clock menu more-arrow no yes mobile

Filed under:

The weird power of the placebo effect, explained

Yes, the placebo effect is all in your mind. And it’s real.

Javier Zarracina/Vox
Brian Resnick is Vox’s science and health editor, and is the co-creator of Unexplainable, Vox's podcast about unanswered questions in science. Previously, Brian was a reporter at Vox and at National Journal.

Over the last several years, doctors noticed a mystifying trend: Fewer and fewer new pain drugs were getting through double-blind placebo control trials, the gold standard for testing a drug’s effectiveness.

In these trials, neither doctors nor patients know who is on the active drug and who is taking an inert pill. At the end of the trial, the two groups are compared. If those who actually took the drug report significantly greater improvement than those on placebo, then it’s worth prescribing.

When researchers started looking closely at pain-drug clinical trials, they found that an average of 27 percent of patients in 1996 reported pain reduction from a new drug compared to placebo. In 2013, it was 9 percent.

What this showed was not that the drugs were getting worse, but that “the placebo response is growing bigger over time,” but only in the US, explains Jeffrey Mogil, the McGill University pain researcher who co-discovered the trend. And it’s not just growing stronger in pain medicine. Placebos are growing in strength in antidepressants and anti-psychotic studies as well.

“The placebo effect is the most interesting phenomenon in all of science,” Mogil says. “It’s at the precise interface of biology and psychology,” and is subject to everything from the drug ads we see to our interactions with health care providers to the length of a clinical trial.

Scientists have been studying this incredibly complex interface in great detail over the past 15 years, and they’re finding that sugar pills are stranger and more useful than we’ve previously imagined. The new science of placebo is bringing new understanding to why alternative treatments — like acupuncture and reiki — help some people. And it could also potentially allow us to one day prescribe smaller doses of pain drugs to help address the opioid crisis currently ravaging America.

Most instructively, the science finds that since we can’t separate a medicine from the placebo effect, shouldn’t we use it to our advantage?

There is no one placebo response. It’s a family of overlapping psychological phenomena.

Belief is the oldest medicine known to man.

For millennia, doctors, caregivers, and healers had known that sham treatments made for happy customers. Thomas Jefferson himself marveled at the genius behind the placebo. “One of the most successful physicians I have ever known has assured me that he used more bread pills, drops of colored water, powders of hickory ashes than of all other medicines put together,” Jefferson wrote in 1807. “It was certainly a pious fraud.”

These days, placebo — Latin for “I shall please” — is much more than a pious fraud.

As Ted Kaptchuk at Harvard, who is regarded as one of the world’s leading experts on placebo, put it to me in a recent interview, the study of the placebo effect is about “finding out what is it that’s usually not paid attention to in medicine — the intangible that we often forget when we rely on good drugs and procedures. The placebo effect is a surrogate marker for everything that surrounds a pill. And that includes rituals, symbols, doctor-patient encounters.”

And it’s not just one thing. “I see the placebo effect as a kind of loose family of different phenomena that are just yoked together by this term,” says Franklin Miller, a retired NIH bioethicist who has edited a volume on the subject. “Sooner or later we’ll get rid of the term,” he says, and talk more specifically about each of its components.

The family of placebo effects ranges from the common sense to some head scratchers. Let’s start off with the simplest.

1) Regression to the mean

When people first go to a doctor or start on a clinical trial, their symptoms might be particularly bad (why else would they have sought treatment?). But in the natural course of an illness, symptoms may get better all on their own. In depression clinical studies, for instance, researchers find around one-third of patients get better without drugs or placebo. In other words, time itself is a kind of placebo that heals.

Sugar pills and active drugs can both change the way patients report symptoms.

2) Confirmation bias

A patient may hope to get better when they’re in treatment, so they will change their focus. They’ll pay closer attention to signs that they’re getting better and ignore signs that they’re getting worse. (Relatedly, there’s the Hawthorne effect: We change our behavior when we know we’re being watched.)

But as we’ve seen, the placebo effect is more than just bias. There’s also:

3) Expectations and learning

The placebo response is something we learn via cause and effect. When we take an active drug, we often feel better. That’s a memory we revisit and recreate when on placebo.

Luana Colloca, a physician and researcher at University of Maryland, has conducted a number of studies on this phenomenon. And they typically go like this: She’ll often hook up a study participant to an electroshock machine. For each strong, painful shock, she’ll flash a red light on a screen the participant is looking at. For mild shocks, she’ll flash a green light. By the end of the experiment, when the participants see the green light, they feel less pain, even when the shocks are set to the highest setting.

The lesson: We get cues about how we should respond to pain — and medicine — from our environments.

Take morphine, a powerful drug that acts directly on neurochemical receptors in the brain. You can become addicted to it. But its analgesic powers grow when we know we’re taking it, and know a caring professional is giving it to us.

Studies show that post-operative patients whose painkillers are distributed by a hidden robot pump at an undisclosed time need twice as much drug to get the same pain-relieving effect as when the drug is injected by a nurse they could see. So awareness that you’re being given something that’s supposed to relieve pain seems to impact perception of it working.

Pain relief is stronger and more immediate when morphine is injected out in the open.
The Lancet Neurology

The research also suggests that fake surgeries — where doctors make some incisions but don’t actually change anything — are an even stronger placebo than pills. A 2014 systematic review of surgery placebos found that the fake surgery led to improvements 75 percent of the time. In the case of surgeries to relieve pain, one meta-review found essentially no difference in outcomes between the real surgeries and the fake ones.

There is such thing as the nocebo effect: where negative expectations make people feel worse. Some researchers think this is what’s fueling the gluten-free diet fad. People have developed a negative expectation that eating gluten will make them feel bad. And so it does, even though they may not have any biological gluten sensitivity.

4) Pharmacological conditioning

This is where things get a little weird.

Colloca has conducted many studies where for several days, a patient will be on a drug to combat pain or deal with the symptoms of Parkinson’s disease. Then one day, she’ll surreptitiously switch the patient over to a placebo. And lo and behold, they still feel healing effects.

On that fifth day, it seems the placebo triggers a similar response in the brain as the real drug. “You can see brain locations associated with chronic pain and chronic psychiatric disease” acting like there are drugs in the system, she says. For instance, Colloca has found that individual neurons in the brains of patients with Parkinson’s disease will still respond to placebos as though they are actual anti-Parkinson’s drugs after such conditioning has taken place.

The brain can learn to associate taking a pill with relief, and produce the same brain chemicals when drug is replaced with placebo.

What’s going on here? Learning. Just like Pavlov’s dogs learned to associate the sound of a bell with food and would start to salivate in anticipation, our brains learn to associate taking a pill with relief, and start to produce the brain chemicals to kick-start that relief.

This pharmacological conditioning only works if the drug is acting on a process that the brain can do naturally. “You can condition pain relief because there are endogenous pain-relieving mechanisms,” Miller says. Painkillers activate the opioid system in the brain. Taking a pill you think is a painkiller can activate that system (to a lesser degree).

And some studies do suggest that the placebo effect’s powers may possibly move beyond the brain.

Researchers have used flavored drinks to condition an immune response to placebo.

In a 2012 study, participants were given a sweet drink along with a pill that contained an immune suppressant drug for a few days. Without notice, the drug was swapped with placebo on one of the trial days. And their bodies still showed a decreased immune response. Their bodies had learned to associate the sweet drink with decreased production of interleukin, a key protein in our immune systems, which is produced in many cells outside the brain.

Results like these show “we are talking about a neurobiological phenomenon,” Colloca says.

5) Social learning

When study participants see another patient get relief from a placebo treatment (like in the electroshock experiment described above), they have a greater placebo response when they’re hooked up to the machine.

6) A human connection

Irritable bowel syndrome is an incredibly hard condition to treat. People with it live with debilitating stomach cramps, and there are few effective treatments. And doctors aren’t sure of the underlying biological cause.

It’s the type of ailment that’s sometimes derided as “all in their head,” or a diagnosis given when all others fail. In the early 2000s, Harvard’s Ted Kaptchuk and colleagues conducted an experiment to see if usually intangible traits like warmth and empathy help make patients feel better.

In the experiment, 260 participants were split into three groups. One group received sham acupuncture from a practitioner who took extra time asking the patient about their life and struggles. He or she took pains to say things like, “I can understand how difficult IBS must be for you.” A second group got sham acupuncture from a practitioner who did minimal talking. A third group was just put on a waiting list for treatment.

A caring provider can create a stronger placebo response than an apathetic one.

The warm, friendly acupuncturist was able to produce better relief of symptoms. “These results indicate that such factors as warmth, empathy, duration of interaction, and the communication of positive expectation might indeed significantly affect clinical outcome,” the study concluded.

Participants in the “augmented” condition — the one in which the caregivers were extra attentive — reported better outcomes at the end of the three-week trial, compared with both participants who received treatment as normal and those waiting for treatment.

This may be the least-understood component of placebo: It’s not just about pills. It’s about the environment a pill is taken in. It’s about the person who gave it to you — and the rituals and encounters associated with them.

What placebos can, and can’t, do

Placebos seem to have the greatest power over symptoms that lie at the murky boundary between the physical and psychological.

A 2010 systematic review looked at 202 drug trials where a placebo group was compared to patients who received neither placebo nor active drug. And it found that placebos seem to move the needle on pain, nausea, asthma, and phobias, with more inconsistent results for outcomes like smoking, dementia, depression*, obesity, hypertension, insomnia, and anxiety. (*Separate literature review on depression meds does find an effect of placebo compared with no treatment.)

“It seems like placebo taps into a family of psychological and brain processes that’s very much something we evolved for,” says Tor Wager, a University of Colorado Boulder neuroscientist who has co-authored many of the key papers on the neuroscience of placebo. “Take pain as an example. If you step on something sharp, there’s pain in your foot. Now, how should you respond to it? Well, if you are running from an attack, you don’t even want to feel that. You keep going.”

Another way to think about it: Placebos tweak our experience of symptoms, not their underlying causes.

A 2011 study elegantly illustrates this. In the experiment, asthma patients were randomly sorted into three groups: One group received an inhaler with albuterol, a drug that opens the airways. Another group got an inhaler with a placebo. A third group got “sham” acupuncture (meaning the needles were withdrawn before they touched the skin). A fourth got nothing. The study authors evaluated lung function on two metrics: self-report from the patients on their asthma symptoms, and an objective measure of lung functioning.

If you go by self-report, it looks like the placebo, albuterol, and sham acupuncture are all equally effective.

The objective measure, however, shows only the albuterol improved airflow. (FEV is a measure of lung function.)

Which isn’t to say that the self-reported improvement on placebo doesn’t matter. In many illnesses, patients would love a greater opportunity to ignore their symptoms.

“In all the objectively measurable illnesses, like cancer, even heart disease, there are components of it that are not [objectively measurable],” Kaptchuk says. And it’s those symptoms that are the prime targets to treat with placebo.

Placebo can only help symptoms that can be modulated by the mind. “There are real limits to what you can condition,” Miller says. You can’t, for example, condition the cancer-killing effects of chemotherapy. Our bodies don’t produce cancer-killing chemicals.

There’s evidence that placebos actually release opioids in the brain

Over the past 15 years, scientists have made some of their most interesting discoveries looking at how placebos have a powerful impact on the brain.

“When I first started studying placebo effects, it kind of seemed like magic — for some reason, your brain mimicked a drug response,” Wager says. “The biggest change in this field in the last 15 years is that neuroscientists are beginning to uncover the underlying neural mechanisms that create the placebo response.”

Placebos, researchers have found, actually prompt the release of opioids and other endorphins (chemicals that reduce pain) in the brain. Other findings:

  • Drugs that negate the effects of opioids — such as naloxone — also counteract the placebo effect, which shows that placebos are indeed playing on the brain’s natural pain management circuitry.
  • The periaqueductal gray matter, a region of the brain key for pain management, shows increased activity under placebo. Regions of the spinal cord that respond to pain show decreased activity under placebo, which suggests either the sensation of pain or our perception of it is diminished under placebo.
  • Patients with Alzheimer’s disease start to show a diminished placebo response. It’s probably due to the degradation of their frontal lobes, the area of the brain that helps direct our subjective experience of the world.

Our understanding of all this is far from complete, Wager says. For one, researchers still don’t completely understand how the brain processes pain. A lot of the brain regions implicated in the placebo response also play a role in emotions. So we don’t yet know if placebo is actually reducing our sensation of pain, or just our interpretation of it. (Also, as with a lot of neuroscience studies, a brain area might “light up” in an experiment, but it’s really, really hard to know what exactly is going on.)

“So really, what we should be concluding from those studies is something like ‘placebo affects the pain you report,’” Wager says. “What does pain mean to you? That’s a decision that’s made in your brain in different circuits, and that’s essential to placebo.”

You can tell people they’re taking a sugar pill for their illness, and they’ll still feel better

Kaptchuk has studied the placebo effect for decades, and something always bothered him: deception. Placebo studies have long relied on double-blind procedures. It ensures scientific rigor but keeps patients in the dark about what they’re actually taking.

“About five years ago, I said to myself, ‘I’m really tired [of] doing research that people say is about deception and tricking people,’” he says.

So he wanted to see: Could he induce a placebo response even when he told patients they were on placebo?

His own randomized controlled trials found that giving patients open-label placebos — sugar pills that the doctors admit are sugar pills — improved symptoms of certain chronic conditions that are among the hardest for doctors to treat, including irritable bowel syndrome and lower back pain. And he wonders if chronic fatigue — a hard-to-define, hard-to-treat, but still debilitating condition — will be a good future target for this research.

“Our patients tell us it’s nuts,” he says. “The doctors think it’s nuts. And we just do it. And we’ve been getting good results.”

Kaptchuk’s work adds a few new mysteries to the placebo effect. For one, he says that the placebo effect doesn’t require patient expectations for a positive outcome to work. “All my patients are people who have been to many doctors before. They don’t have positive expectations about getting better,” he says. “They’ve been to 10 doctors already.”

Colloca has a different interpretation of his results. She says there’s a difference between belief and expectation, so while the patients may not believe the pill will work, they still unconsciously expect it to.

That’s because, she says, they still have a deep-seated conditioned memory for what it means to take a pill. They have a conditioned memory for what it means to be in the care of another person. And that memory is indeed an expectation that can kick-start the analgesic effect in the brain. They don’t have to be aware it’s happening.

Some doctors wonder if placebos can be integrated into mainstream medicine

The researchers I spoke to for this story are overall optimistic that these discoveries can be used in the clinical settings. There’s a lot of work left to do here, and certainly some of the findings are easier to implement than others. For instance, we could start with reminding doctors that they can relieve pain simply by being warm and caring to their patients.

Colloca wonders if the placebo effect can also be harnessed so that the millions living with chronic pain can feel the same therapeutic effects with a lower dosage of opioid treatments that are both ineffective and deadly.

The NIH’s Miller says it’s too soon to start prescribing placebos, or using the effect, to decrease the dosage of a drug. For one, most of these studies are short-term and conducted with healthy volunteers, not actual patients.

“There’s still lot we don’t know,” he says. Like side effects: Just as a placebo can mimic a drug, it can also mimic a side effect. “We haven’t done the kinds of studies that will indicate that you can maintain therapeutic benefit at lower side effect burden.”

More broadly, Kaptchuk says, for years researchers have seen the placebo as a hurdle to clear to produce good medicine. But placebo is not just a hurdle. “It’s basically the water that medicine swims in,” he says. “I would like to see the bottom line of my research change the art of medicine into the science of medicine.”