How real is the placebo effect?

We investigate

16th December 2017

We’ve all heard of the placebo effect, the idea that unwitting patients taking fake medicine – replacing it with sugar pills or some other harmless substitute – can miraculously get better, giving the impression that taking placebos can be as effective as real medicine.

A joint survey by The University of Oxford and The University of Southampton 2013 found that 97% of GPs had given patients an “impure placebo” – medicine with an active ingredient that isn’t usually prescribed for the condition being treated – and 10% had prescribed “pure placebos” – medicine with no active ingredient, such as sugar pills or water.

Recently, the power of the placebo has been given a boost, with numerous news outlets, ranging from The Times and The Daily Mail to Gizmodo and Vox, all running the story that placebos work even when patients know they’re receiving a placebo.

On top of which, the placebo effect has also been used to defend scientifically unsound alternative treatments such as homeopathy; after all, if it makes the patient feel better, what’s the harm?

The problem is that the placebo effect may not exist, at least not as we understand it.


Cochrane is an organisation that provides healthcare professionals with research and evidence to help them make the informed decision for their patients. In an article from 2010, Cochrane found that “placebo treatments produced no major health benefits, although on average they had a modest effect on outcomes reported by patients, such as pain. However, the effect on pain varied from large to non-existent”.

Further to this, Ted Kaptchuk, Professor of Medicine at the Harvard-affiliated Beth Israel Deaconess Medical Centre published an article in the New England Journal of Medicine investigating the effects of placebo treatments on asthma sufferers. In it, he states that the “placebo had no effect on measurable biological factors but was indistinguishable from medication with regards to subjective outcomes”.

Subjective outcomes in this sense refer to how the patients feel about their treatments regardless of whether or not they are actually better. It’s a psychological effect rather than a physiological one. In short, placebos don’t make patients better; it makes them believe that they feel better. This might seem like splitting hairs, but feeling better and getting better are two separate things. A patient who feels better but isn’t getting better puts themselves at risk.

The scratch of a routine flu jab can become a merciless impaling in the eyes of someone afraid of needles

The only area where placebo seems to have an effect over no treatment at all is in psychological reactions to pain. There is evidence to suggest that the amount of pain we feel can be influenced by factors such as anxiety. The scratch of a routine flu jab can become a merciless impaling in the eyes of someone afraid of needles. But as mentioned by both Cochrane and Kaptchuk, the perceived effects of placebos on pain are no different than the effects of actual medicine – though, in the case of taking medicine, you also have the observed physiological effect of the medicine as well.

If placebos have no physiological effect and regular treatment is just as effective in regards to psychology, why does the myth of the power of placebos continue? 


In some instances it’s doctors themselves. Like anyone else they are subject to confirmation bias, accidentally seeking out evidence for what they already believe or perhaps they simply aren’t up to date with the latest research.

In some cases, this bias may be deliberate. A number of authors have questioned Kaptchuk’s work, claiming he has a habit of overselling his results and the efficacy of placebos. In the same paper that he wrote “placebo treatments produced no major health benefits,” he concluded that the “Placebo effects can be clinically meaningful and can rival the effects of active medication in patients with asthma.”

From the paper’s context, this may be referring to subjective accounts but this is not explicitly stated. The Overtake reached out to Professor Kaptchuk but at the time of publication had received no comment on his paper’s conclusion.

News sites across the globe have used Kaptchuk’s research and regardless of his intentions, for the large part, focusing on the apparent benefits of placebos. The media can often misrepresent scientific findings with nuance being lost in the translation of scientific jargon into something easily digested by the average reader. In some cases it may be a matter of authors trying to make the story more relevant or compelling for their specific audiences or facts being mistakenly pulled from other reports on the same story rather than the source material.

This kind of scientific illiteracy means there is still a strong belief in the power of the mind in medical matters and the idea that if you feel like you are getting better, you will get better. But even that doesn’t hold water.


There are a broad range of effects that can alter your belief about your treatment and the effectiveness of placebos. The simplest explanation you were going getting better anyway.  As your body gets better, it’s easy for doctors to incorrectly attribute that healing to placebos. In “open label-placebos” trails, cases where patients know they are taking a placebo, patients are often explicitly told about the healing potential of placebos, some may already believe that to be the case and the same thing happens. The body’s natural recovery is assumed to have been the effect of the placebo.

Confirmation bias means patients are more likely to notice and report changes that they are expecting and ignore changes they aren’t

Further to that, clinical tests are subject to a number of confounders – biases that can affect an experiment’s results – creating the appearance of the placebo effect. Mike Hall of the Merseyside Skeptics has been investigating the placebo effect, or as he prefers, the placebo effects.

“There are various biases to account for. Confirmation bias means patients are more likely to notice and report changes that they are expecting and ignore changes they aren’t,” he says.

“The Hawthorne Effect means patients may alter their behaviour, simply because they are aware they are being observed,” he adds.

Apparent improvements in the control group may, in part, be due to this.

“Then there is the Observer-Expectancy Effect, Recall Bias, the Clever Hans Effect, improved compliance, selection bias, and more and more. All these effects and biases can alter the recorded data, even if they don’t change the condition of the patient. This makes the patients in the placebo arm look as if they’ve improved more than they have.”

All these factors together indicate that placebos are unlikely to have value in medical treatment outside of clinical tests, where they act as a baseline against which actual treatment is measured. In tests, the ethics of lying to patients about their medication is somewhat justified as a part of avoiding biasing the experiment. But extending this to general medical practice seems unreasonable.

Lying to patients erodes trust in doctors and prescribing placebos is a matter of ethics. Add to that the additional cost of placebo treatments and the potential long-term implications of antibiotic resistance in impure placebo prescriptions and the benefits of placebos become extremely difficult to justify. When placebos offer no advantage over actual medicine, prescribing them to a patient to placate them or for peace of mind serves only to mislead them, and there is no sugar coating that.

16th December 2017