Does the placebo effect exist?

False perception of a placebo effect may be due to the fact that patients who have been given a placebo report improvement earlier and more eagerly in order to please and thank the care giver. These patients may even do this when there is no real physical improvement attained. One quoted figure is that about one third of patients improve on a placebo, but a recent study has called that number into question (A. Hróbjartsson & P. C. Götzsche, 2001), claiming that the effect is much smaller, if it exists at all. The 30 percent figure derives from a paper by Henry Beecher, published in 1955 (H. Beecher, 1955). Beecher was one of the leading advocates of the need to evaluate treatments by means of double-blind trials and this helps to explain why it has been so widely quoted.

It's been argued that the placebo effect may simply be an experimental artifact that arises because the patients in any study are selected to on the basis that they have certain symptoms. The observation of placebo effect is then attributed to natural healing of that symptom since the study began.

The Hróbjartsson & Götzsche study found that in many studies where a control group was used that did not get any treatment at all, the effects in the no-treatment group were almost equal to the effects in the placebo group for studies with binary outcomes (e. g. well treated or poorly treated). The authors concluded that the placebo effect does not have "powerful clinical effects," and conceded that placebos have "possible small benefits in studies with continuous subjective outcomes and for the treatment of pain." Their study suggested that there was no use of placebos outside of clinical trials.

In a follow-up study (A. Hróbjartsson & P. C. Götzsche, 2004) the same authors were able to confirm their previous results and concluded: "We found no evidence of a generally large effect of placebo interventions. A possible small effect on patient-reported continuous outcomes, especially pain, could not be clearly distinguished from bias " .