Irving Kirsch

In 1969, William McGuire described the three stages of what he called the life history of an artefact [1]. The first stage is ignorance. With respect to the placebo effect, this characterizes the history of medicine up to the middle of the twentieth century. Before then, placebos might be used to mollify persistent patients, but the idea that they might produce actual benefits was rarely considered. In the 1950s and 1960s, a spate of studies reported beneficial effects of placebos, not only on subjective well-being, but also on concomitant physiological function. This signalled the transition from the ignorance stage to what McGuire called the coping stage in the life history of an artefact. The coping stage is the period in which the existence of an artefact is recognized and research methods are developed to control it. In the last half of the twentieth century, the use of placebo-controlled randomized clinical trials (RTCs) became the norm.

The papers in this theme issue of Philosophical Transactions indicate that the concept of placebo is maturing from the coping stage into the final stage in the history of an artefact: the exploitation stage. It is here that interest in a concept evolves from the view of it as an artefact that needs to be controlled to a phenomenon that is interesting in its own right. The focus shifts towards understanding the placebo effect. The questions being asked include: What are the psychological and physiological mechanisms by which placebo effects are produced? What are the conditions that placebo effects can or cannot be affected by placebo? What factors maximize or minimize the placebo effect? How reliable is it? Can placebo responders be identified and, if so, who are they? And finally, the sixty-four thousand dollar question: how can placebo effects be harnessed in clinical practice?

This latter question is particularly important in light of the following observations. There are conditions (e.g. depression) for which placebos can be as effective—or almost as effective—as approved medications [2,3]. Placebos differ from these medications, however, in at least one important way: they are virtually free of dangerous side effects. I say virtually free because placebos can cause side effects. This is one type of the negative effects of placebos that has given rise to the concept of the nocebo effect, which is an undesirable effect produced by placebos or placebo-like stimuli (e.g. verbal information and conditioning). The best example of placebo-induced side effects is a study reported in The Lancet in 1981, in which telling people that headaches are a side effect of lumbar puncture resulted in significantly more self-reported headaches following the procedure [4]. Still, the incidence of adverse events following placebo treatment is substantially less than their incidence following administration of active drugs.

The biggest barrier to the use of placebos in clinical practice is the almost universal perception that for a placebo to be effective it must be administered deceptively. Since expectancies appear to play a major role in placebo responding, informing people accurately that they are being given a placebo should prevent it from being effective. There is a way around this, however. If a convincing rationale can be presented, perhaps placebos can be prescribed openly without deception. Recently, my colleagues and I developed and tested such a rationale [5]. We told patients suffering from irritable bowel syndrome (IBS) that placebos have been shown to be effective for their condition, that their effects are induced at least in part by a well-known mechanism (that of classical conditioning), and that for that reason, the act of taking a placebo pill could work as a new mind–body treatment that could reduce IBS symptoms. We found that the patients in our study accepted this rationale, took their placebo pills as prescribed (two pills twice a day), and got better in comparison with patients in a control group that were not given the placebo pills. The effect was as large as that produced by commonly prescribed medication for IBS. Furthermore, it could not be completely accounted for by the therapeutic relationship, because time and attention were held constant across the two groups.

Of course, there are many conditions for which we have effective physical treatments, but the placebo effect can play a role even for these disorders. As shown most convincingly in the work by Benedetti et al. using the hidden infusion design [6,7], the placebo effect can be a component accounting for varying proportions of the response to active drugs. Other studies have shown that manipulating the ‘dose’ of placebo factors (e.g. the amount and quality of the time the clinician spends with the patient) can influence the treatment outcome significantly [8,9]. These studies are among the important ideas described and discussed extensively in various contributions to this Theme Issue of Philosophical Transactions. Taken together, these contributions suggest that we may be on the cusp of a new era, in which treatments aimed at harnessing and exploiting the placebo effect ethically and without deception can be developed, tested and implemented as an adjunct—and sometimes even as an alternative—to drug treatment.