Survey finds 97% of GPs prescribe placebos


"Most family doctors have given a placebo to at least one of their patients," BBC News reports. The news is based on a large survey of UK GPs...

"Most family doctors have given a placebo to at least one of their patients," BBC News reports.

The news is based on a large survey of UK GPs. For the purpose of the study, placebos were put into one of two categories:

  • pure placebos – treatments containing no active ingredients, such as sugar pills
  • impure placebos – treatments that contain active ingredients but are not recommended for the condition being treated, such as antibiotics for flu

The survey found that 97% of doctors admitted to giving an impure placebo at some point during their career, while 10% had given pure placebos.

The survey found that more than 1% of GPs used pure placebos at least once a week, and more than three-quarters (77%) used impure placebos at least once a week. Most doctors said placebos were ethical in some circumstances.

Placebos are often used in the control group in trials looking at the effectiveness of treatments. It is widely recognised that they can result in an improvement in a patient’s condition – a phenomenon known as the placebo effect.

However, there is an ongoing and vigorous debate about whether using placebos in normal medical practice is ethical.

Where did the story come from?

The study was carried out by researchers from the University of Oxford and the University of Southampton. It was partly funded by the University of Oxford and the Southampton Complementary Medical Research Trust (a registered charity).

The study was published in the peer-reviewed journal PLOS ONE, which is freely available to read on an open access basis.

The study was covered fairly in the media.

What kind of research was this?

This was a cross-sectional survey of a random sample of GPs in the UK. The survey used a web-based questionnaire asking about the GPs’ use of placebo treatments. The researchers say that although surveys in various countries suggest that 17-80% of doctors have routinely prescribed placebos, their use by UK GPs is unknown. They also aimed to find out under which conditions GPs think placebo use is ethical.

What did the research involve?

In April last year, researchers emailed their survey to a random sample of 1,715 GPs registered with (a commercial website for doctors). Email reminders were sent twice and the survey closed about a month later. The questionnaire asked GPs to note how frequently (if at all) they used placebo treatments. It also asked their reasons for placebo use, the circumstances under which they felt a placebo was ethically acceptable and what they told patients when they prescribed a placebo.

Understanding of what is meant by placebo is important in this study. The researchers classified the placebos as “pure placebos” or “impure placebos”.

Pure placebos were defined as interventions that had no active ingredients, such as sugar pills or fresh water injections.

Impure placebos were defined as substances, interventions or ‘therapeutic’ methods that have known value for some ailments but lacked specific effects or value for the condition for which they were prescribed. The examples given included:

  • positive suggestions (this is not explained in the study)
  • nutritional supplements 
  • probiotics for diarrhoea
  • peppermint pills for pharyngitis
  • antibiotics for suspected viral infections 
  • sub-clinical doses of otherwise effective therapies 
  • ‘off-label’ uses of potentially effective therapies
  • complementary and alternative medicine, such as homeopathy, the effectiveness of which is not evidence-based
  • conventional medicine where the effectiveness is not evidence-based
  • non-essential diagnostic practices, such as X-rays or blood tests, based on the patient’s request or for reassurance

For each type of placebo, prevalence of use was categorised as frequent (daily or about once a week), occasional (about once a month) and rare or never (more than once a year or never).

What were the basic results?

Of the 1,715 GPs contacted, 783 (46%) completed the questionnaire. The researchers found that:

  • 12% (95% confidence interval (CI): 10 to 15%) had used pure placebos at least once in their career
  • 97% (95% CI: 96 to 98%) had used impure placebos at least once in their career
  • 1% used pure placebos at least once a week
  • 77% (95% CI: 74 to 79%) used impure placebos at least once a week
  • most doctors (66% for pure, 84% for impure) believed placebos to be ethical in some circumstances

At least a quarter of GPs used certain impure placebos frequently. These included non-essential physical examinations, conventional medicine where the effectiveness was not evidence-based and (somewhat worryingly given the growing problem of antibiotic resistance) antibiotics for viral infections.

The reasons GPs gave for prescribing both pure and impure placebos varied. They included possible psychological treatment effects, requests by the patient for a therapy and the treatment of non-specific complaints.

Half the GPs who used a placebo treatment told patients it had helped other patients, without specifically telling them it was a placebo. However, a large majority of doctors (about 80%) thought that pure or impure placebos were not acceptable when they involved deception. More than 90% thought they were not acceptable if they endangered the trust between doctor and patient.

How did the researchers interpret the results?

The researchers say that placebo use is common in primary care but that questions remain about the benefits, harms and costs of placebos, and whether they can be delivered ethically. Further research is required to investigate ethically acceptable and cost-effective placebo interventions, they argue.

Researchers also say the survey was a representative sample of GPs – and that the response rate was high enough to reflect the GP population.


This survey suggests that about three-quarters of GPs use an “impure” placebo at least once a week and that most think placebos have a useful role to play in treatment. Importantly, almost all think that any risk of damaging a trusting relationship is unacceptable. However, it is possible to prescribe a placebo to a patient without actively lying to them.

Limitations of the study include:

  • selection bias – GPs with strong views (either for or against placebos) may have been more likely to respond to the survey
  • recall bias – whether GPs recalled their use of placebos, or responded to the survey, accurately is uncertain

However, this survey remains of interest, particularly the finding that GPs regularly use “impure” placebos. Aside from the ethical issue, such placebos can be costly and they may also be harmful. Sometimes they contain harmless substances that cause ill effects – these are referred to as "nocebos". For example, antibiotics can have side effects and used inappropriately they also promote antibiotic resistance, resulting in increasing ineffectiveness, which has been highlighted recently in a report on antibiotic resistance by the Chief Medical Officer.

A clear and agreed definition of placebo types is clearly required. As the authors argue, further research is needed into the benefits and harms of using placebos, and their cost. Consultation on whether they are ethically acceptable could also be considered.

Article Metadata Date Published: Tue, 15 Aug 2017
Author: Zana Technologies GmbH
NHS Choices