Published WINTER 2006

Mind over matter?

Reproduced from Issue 131 of Arthritis Today

Edzard Ernst, one of the UK 's leading experts on complementary medicine offers his opinion on the phenomenon that is the placebo effect.


Edzard Ernst

J N Blau once wrote that “the doctor who fails to have a placebo effect on his patients should become a pathologist”. This little ‘bon mot' tells us a lot about placebo effects:

  • They are pretty much ubiquitous in clinical medicine
  • They benefit patients and are therefore desirable
  • We don't need a placebo to generate a placebo-effect

In the following short article, I will elaborate on these and several other themes related to the enigmatic placebo response.

Perhaps placebo-effect is not an ideal name for what I want to address. It has negative connotations. “This is only a placebo”, implies that the therapy in question does not really work. Moreover the term placebo-effect gives the impression that we are talking about one singular effect, possibly even one singular mechanism. The truth is that we are dealing with an array of phenomena and mechanisms. Many experts therefore prefer to speak of ‘context-effects'.

Context effects summarise everything that may affect health outcomes other than the specific effects of the therapy itself, in the simplest case the pharmacological actions of a drug. If a patient's pain improves after taking a pain-killer, this is usually interpreted as the result of the analgesic properties of the drug. But context effects could also play a significant, additional role.

For instance, the patient might have hoped to get better; thus his or her expectation could have contributed to the pain reduction. Or the doctor could have spent some time reassuring the patient, building up a sound therapeutic relationship with them; we know that the nature of the therapeutic relationship can affect the success of many treatments. Or the patient had prior experience and knew that, a little while after taking the tablet, the pain would subside. Just like Pavlov's dog's, we can be trained to have certain experiences even when the actual reason is lacking (e.g. even if the tablet contained only sugar), a process called conditioning.

Because context effects can be complex and are not based on one but several phenomena, it is understandable that there is more than one mechanism by which they operate. One theory holds that placebos work because they increase endorphins in the brain. These are chemical substances that decrease pain and make us generally feel good. Experiments have confirmed that patients who take placebos increase their endorphin levels in the brain. So here we have a mechanism that makes us understand how placebos work. But it would be naïve to think that this explains everything.

One observation has puzzled researchers and clinicians for many years. Why are placebo-effects so unreliable? Sometimes figures of around 30 per cent are cited to describe the magnitude of the placebo effect: around 30 per cent of people respond to placebo, or about 30 per cent of the total therapeutic effect is due to placebo. These are approximate averages; they do not mean that we can rely on the 30 per cent figure in individual cases. The somewhat confusing truth is that one patient may respond to placebo today but not tomorrow. Similarly there is no such thing as a placebo responder (someone who always benefits from placebo) and a placebo non-responder (someone who never benefits from it). This unreliability makes it problematic to count on placebo effects in clinical practice.

Yet when placebo-effects do happen they are generally (not always) beneficial; that is to say patients profit from them. Does that justify using treatments which are free of specific therapeutic effects (homeopathic remedies are often thought to fall in this category)? I don't think so, but if it is true that patients' symptoms can be improved by such therapies, why not use then for that purpose? The reason, I think, is that we don't need placebos to generate placebo effects. This may sound paradoxical but, actually, is quite simple. The placebo-effect is a bonus that comes ‘free' (so to speak) with any treatment regardless whether it also has specific effects or not. It is therefore neither logical nor appropriate to use pure placebos that only rely on placebo effects – one might as well use treatments that have both specific effects and placebo-effects. In this way one makes optimal use of the “free bonus”.

In a recent study sponsored by the Arthritis Research Campaign published in the British Medical Journal a year ago, we tried to find out whether magnetic bracelets are more effective than a placebo for osteoarthritis pain These devices are worn by many patients and we therefore felt that it was worth determining whether they actually work. Controlling for placebo-effects is difficult in this situation as patients can, of course, tell whether a bracelet is magnetic or not. We opted for testing placebo bracelets against weak and normal strength bracelets. Our results were not as clear as we had hoped. We found that magnetic bracelets reduced pain, but we are uncertain whether this response was due to specific or to placebo effects.

What seems important for maximising placebo effects is to administer treatments with empathy, sympathy, conviction etc – “the doctor who fails to have a placebo effect on his patients should become a pathologist”. A good therapeutic relationship may not be the only precondition for generating a sizable placebo response but it certainly helps.

Complementary practitioners often seem to be particularly skilled at building up strong therapeutic relationships with their patients. They certainly have more time but perhaps also more empathy and intuition. On the one hand, this is good news. It might enable them to be successful and have satisfied customers. On the other hand, there may also be a downside to it. Writing in The Guardian , Ben Goldacre explains: “Whether mainstream medics would want to go back to the old ways and embrace the placebo-maximising wiles of the alternative therapists is an easy question: no thanks. The didactic, paternalistic, authoritative, mystifying mantle has passed to the alternative therapist, and to wear it requires one thing most doctors are uncomfortable with, dishonesty”.

  • Edzard Ernst is Professor of Complementary Medicine at the Peninsula Medical School at the University of Exeter and Plymouth.