The Placebo Effect: What Is It and How Does It Work?

Jane D. was a frequent visitor to our emergency service. She usually came late at night, asking for an injection of Demerol, a narcotic that was the only thing that relieved her unbearable headaches. One day, our psychiatrist suggested that the nurse give Jane a saline injection instead of Demerol. It worked! He explained to Jane that she had responded to a placebo, thinking this would help her understand the psychological nature of her problem. But as he was about to say goodbye, Jane asked, “Can I have that same medicine next time instead of Demerol? It helped me so much!”

The term “placebo effect” is somewhat misleading and causes confusion. The word “placebo” comes from Latin and means “I shall please.” You might think its opposite is “I benefit,” but the placebo effect is not about the mysterious action of a harmless medicine. Instead, it’s a complex web of psychosocial influences that arise during treatment. These influences occur with both active medications and inactive substances.

Dr. Mark Crislip believes the placebo effect is a myth. “I think the placebo effect in pain relief is a good example of cognitive behavioral therapy; the pain doesn’t decrease, but the patient’s emotional response changes… There’s no such thing as a placebo effect, only changes in sensation and perception.” He’s right that the placebo effect doesn’t alter pain signals in nerve tissue. But most people believe that changes in sensation are the result of the placebo and see value in it.

There’s a big difference between pain and suffering. Labor pains cause incredible pain, but a woman who knows the outcome will be positive suffers less than, say, a man with a bad leg who fears amputation. Some say morphine doesn’t so much remove pain as allow a person to ignore it. The perception and significance of pain are as important as the pain signal itself. If a placebo can distract a patient or help them reframe the meaning of their pain, their changed perception can reduce suffering.

Beecher Claims Placebo Has an Effect

In 1955, Dr. Henry Beecher published a landmark paper called “The Powerful Placebo” in the Journal of the American Medical Association. He reviewed studies comparing active treatments to placebos and found that, on average, 35% of patients improved with a placebo. This meant that any study not including a placebo group might give a false positive result. Placebo-controlled trials have since become the cornerstone of medical science. It’s not enough to show that a miracle cure works; it must be better than a sugar pill that looks the same.

Beecher’s work is often cited as proof that 35% of patients respond to placebos, but his results weren’t entirely accurate. He didn’t measure the pure placebo effect—instead, he looked at a mix of placebo effect, natural disease progression, and other factors. Some patients who seemed to respond to placebo were actually getting better for other reasons, such as spontaneous improvement, symptom changes, or even just wanting to please their doctors.

Hróbjartsson Says Placebo Doesn’t Work

To determine how much of the 35% improvement was due to placebo, we need to know how many untreated patients would have improved anyway. In 2001, Danish researchers Asbjørn Hróbjartsson and Peter Gøtzsche published “Is the Placebo Powerless?” in the New England Journal of Medicine. They analyzed studies comparing placebo treatment to no treatment and found little evidence that placebos have significant therapeutic effects.

In experiments using a simple “improved/not improved” scale, there was no significant difference between placebo and no treatment. Some placebo effect was seen in subjective symptoms like pain, but not in objective measures like blood pressure. The effect was more noticeable in small, biased, or subjective studies.

But Placebos Do Work, Don’t They?

It’s hard to reconcile these findings with what we know about placebos, like Jane D.’s case. In studies of pain after dental surgery, patients were given either morphine or saline. If told the saline was a powerful new painkiller, they felt as good as those who got morphine. In another study, all patients received morphine for post-op pain, but only half were told what they were getting. Those who didn’t know felt less relief. In acupuncture studies, patients were divided into “real” and “placebo” groups, but the procedures were identical. Those who believed they had real treatment reported more improvement, even though everyone was treated the same.

We not only know that placebos help, but we can rank their effectiveness:

• Placebo surgery works better than placebo injections.
• Injections work better than pills.
• Sham acupuncture is better than placebo pills.
• Capsules are better than tablets.
• Large pills are better than small ones.
• The more often a person takes a placebo, the more effective it is.
• The more expensive, the better.
• Pill color matters.
• It’s better to tell a patient, “This will help you,” than “It might help.”

In one study, patients were given the same aspirin from bottles with and without labels. Treatment was more effective when patients saw a labeled bottle and recognized the manufacturer. Our pharmacy had two types of allergy pills from the same maker, differing only in color—green and blue. When a patient said the green pills stopped working, we gave them the blue ones, and they got better.

Nocebo Effect and Subjectivity

Alongside the placebo effect is the nocebo effect (“I will harm”). People taking inert substances often report new symptoms. A friend of mine stopped taking a homeopathic remedy for insomnia because she thought it caused side effects (homeopathy is pure placebo, as it contains only water). In a study of postmenopausal hormone therapy, 63% of women on hormones and 40% on placebo reported withdrawal symptoms. If you tell patients a drug may cause nausea, they’re more likely to feel it.

The placebo effect is mostly subjective. It doesn’t work if patients are asleep or unconscious—you have to know you’re being treated. Placebos won’t replace birth control, cure cancer, heal fractures, or change anything objectively measurable. They work for subjective complaints like headaches, depression, itching, shortness of breath, tension, indigestion, and other self-reported symptoms.

This doesn’t mean such symptoms are imaginary. Some doctors have tried to use placebos to see if a patient is “really” sick. This doesn’t work and is unethical.

Some researchers believe placebos can have objective effects. For example, when doctors painted skin growths with inert dye and told patients they’d disappear when the dye wore off, the growths did vanish. Patients with new pacemakers improved even before the devices were turned on. Asthmatics’ airways widened when told they were given a bronchodilator. Colitis patients on placebo not only reported feeling better, but sigmoidoscopy showed reduced inflammation. Ulcer patients recovered faster when given two placebo pills instead of one.

According to Harvard researcher Herbert Benson, the placebo effect stimulates healing in 60–90% of illnesses, such as angina, asthma, herpes, and ulcers. Some uncritical studies suggest placebo can affect tumors, movement disorders, temperature, pulse, blood pressure, cholesterol, blood sugar, and exercise tolerance.

However, evidence for objective placebo effects is weak. Another theory is that the placebo effect is subjective but can indirectly influence objective symptoms. For example, if you’re in pain and a placebo reduces your sensitivity, your pulse and blood pressure may drop. If you’re an asthmatic and feel calmer, your breathing may improve. Thus, the placebo effect doesn’t directly cause objective changes but alters your perception, which can indirectly affect your physiology. It’s a kind of meaning-making.

Placebo Surgery and Its Implications

Placebo surgery is another topic of debate. Forty years ago, Seattle cardiologist Leonard Cobb performed a popular angina surgery where doctors made chest incisions and tied off two arteries to increase blood flow to the heart. Ninety percent of patients said it helped. But when Cobb performed sham surgery (incisions only, no arteries tied) and compared results, the fake operation was just as effective. The procedure was soon abandoned.

In a recent study, sham arthroscopic knee surgery was compared to real surgery for osteoarthritis. Patients who only had incisions recovered just as well. One was told about the sham surgery but still believes it saved him. Some question whether these experiments prove surgery has a placebo effect—maybe the surgery wasn’t needed, and patients would have recovered anyway.

How Does the Placebo Effect Work?

If the placebo effect is real, what’s the mechanism? We can’t just blame it on suggestible patients. Several explanations exist: expectation, motivation, conditioned reflexes, and endogenous opioids.

• Expectation is a proven psychological phenomenon that even affects visual perception—we’re more likely to see what we expect. Wine tastes better if it’s expensive. Kids prefer food from McDonald’s packaging. If we expect pain, we’re more likely to feel it. If told we’re getting a strong painkiller, we’ll likely feel relief sooner.
• Motivation—the desire to get well or be free of pain—directly affects recovery with placebos. Highly motivated patients are more compliant and follow doctors’ advice more closely. Those who agree to take placebo pills regularly develop a stronger response.
• Conditioned reflexes were studied by Pavlov in dogs. People associate pills and medicine with recovery. The body even develops physiological responses: dogs salivate when given morphine; after conditioning, a placebo injection causes similar, though weaker, symptoms.
• Endogenous opioids are natural painkillers produced in the brain, similar to opium drugs like morphine. There’s evidence that placebo responders produce more of these chemicals. Brain scans show opioid receptors activate when people are told a placebo is a painkiller. Drugs that block narcotics can neutralize the placebo effect.

Dopamine levels in Parkinson’s patients’ brains rose after taking a placebo; those who reported feeling better also had higher dopamine. In another brain scan study, patients played a game and were told their chances of winning. The same brain areas lit up in those confident of winning as in people who believed in the placebo. In a Coke vs. Pepsi experiment, information about the brand was processed in a different brain area than taste. We may be getting closer to understanding how expectation, optimism, and prior experience are organized in the brain. There may be genetic differences or differences in dopamine receptor sensitivity. Brain imaging is giving us lots of data, and making sense of it is a key challenge for future science.

Can Animals Respond to Placebos?

Supporters of homeopathy and acupuncture claim animals respond to these treatments but not to placebos, suggesting placebo only works in humans. But veterinary manuals say animals can respond to placebos, and there are reasonable explanations:

• Animals can develop conditioned physiological responses to drugs, which also occur when the drug is replaced with a placebo.
• They respond to human attention and care.
• Handlers may misinterpret animal behavior as a placebo effect.
• Since animals can’t talk, we judge their pain by behavior, which isn’t always reliable.

Ethical Considerations

Some people don’t care whether they’re getting a placebo or not. If they feel better, nothing else matters. In “Snake Oil Science: The Truth About Complementary and Alternative Medicine,” R. Barker Bausell says the main benefit of alternative therapies is the placebo effect, often enhanced by rituals and pseudoscientific jargon.

If we give a patient a placebo and they improve, what’s wrong with that? If a little deception helps, why not lie? The problem is that it’s unethical for doctors to lie or prescribe inactive drugs, and deception can permanently damage doctor-patient trust. As Bob Carroll notes in “The Skeptic’s Dictionary,” the placebo is a direct path to quackery.

A recent Danish survey found that 48% of doctors admitted to using placebos about 10 times in the past year, prescribing things like antibiotics for viral infections and vitamins for fatigue. Specialists and hospital staff used placebos less often. A 2004 Israeli study found 60% of doctors prescribed placebos to “deflect” patient requests for unproven drugs or to calm patients.

What if doctors are honest about placebos? If they tell patients, will the effect still work? Maybe not. After medical experiments, patients told about the placebo asked if they could keep taking it. In another case, patients were told a pill was inactive and would only “reduce the dose” of an active drug; they agreed and reduced their medication. What if a doctor says, “This remedy isn’t recognized by science, but it’s helped many people”? Placebos raise many ethical questions.

Recently, a court case involved a device with “Q-rays.” The makers admitted it was a sham, but argued that since the placebo effect exists, they could legally sell it. The court disagreed.

Is there an ethical way to use the placebo effect? Absolutely. Doctors already do. The placebo effect is an essential part of any doctor-patient interaction. Good doctors inspire trust, give hope, and instill confidence.

It’s not the placebo itself that’s effective, but the meaning of treatment. We enter a relationship with a caring person who promises to help. The prescription may symbolize that care. We may develop a conditioned expectation of recovery because this person has helped us before. We’re told why we’re sick and what to do to get better. We’re given hope, support, kindness, and attention. All these factors can trigger real physiological responses—our pulse drops, we relax, stress hormones decrease, and many other changes occur that promote healing or at least improve well-being. One study found patients recovered faster if their window view was of trees rather than a brick wall. Even if it’s not true, it’s nice that patients had something pleasant to look at. Even if we can’t measure changes, quality of life matters.

Placebo is present in effective treatment. A significant percentage of antidepressant effectiveness is due to placebo. Morphine works even better if the doctor tells you it’s a strong drug.

We shouldn’t separate the placebo effect from traditional medicine. That’s a mistake. As neurologist Robert Burton put it, even with today’s medical knowledge, treatment for most common illnesses—from back pain to the common cold—relies primarily on patient support and belief, not drugs. We need to rethink the meaning of the placebo effect and learn how to use it safely, affordably, and without deception.