Top 10 Cases Where the Placebo Effect Misled Science

Sham operations, colored pills and expectancy tests have repeatedly shown that a patient's belief can look like a drug, and can lead researchers astray. The placebo effect is not a single trick, and it is not proof that an illness is imaginary. It is a set of measurable changes — in pain, in some symptoms, sometimes in physiology — that follow a treatment because of expectation, ritual, conditioning, or the ordinary course of a disease ^[1][6]. This list is a curated set of ten cases and lines of research in which those factors complicated what scientists thought they were seeing.

Some of the entries are controlled trials that changed practice. Others are famous anecdotes that should be handled with care, or laboratory designs that isolate expectancy from pharmacology ^[3][4]. The point is not to mock patients. It is to show why modern trials use blinds, shams, and control arms.

Key Facts

A placebo is an inactive or sham intervention used for comparison. The "placebo effect" is the response that can still appear. Good trials try to separate that response from a drug's specific action.

Ten Placebo Lessons at a Glance

1. Sham Knee Arthroscopy

In 2002, J. Bruce Moseley and colleagues reported a randomized, placebo-controlled trial of arthroscopic surgery for osteoarthritis of the knee in the New England Journal of Medicine ^[1]. Patients were assigned to débridement, to lavage, or to a sham procedure in which incisions were made but the arthroscope was not used to treat the joint. Over a two-year follow-up, the real operations did not outperform the sham on the trial's pain and function measures ^[1].

The paper did not say that no one ever needs knee surgery. It said that, in this group and for this indication, the ritual and the expectation that go with an operation could account for much of the benefit that had been attributed to the arthroscopy itself. The result forced a more careful look at a common procedure and remains a standard example of why surgical trials, not only drug trials, need sham or other strong controls.

2. Suggestion and a Skin Rash

Dermatology and psychosomatic research in the mid-twentieth century included experiments in which a harmless leaf or a suggested "exposure" was presented as a plant that would cause a rash, and in which a real irritant was sometimes presented as harmless ^[2]. In some of these designs, a portion of volunteers developed or reported a reaction that followed the suggestion more closely than the chemical facts. The literature is older than today's reporting standards, and not every famous classroom story is a well-controlled paper, so the claim should be kept modest ^[2].

What can be said, without dressing the work up as a single definitive "blindfold test," is that expectation can influence skin symptoms and the reporting of those symptoms. The finding sits with a wider literature on nocebo effects: warning people about a side effect can make the side effect more likely to be felt. A rash is a vivid example because it looks so plainly "physical."

Related: Top 10 Psychological Facts About Human Behavior

3. The Case of Mr. Wright

In 1957 the psychologist Bruno Klopfer published an account of a patient, known as Mr. Wright, who had advanced cancer and appeared to improve dramatically after receiving Krebiozen, a much-hyped experimental treatment, then to relapse when he learned of sceptical press coverage, and to rally again after a sham "improved" injection ^[3]. The narrative is one of the most often retold placebo stories in medicine. It is also a single case, reported by a clinician after the fact, without the controls that would make it a scientific demonstration ^[3].

The honest use of the story is as a warning about how powerfully hope, authority, and news can seem to move a serious illness in one person, and about how carefully such reports must be read. It is not a reason to prefer anecdote over trials, and it is not a reason to sell false hope. Later writers have used it, rightly, as a prompt to study expectation, not as a substitute for evidence.

4. When People Think They Have Drunk

Alcohol researchers have long used a "balanced placebo" design: some participants receive alcohol and are told that they have, some receive alcohol and are told that they have not, some receive a non-alcoholic drink and are told that it is alcohol, and some receive a non-alcoholic drink and are told the truth ^[4]. In a number of studies, especially of social and aggressive behaviour, what people were told explained as much or more than what they actually drank ^[4].

The design misled informal common sense, which tends to attribute every change at a party to chemistry. Pharmacology is real; so is expectancy. The two can be separated only if the experiment is willing to deceive, carefully and with consent, about the contents of a glass. That is why the finding belongs on a list of ways belief can impersonate a drug.

5. The Color of a Pill

A scattered but persistent literature has found that the color, size, and form of a tablet can change the effect people expect, and sometimes the effect they report ^[5]. Warm colors have often been associated, in such studies, with stimulation; blue and other cool colors with sedation; and injections or capsules with a stronger anticipated effect than a small plain tablet. The results are not identical in every country or every sample, because color is cultural as well as perceptual ^[5].

Related: Top 10 Most Bizarre Phobias: Strange Fears That Exist

For trial design, the lesson is practical. If a "placebo" does not look and feel like the drug, part of the difference between arms may be packaging rather than molecule. If two active drugs are compared in different colors, marketing and medicine can become hard to pull apart. Appearance is not a detail. It is part of the treatment.

6. Open-Label Placebos

Ted Kaptchuk and colleagues have reported trials in which patients were told, honestly, that they were receiving a placebo — a pill with no active drug — and still showed improvement on some symptom measures, including in irritable bowel syndrome and other conditions ^[6]. The finding is surprising only if one assumes that a placebo works solely by hidden deception. If ritual, a caring clinical encounter, and a plausible story about the brain's predictive machinery also matter, then honesty and a placebo are not a contradiction ^[6].

Open-label work does not mean that placebos should replace proven drugs for diseases that need them. It does mean that the older idea of the placebo as a mere lie is too narrow, and that some of what looks like "being fooled" may be a treatable feature of how symptoms are generated. The science is still being tested. It has already complicated a simple picture.

7. Ligating the Internal Mammary Artery

In the 1950s some surgeons treated angina by tying off the internal mammary artery, on the theory that blood would be redirected to the heart ^[7]. Early uncontrolled series looked encouraging. Then, around 1959–1960, randomized comparisons that included a sham incision, associated with work by Leonard Cobb, E. Grey Dimond, and others, found that the real ligation did not clearly outperform the fake operation ^[7]. The procedure fell out of use.

This is one of the cleanest historical examples of a surgical fashion being corrected by a sham control. Patients in both arms could feel better, which had been enough to sustain the operation until someone measured it against a performance of surgery. The episode is a precursor of later sham-surgery trials, including the knee study above, and a warning that a plausible mechanism plus grateful patients is not a result.

Related: Top 10 Oldest Buildings in the World

8. Placebo Sleep

In 2014, Christina Draganich and Kristi Erdal reported that adults who were given false feedback about how much they had slept the previous night — that they had slept better or worse than they had in fact slept, based on a sham reading — then performed differently on attention tests in the direction of that story ^[8]. The study is small by the standards of a drug trial, and it is about a night's belief, not about a treatment for insomnia, but it is a neat demonstration of expectancy in a domain people treat as purely physiological ^[8].

If a person can be "better rested" or "worse rested" in performance partly because of a number they were told, then fatigue, like pain, is not only a meter of hours. Sleep science still needs objective measures. It also needs to remember that the story of a night can leak into the day.

9. Sham Surgery and Parkinson's Disease

Surgical and cell-based treatments for Parkinson's disease have been tested, in some trials, against sham operations in which a patient is prepared and even has a burr hole started, but does not receive the experimental implant or lesion ^[9]. In more than one early study, substantial improvements appeared in the sham arm, and some hoped-for advantages of the real procedure shrank or vanished in the stricter comparison ^[9]. The ethics of sham brain surgery are seriously debated, because a control arm still carries risk.

The scientific point is harder to dodge than the ethical one is to resolve. Parkinson's symptoms, including some motor features, can respond to expectation and to the intense ritual of a neurosurgical program. Without a sham, a dramatic before-and-after video is not enough. The field has had to learn that lesson in public.

Related: 10 Sikh Gurus: The Pillars of Sikhism Who Shaped Faith, Equality, Courage, and Humanity

10. Conditioned Placebo Analgesia

Work associated with Fabrizio Benedetti, and earlier conditioning research in the tradition of Robert Ader, has shown that if a drug such as morphine, or another active analgesic, is repeatedly paired with a context or a cue, the cue alone can later produce part of the pain relief ^[10]. The brain has learned an association. The "placebo" is then not only a story the patient has been told; it is a learned physiological response ^[10].

This line of research helps explain why placebos are not all equally empty, and why a patient's history with a clinic, a syringe, or a brand can matter. It also warns researchers not to treat a washout as if it had erased every trace of a prior drug. Conditioning can mislead a trial, or it can be used, carefully, as part of a treatment plan. Either way, it is one of the mechanisms that make belief look like pharmacology.

Conclusion

The placebo effect has misled science when uncontrolled operations, colorful pills, hopeful announcements, and learned cues were mistaken for specific treatments. It has also instructed science, once trials were designed to catch it. Sham surgery, balanced alcohol experiments, open-label pills, and conditioning studies are different windows on the same problem: a human being is not a passive container for a molecule.

None of this is an argument against medicine. It is an argument for better controls, for modest interpretation of single cases, and for taking symptoms seriously even when expectation is part of their cause. The ten examples above are a starter set, not a complete map of a very large literature.