-This paper reviews the psychological mechanisms that lead to placebo responses and the physiological basis for reduction in symptoms. Some psychologica l mechanisms (expectancy, conditioning) lead to symptom reduction but are unlikely to reduce underlying pathology. Other mechanisms (therapeutic relationship, empowerment) may additionally reduce later pathology. The nature, size and duration of the placebo response depends on the placebo inducing context. In clinical practice, the placebo response creates an adjunctive response to that of active treatment. It is a useful, but fickle, boon as it is difficult to predict when it will occur.
KEY WORDS: empowerment, expectancy, placebo, psychology, therapeutic relationshipThe term placebo response has been defined recently as 'the reduction in a symptom as a result of factors related to a subject's/patient's perception of the therapeutic intervention' . 1 Research on placebos often involves psychological symptoms such as depression or pain, so there is a tendency to treat placebo responses as 'not real' or only psychological. However, placebo responses can create a variety of physiological changes. Not only is placebo pain reduction mediated in part by endorphin release, 2 but placebos create endogenous dopamine release in Parkinson's disease 3 and changes in bronchial muscle tone (peak expiratory flow rate) in asthmatics. 4 Placebo responses are just as real as any other therapeutic response -what distinguishes them is their cause.When evaluating pharmacological and other treatments, placebo responses are an inconvenience that must be controlled by using randomised controlled trials (RCTs). However, in clinical practice, placebo responses have an impact on patient outcome. It is increasingly recognised that, when properly handled, the placebo effect is a 'boon to busy clinicians and their patients' . 5 The aim of this paper is to examine how this boon can be exploited in clinical practice.
How large is the placebo response?One way of evaluating the overall impact of placebo responses is to compare the outcome in the placebo arm of an RCT with that of the treatment or verum arm. Table 1 reproduces the results of a meta-analysis of improvement rates in the placebo and verum (active) arms of several RCTs. 6 The authors conclude that: placebo responses vary considerably between different diseases, and clinically significant numbers of patients improve without active treatment. It was also concluded from RCT data that, contrary to common belief, placebo responses can be long term, at least 12 weeks in duration.One difficulty in evaluating the long-term impact of the placebo response is that some chronic diseases remit spontaneously. The true effect of the impact of placebos requires comparison with the natural history of the disease rather than with a baseline of no change. With this comparison, the extent of placebo response is reduced. 7 A review 1 comparing placebo with natural history in RCTs of analgesia which included a natural history arm suggests a mean...