Placebos ease how you feel but rarely change what doctors measure

Key takeaways

• A Cochrane review of placebo-versus-no-treatment trials found no important effects in general, but modest relief of self-reported symptoms such as pain and nausea.
• In a 2010 Harvard-led trial, 59% of people with irritable bowel syndrome who knowingly took sugar pills reported adequate relief, compared with 35% who took nothing.
• In studies of healthy volunteers, honest placebos changed how people said they felt but not objective measures of the body.
• In a 2002 NIH-funded depression trial, placebo did as well as both St. John’s wort and the drug sertraline, a reminder of why fair tests need a placebo group.
• Before you credit a remedy that “works for me,” rule out the natural course of the illness, starting at your worst and other changes in your life.

In 2010, a Harvard-led team tried something that sounded like a contradiction. They gave 80 people with irritable bowel syndrome, or IBS, either nothing or a bottle of pills clearly labeled “placebo”¹.

The researchers told the pill group the truth. The capsules held no active ingredient, much like sugar pills. Patients didn’t need to believe in them. They only had to take them twice a day¹.

After three weeks, 59% of the pill group reported adequate relief of their symptoms. Only 35% of the group that got nothing did¹. The authors called their study small and said larger trials were needed¹.

It’s tempting to read that as proof that belief can heal anything. The full research picture is narrower, and more useful.

Placebos mostly change how symptoms feel, especially pain and nausea. They rarely change the things a doctor can measure, such as blood sugar or the disease process underneath. If you care about judging remedies fairly, natural or not, that difference is the heart of the matter.

The biggest review of placebos found small effects, mostly on symptoms

Many people picture “the placebo effect” as the improvement seen in a trial’s placebo group. That improvement is real, but much of it would have happened anyway.

So two researchers, Asbjørn Hróbjartsson and Peter Gøtzsche, asked a sharper question. What happens when people get a placebo compared with people who get no treatment at all? Their Cochrane review, last updated in 2010 with searches through early 2008, found 234 such trials. Of these, 202 had usable data, covering 60 conditions².

Their main conclusion was sobering. Placebos had no important clinical effects in general².

The exceptions were symptoms that patients report themselves. Placebos produced small benefits for pain and a small but steady benefit for nausea². For pain, the effect ranged from barely there to clinically important, even among well-run trials².

On rating scales, the effect on what patients reported was about twice the effect on what observers recorded². Lab measurements, such as blood sugar, showed no benefit; if anything, they leaned slightly the wrong way².

The authors added a caution. Some of the reported relief may be politeness. People who receive a treatment may tell researchers what seems expected². That makes a “true” placebo effect hard to separate from biased reporting.

Two details stand out. Physical placebos, such as sham acupuncture, tended to show larger effects than pills. So did trials that didn’t tell patients a placebo might be used².

Much of what looks like a placebo effect is something else

When a placebo group gets better, several forces are usually at work. Only one of them is expectation.

• Natural course. Colds clear up, back pain flares settle and many rashes fade on their own. Without a group that gets no treatment, the Cochrane authors note, a placebo’s effect can’t be told apart from this natural course².
• Regression to the mean. People tend to seek help when symptoms are at their worst. Symptoms that swing up and down will often drift back toward their usual level, whatever you do. The Cochrane authors list this among the forces that can mimic a placebo effect².
• Polite reporting. As noted above, people may report feeling better because it seems expected².
• Other changes. People in a study may also sleep more, worry less or try other remedies.

Psychiatry shows how big these combined forces can be. A 2024 umbrella review in Molecular Psychiatry, a review of reviews, gathered 1,691 trials of medicines and brain-stimulation treatments, with 261,730 patients³.

In depression trials, people on placebo improved by a large amount³. Across disorders, placebo groups tended to improve more in later trials, in larger studies run at more sites and among people who started out sicker³. That last pattern is what regression to the mean would predict, though the review did not test the reason.

The authors rated 18 of the 20 reviews they pooled as low quality, and placebo responses varied widely from one disorder to another³.

Honest placebos can ease symptoms, within limits

Since the 2010 IBS study, researchers have run more trials of “open-label” placebos, given with full honesty about what they are.

A 2021 review in Scientific Reports pooled 11 randomized trials comparing open-label placebos with no treatment⁵. The conditions included back pain, cancer-related fatigue, allergies, depression, IBS and menopausal hot flashes⁵.

The pooled effect was moderate to large. But results varied a lot from trial to trial, and the risk of bias was moderate⁵. The authors called the approach promising and the research early⁵.

A 2025 review of four studies of chronic low back pain, with 332 patients in all, gives a sense of scale. Open-label placebos lowered pain by about 0.6 points on a numerical pain rating scale, and the certainty of that evidence was very low⁷.

For comparison, a 2001 analysis of 10 pain trials, using the same kind of 0-to-10 scale, found that people on average rated themselves “much improved” after a drop of about 2 points, or about 30%⁸. By that yardstick, the average back pain benefit was small.

The most revealing study looked at healthy volunteers. A 2023 review of experiments in people without a disease pooled results two ways⁶:

• On what people reported about themselves, open-label placebos had a modest effect⁶.
• On objective measures of the body or behavior, the effect was essentially zero⁶.

The authors rated the certainty of that evidence as low to very low⁶. Still, the pattern matches the Cochrane findings: expectation changes the report more readily than the body²⁶.

One limit applies to all open-label trials. People know whether they got pills or nothing, and that knowledge can shape how they answer questions about their symptoms.

The ritual and the relationship matter

If a pill with nothing in it can help, what’s doing the work? A 2008 trial in the BMJ tried to pull the pieces apart⁴.

Ted Kaptchuk, who later led the 2010 study, and his colleagues enrolled 262 adults with IBS and split them three ways⁴:

1. A waiting list, with no treatment yet.
2. Sham acupuncture, a fake version of the treatment, with a practitioner who kept the first visit under five minutes and was told not to chat.
3. The same sham acupuncture with a warm, confident practitioner, a first visit of about 45 minutes, careful listening and questions about the patient’s life.

After three weeks, 28% of the waiting list reported adequate relief. So did 44% of the brief-visit group and 62% of the warm-visit group⁴. Quality of life followed the same pattern, though only the gap between the two sham groups was statistically clear. The authors judged the relationship the strongest single ingredient⁴.

This finding speaks directly to natural-health values. Long visits, careful listening and being treated as a whole person are a big part of what many natural-health practitioners offer. That care is not trivial. It appears to ease real symptoms.

It also means a remedy given with that kind of care will look stronger than it is unless it’s tested against the same care plus a placebo. Honoring the relationship and testing the remedy are two parts of the same honest standard.

Why fair tests compare remedies with a placebo

In 2002, a trial funded by the National Institutes of Health compared St. John’s wort, a popular herb for mood, with sertraline, a prescription antidepressant, and with placebo⁹. It enrolled 340 adults with major depression for eight weeks at 12 U.S. clinics⁹.

The share of patients who improved enough to count as a “full response” was 23.9% on the herb and 24.8% on sertraline. On placebo, it was 31.9%⁹.

Neither treatment beat placebo on the trial’s two main measures. Sertraline did better than placebo only on a secondary rating⁹. Because a proven drug also failed, the authors noted that the trial may not have been sensitive enough to detect a real effect⁹.

Picture that trial without a placebo group. Both the herb and the drug would have seemed to work, since about a quarter of patients fully improved on each.

The trial was not the last word. In its 2025 summary of the research, the National Center for Complementary and Integrative Health (NCCIH) says St. John’s wort appears to work better than placebo for mild to moderate depression. NCCIH adds that it’s about as effective as standard antidepressants for those milder cases, and less is known about severe depression¹⁰.

NCCIH also warns that the herb can interact dangerously with many medicines and can weaken some, including birth control pills and warfarin¹⁰. The lesson cuts both ways. A placebo group is what lets anyone, drug company or herbalist, see whether a treatment adds anything beyond hope, time and care.

Doctors use placebos too, often without saying so

Placebos aren’t only a research tool. In a 2008 survey in the BMJ, researchers mailed questions to 1,200 U.S. internists and rheumatologists. Of the 679 who answered, about half said they prescribed “placebo treatments” regularly¹¹.

Sugar pills were rare, used by 2%. Instead, 41% had used over-the-counter painkillers and 38% had used vitamins as placebos in the past year¹¹. Some had used antibiotics or sedatives the same way¹¹.

Most described these treatments to patients as something that might help but wasn’t usually used for their condition. Only 5% told patients outright that it was a placebo¹¹. Yet 62% considered the practice ethical¹¹.

The American Medical Association’s Code of Medical Ethics sets firmer rules. A doctor may use a placebo only with the patient’s cooperation and general consent, though not necessarily with notice of each dose¹². A placebo must never be used just to quiet a difficult patient¹².

The AMA council’s 2006 background report explains why. Passing off a placebo as real medicine risks the patient’s trust, and it can mask a problem and delay real treatment. Some patients also get side effects from placebos, the nocebo effect described next¹³.

The nocebo effect works in reverse

Expectation can make people feel worse, too. That’s the nocebo effect.

When Harvard researchers pooled 12 COVID-19 vaccine trials, more than 35% of people who got a placebo shot reported whole-body symptoms after the first dose¹⁴. Among people who got the real vaccine, the team estimated that nocebo responses accounted for about three-quarters of the whole-body symptoms reported after dose one¹⁴.

A companion article, “Side effects are real, and so are the ones fear creates,” covers the nocebo effect and medicine side effects in depth.

Before you credit a remedy: a six-question self-check

When something “works for me,” that experience matters. It’s also worth checking before you spend more money or tell a friend. Use this list the next time you start anything new, from an herb to a prescription.

1. Did I start at my worst? If you began on a peak day, some improvement was likely anyway.
2. Does this problem usually ease on its own? Colds, sprains, back pain flares and many rashes do.
3. What else changed? Note sleep, stress, season, diet and any other treatment you started.
4. What exactly improved? Write down whether it was a feeling, such as less pain, or a measurement, such as blood pressure. Placebos mostly move feelings²⁶.
5. Can I test it more fairly? Keep a daily symptom log for a few weeks before you start, during use and after a planned pause. Doctors sometimes run a formal version, called an N-of-1 trial, that rotates a treatment and a placebo in one patient¹³.
6. Is the ritual the active part? If calm, care and routine are what help, you may get the same benefit from a cheaper version of the ritual.

Stop and get checked instead of experimenting if symptoms are severe, keep getting worse or come with warning signs such as unexplained weight loss, bleeding, fever or new shortness of breath. Never stop a prescribed medicine to test a remedy without talking to your prescriber.

What the evidence shows

Claim	Evidence rating	What it rests on
People given a placebo report a little less pain and nausea than people given nothing	Strong evidence	Cochrane review of placebo-versus-no-treatment trials; the effect is small, varies, and may partly reflect biased reporting2
Placebos improve lab results and other objective measures	Evidence says no	No benefit on lab measures in the Cochrane review; no effect on objective measures in volunteer studies26
Open-label placebos ease symptoms such as IBS and back pain	Promising but early	Small trials, high variation, low certainty157
A warm, attentive practitioner adds symptom relief	Promising but early	One well-designed trial of 262 people4
All the improvement in a placebo group is caused by the placebo	Evidence says no	Natural course and other forces also play a part; placebo beat no treatment by only a little, while placebo groups often improve a lot23
Expectation can cause side effects (nocebo)	Strong evidence	Placebo groups in large vaccine trials14

Questions readers ask

The bottom line

Placebos are real, but their reach is narrow. They mostly ease how symptoms feel, especially pain and nausea, and the care and ritual around a treatment carry much of that power. They rarely change what a lab test would show.

That’s the reason fair trials compare remedies with a placebo, and the reason “it worked for me” deserves a second look. Value the relief, keep the rituals that help, and save your trust and money for treatments that also beat a sugar pill.

Related reading

• Side effects are real, and so are the ones fear creates
• Stress can reach your immune cells, but positive thinking still can’t cure cancer
• A fake inhaler made asthma patients feel better while their lungs showed no change
• Most health products aren’t scams, but many just haven’t been proven
• How to weigh healing stories, which move us most when the stakes are highest

Sources

1. Harvard Medical School. “Placebos Work, Even Without Deception” (on Kaptchuk TJ, Friedlander E, Kelley JM, et al., “Placebos without Deception: A Randomized Controlled Trial in Irritable Bowel Syndrome,” PLoS ONE, 2010). HMS News, December 2010. hms.harvard.edu
2. Hróbjartsson A, Gøtzsche PC. “Placebo Interventions for All Clinical Conditions.” Cochrane Database of Systematic Reviews, 2010, Issue 1 (reissued 2022, Issue 4), CD003974.pub3. cochrane.org
3. Huneke NTM, Amin J, Baldwin DS, et al. “Placebo Effects in Randomized Trials of Pharmacological and Neurostimulation Interventions for Mental Disorders: An Umbrella Review.” Molecular Psychiatry, 2024;29(12):3915-3925. pubmed.ncbi.nlm.nih.gov
4. Kaptchuk TJ, Kelley JM, Conboy LA, et al. “Components of Placebo Effect: Randomised Controlled Trial in Patients With Irritable Bowel Syndrome.” BMJ, 2008;336:999-1003. pubmed.ncbi.nlm.nih.gov
5. von Wernsdorff M, Loef M, Tuschen-Caffier B, Schmidt S. “Effects of Open-Label Placebos in Clinical Trials: A Systematic Review and Meta-Analysis.” Scientific Reports, 2021;11. pubmed.ncbi.nlm.nih.gov
6. Spille L, Fendel JC, Seuling PD, Göritz AS, et al. “Open-Label Placebos: A Systematic Review and Meta-Analysis of Experimental Studies With Non-Clinical Samples.” Scientific Reports, 2023;13:3640. nature.com
7. Flávio-Reis VHP, Pessoa-Gonçalves YM, Diaz CAV, et al. “Open Label Placebo for Chronic Low Back Pain: A Systematic Review and Meta-Analysis of Randomized Controlled Trials.” Taylor & Francis (figshare record), 2025. pubmed.ncbi.nlm.nih.gov
8. Farrar JT, Young JP Jr, LaMoreaux L, Werth JL, Poole RM. “Clinical Importance of Changes in Chronic Pain Intensity Measured on an 11-Point Numerical Pain Rating Scale.” Pain, 2001;94(2):149-158. pubmed.ncbi.nlm.nih.gov
9. Hypericum Depression Trial Study Group. “Effect of Hypericum perforatum (St John’s Wort) in Major Depressive Disorder: A Randomized Controlled Trial.” JAMA, 2002;287(14):1807-1814. dcscience.net
10. National Center for Complementary and Integrative Health. “St. John’s Wort.” NIH, updated May 2025. nccih.nih.gov
11. Tilburt JC, Emanuel EJ, Kaptchuk TJ, Curlin FA, Miller FG. “Prescribing ‘Placebo Treatments’: Results of National Survey of US Internists and Rheumatologists.” BMJ, 2008;337:a1938. scholars.duke.edu
12. American Medical Association. “Use of Placebo in Clinical Practice.” Code of Medical Ethics, Opinion 2.1.4. code-medical-ethics.ama-assn.org
13. American Medical Association Council on Ethical and Judicial Affairs. “Placebo Use in Clinical Practice.” CEJA Report 2-I-06, 2006 (background report for Opinion 2.1.4). Use of placebo in clinical practice — background reports.pdf
14. Harvard Medical School. “Power of Placebo” (on Haas JW, Kaptchuk TJ, et al., JAMA Network Open, 2022). HMS News, January 18, 2022. hms.harvard.edu