The placebo effect is a genuine, measurable improvement in symptoms that follows an inert treatment, driven by expectation, conditioning, and the clinical ritual rather than any active drug. It is strongest for self-reported experiences like pain, nausea, and fatigue, and it leaves documented fingerprints in the brain.
Published: June 5, 2026. Last reviewed: June 5, 2026.
What is the placebo effect?
The placebo effect is a real, repeatedly measured change in a patient’s symptoms after an inert treatment, first quantified at scale by anesthesiologist Henry Beecher in his 1955 paper “The Powerful Placebo” in the Journal of the American Medical Association [1]. The word comes from the Latin for “I shall please,” a phrase that once described sham mourners hired for funerals.
Beecher reviewed fifteen clinical trials and estimated that 35 percent of patients were “satisfactorily relieved” by a dummy treatment [1]. That figure traveled far. It also turned out to be soft. In 1997, German researchers Gunver Kienle and Helmut Kiene reanalyzed Beecher’s source studies and found that most of the improvement could be explained by spontaneous recovery, regression to the mean, and other ordinary statistical artifacts rather than belief [2]. A larger 2001 review by Asbjørn Hróbjartsson and Peter Gøtzsche in the New England Journal of Medicine compared placebo against doing nothing across 130 trials and found little sign of a large clinical placebo effect, with one stubborn exception: continuous, self-reported outcomes like pain [3].
So the honest definition is narrower than the folklore. A placebo does not cure. It shifts how a symptom is experienced and reported, and that shift, in the right conditions, is large enough to measure against a control and trace inside the nervous system. This piece sits inside the wider study of science and natural anomalies, where the question is what the third significant figure actually shows.
A history under reanalysis
| Study | Year | Condition | Core finding |
|---|---|---|---|
| Beecher, JAMA | 1955 | Pain, anxiety, nausea | ~35% “satisfactory relief” from placebo |
| Levine, Gordon & Fields, Lancet | 1978 | Dental surgery pain | Naloxone cancels placebo relief |
| de la Fuente-Fernandez, Science | 2001 | Parkinson’s disease | Placebo releases striatal dopamine |
| Moseley, NEJM | 2002 | Knee osteoarthritis | Sham surgery equals real surgery |
| Kaptchuk, PLoS ONE | 2010 | Irritable bowel syndrome | Open-label placebo beats no pill |
How belief becomes biology
In 1978, rheumatologist Jon Levine and neurologist Howard Fields at the University of California, San Francisco showed that the opioid blocker naloxone cancels placebo pain relief, the first hard evidence that a sugar pill recruits real brain chemistry [4]. The implication was sharp. If blocking the body’s own opioids erases the benefit, then the benefit was chemical all along.
Placebo analgesia and the opioid system
Levine and Fields gave patients recovering from molar extraction either a placebo or naloxone under double-blind conditions. The naloxone group reported more pain, and the extra pain was concentrated entirely among the people who would otherwise have responded to placebo [4]. The natural reading is that expectation triggers the release of endorphins, the body’s internal opioids, producing genuine placebo analgesia. Fabrizio Benedetti, a neurophysiologist at the University of Turin and one of the field’s central figures, later showed the system is not all-or-nothing: some placebo pain relief runs through opioid pathways that naloxone blocks, and some runs through separate routes that it does not.
Dopamine and Parkinson’s disease
The cleanest picture comes from movement disorders. In 2001, Raul de la Fuente-Fernandez and colleagues at the University of British Columbia scanned Parkinson’s patients with positron emission tomography and found that a placebo injection triggered substantial dopamine release in the striatum, comparable to a dose of levodopa [5]. On the math: the patients were not imagining a smoother gait. Their damaged dopamine system was firing because they expected relief, and the expectation alone moved the chemistry. The same neurotransmitter that the disease destroys is the one that hope recruits.
Open-label placebos: honesty that still works
In a 2010 randomized trial of 80 irritable bowel syndrome patients, a Harvard team led by Ted Kaptchuk found that pills openly labeled as placebos still beat no treatment, with 59 percent reporting adequate relief versus 35 percent in the control group [6]. Patients were told, in plain language, that the capsules contained no active medicine and worked through mind-body processes. They improved anyway.
This result, published in PLoS ONE, broke the assumption that deception is the engine of the effect. On the IBS Global Improvement Scale the open-label placebo group scored 5.0 against 3.9 for controls, a gap unlikely to be chance (p = 0.002) [6]. A 2016 trial by Cláudia Carvalho and colleagues in the journal Pain repeated the pattern in chronic low back pain: adding honestly labeled placebo pills to usual care cut pain and disability with moderate to large effect sizes across 97 enrolled patients [7]. A 2021 follow-up led by Anthony Lembo and Kaptchuk found open-label placebo statistically non-inferior to concealed placebo in IBS, which is a strange and useful sentence to be able to write [8].
Why would a known sugar pill help? The leading account is conditioning plus ritual. Taking a pill twice a day is itself a learned cue, and the act of being treated, the appointment, the explanation, the care, carries therapeutic weight independent of the contents.
The nocebo effect: belief’s dark twin
The nocebo effect is the placebo’s mirror image, in which negative expectation produces real symptoms and side effects, a phenomenon Fabrizio Benedetti and others have tracked across migraine, surgery, and drug-withdrawal studies for more than two decades. The Latin here means “I shall harm.” Where belief in benefit can summon endorphins, dread can summon the opposite.
What the data rules out is the comforting idea that this only happens to the suggestible. In drug trials, the side effects reported by patients on placebo tend to mirror the warned-about side effects of the real drug in the same study, even though the placebo group swallowed nothing active. Benedetti’s work traces part of the nocebo effect to anticipatory anxiety and the brain chemical cholecystokinin, which amplifies pain. The practical cost is real: nocebo responses inflate dropout rates, push up the doses and sample sizes trials need, and can make a useful medicine look more dangerous than it is. Reading a long warning label, it turns out, is not a neutral act.
What shapes the size of a placebo response?
The size of a placebo response is not fixed; it scales with the dose form, the color, the price, the ritual, and even the patient’s genes, according to converging evidence from pharmacology and the field Kathryn Hall named the placebome. Two identical sugar pills can produce different effects depending entirely on how they are presented.
Reviews of placebo design have found that injections tend to outperform capsules, capsules outperform tablets, more pills beat fewer pills, and branded packaging beats plain. Color carries suggestion too: warm-colored placebos read as stimulants and cool-colored ones as sedatives. Hall, a molecular biologist at Brigham and Women’s Hospital, went a layer deeper and linked variation in the COMT gene, which sets baseline dopamine levels, to how strongly people respond to placebo [9]. Her review with Joseph Loscalzo and Ted Kaptchuk coined the placebome, the proposed network of genes that tunes each person’s response [9]. The number that matters here is not one statistic but a direction of travel: the placebo response is partly heritable, which means it is biology, not credulity.
| Factor | Direction of effect | Basis |
|---|---|---|
| Delivery form | Injection > capsule > tablet | Trial-design reviews |
| Quantity | More pills > fewer pills | Dose-expectation studies |
| Branding and price | Branded/expensive > plain/cheap | Marketing-cue studies |
| Color | Warm = stimulant, cool = sedative | Suggestion research |
| Genetics (COMT) | Dopamine variants alter response | Hall et al., 2015 [9] |
Why placebo responses are growing
Placebo responses in United States pain trials roughly doubled between 1990 and 2013, reaching an average 30 percent reduction in reported pain, according to an 84-trial analysis by Jeffrey Mogil’s group at McGill University published in 2015 [10]. The trend appeared only in American trials. European and Asian studies over the same window held flat.
Mogil’s team linked the rise to trials growing longer and larger: the average United States study stretched from about four weeks and fewer than 50 patients in 1990 to roughly twelve weeks and more than 700 patients by 2013 [10]. Bigger, longer, more elaborate trials seem to deepen the ritual and the expectation. A similar climb had already been documented in antidepressant trials, where the placebo arm has steadily narrowed the gap with active drugs. For drug developers this is expensive news, because a swelling placebo response can sink an effective compound that simply cannot beat belief by a wide enough margin.
Surgery is not immune. In 2002, orthopedic surgeon Bruce Moseley randomly assigned 180 patients with knee osteoarthritis to arthroscopic cleanup or a sham operation with only skin incisions, and found no difference in pain or function out to two years [11]. The scalpel added nothing the theater of surgery had not already delivered.
What the placebo effect means for medicine
The placebo effect reliably changes how patients feel, but it does not shrink tumors, clear infections, or repair the objective machinery of disease, a boundary that decades of controlled trials have drawn with unusual clarity. It moves symptoms, not pathology. A placebo can ease the ache of an arthritic knee and do nothing for the cartilage underneath.
That boundary is exactly what makes the effect useful rather than magical. It is a tool for the experiential side of illness, the part made of pain, nausea, fatigue, anxiety, and bowel distress, where suffering is real even when tissue is intact. Open-label trials hint that clinicians might one day prescribe placebos honestly for those symptoms, and the conditioning research suggests placebos could let patients taper real drugs while keeping some benefit. We don’t yet know how durable those gains are, or which patients they suit, and the genetics is still young. What is settled is the headline: expectation is a physiological event, traceable to dopamine, endorphins, and learned cues, and any honest account of medicine has to include it. For more frontier-science investigations in the same register, see other work from Dr. Felix Chen and the broader esovitae coverage of scientific anomalies and phenomena.
Frequently asked questions
Is the placebo effect real or just imagination?
It is real and measurable, though narrower than popular belief suggests. Brain imaging shows placebos release endorphins and dopamine, and controlled trials show genuine symptom relief. What placebos do not do is alter the underlying disease. They change the experience of illness, especially pain, nausea, and fatigue.
Can a placebo work if you know it is a placebo?
Yes. In a 2010 Harvard trial, irritable bowel syndrome patients given pills openly labeled as placebos still improved more than untreated controls, 59 percent versus 35 percent [6]. Later trials in back pain and IBS repeated the result. The leading explanation is conditioning and the ritual of being treated, not deception.
What is the difference between the placebo and nocebo effects?
The placebo effect is benefit from positive expectation; the nocebo effect is harm from negative expectation. Reading a drug’s warning label can produce its listed side effects in patients taking only a sugar pill. Both run on the same machinery of expectation, conditioning, and brain chemistry such as cholecystokinin and dopamine.
How strong is the placebo effect?
It varies by condition. In United States pain trials, placebo groups averaged roughly a 30 percent reduction in pain by 2013 [10]. In Parkinson’s disease, placebo can release striatal dopamine on a scale comparable to medication [5]. For objective outcomes like infection or tumor size, the effect is essentially zero.
Why does the placebo effect seem to be getting stronger?
A 2015 analysis of 84 trials found placebo responses rising only in United States pain studies, tied to trials becoming longer and larger over time [10]. More elaborate trials appear to heighten patient expectation. The same drift has appeared in antidepressant trials, complicating drug approval.
Does the placebo effect work on animals or infants?
Partly. Conditioning, one root of the effect, can be demonstrated in animals, and caregivers can report placebo-by-proxy improvements in infants and pets. But the expectation component depends on conscious belief about treatment, so the human placebo response is richer than anything seen in non-verbal subjects.
Can placebos replace real medicine?
No. Placebos relieve symptoms but do not treat disease, so they cannot replace antibiotics, insulin, or chemotherapy. Researchers instead study how to harness placebo effects alongside real treatment, for example using open-label placebos for chronic pain or conditioning to reduce needed drug doses.
Who studies the placebo effect today?
Major centers include Harvard’s Program in Placebo Studies, led by Ted Kaptchuk, and Fabrizio Benedetti’s neurophysiology group at the University of Turin. Kathryn Hall at Brigham and Women’s Hospital studies the genetics of placebo response, a field she named the placebome [9].
