger in the group that had received their treatment from a caring provider.

“It is unlikely that a single locus like COMT fully accounts for a complex behavioral phenotype like placebo response,” the authors wrote, but it seemed that it was at least a part of the larger puzzle.

But as researchers continue to piece that puzzle together, the authors of the new Molecular Medicine paper argue, they also face a new set of ethical concerns. “The immediate [issue] is autonomy—you shouldn’t be tested if you don’t want to be tested. And how is that information shared?” Hall said. “You can build any kind of scenario with that, but can you imagine if insurance companies give you [fewer] drugs because you’re a placebo responder? … Genetic information that might affect how people treat you or interact with you needs to be managed in a really careful way to ensure the patient’s safety.”

Another issue is who should own that information in the first place. Out of necessity, a doctor—as the one to facilitate the genetic testing—would be the knowledge gatekeeper when it came to a patient’s placebo sensitivity. Bioethical standards, though, dictate that if doctors use placebos on their patients, then the patients have to be fully informed every step of the way—even if it might lessen the effects. The American Medical Association warned in a 2006 report that “in the clinical setting, the use of a placebo without a patient’s knowledge may undermine trust, compromise the patient-physician relationship, and result in medical harm to the patient.”

On the other side of those ethical pitfalls, though, the study of placebo genetics also holds promise: for scientists hoping to better understand the results of clinical drug trials, and for helping doctors to better tailor treatment plans to each individual patient. “As we think about personalized medicine … I think it’s important that we don’t forget the element that comes from the placebo response,” Hall said. “In a way, [placebo] is the most well-studied treatment that has never been studied.”