EDITOR, - Vernon M S Oh leaves us with a medical dilemma: on the one hand he tells us that we “should consider giving a placebo to appropriate patients,” but, earlier in the review, he tells us that “selecting the patients most likely to benefit from placebo is difficult.”1 In practice, doctors tend to give placebos to the least appropriate patients. These include neurotic patients, who are least likely to respond2; demanding patients who have an investment in not responding; and chronically ill patients who suffer from what psychologists have called placebo sag. Most placebos are probably given unwittingly by the doctor,3 and this folie a deux makes the placebo particularly effective. In this context, and with many active treatments, the placebo effect may be at its greatest but goes unrecognised.

Placebos' effectiveness is largely a function of the attitudes and skills of the doctor and his or her therapeutic authority and relationship with the patient. These factors are also important in consultations in which active drugs or no drugs are given. Indeed, a therapeutic consultation can be equally as effective as one in which either a placebo or symptomatic medication is given.4 Maximising the placebo effect should therefore be the aim of every consultation, and learning the necessary skills should, as Robert Buckman and George Lewith suggest,5 be part of the medical curriculum.

Neither drug companies nor medical reputations have much to gain from the placebo effect. This may partly explain why medical attitudes to it and, indeed, to doctors' role as healers have failed to …