Even if given a dummy medicine, some people get better, thanks to the miraculous placebo effect. Could it ever be used deliberately as a form of medicine?

Does a medicine depend only on the active ingredients within it? Oddly, the answer is no. No matter what is in them, painkilling medicines are more effective when wrapped up in 'proper' pharmaceutical packaging. Red, yellow or orange pills provoke a strong stimulatory or antidepressant effect; blue, purple or green tablets are good for tranquillising reactions.

In 1955, Henry Beecher tried to put a size on the placebo effect. He found that an average of 32 per cent of patients across 26 studies responded to placebo. Remarkably, then, nearly one in three patients improves even though they are not receiving a medicine.

What might be going on? In reality there are likely to be several 'placebo effects'. For example, disease may spontaneously regress in some cases.

But there is growing evidence of physiological changes linked to the placebo effect. Brain imaging studies, for example, show differences in brain areas associated with pain when people are anticipating pain relief (even if it never actually arrives).

But it is not just the brain that is influenced by the placebo effect. Changes in the immune system and cardiovascular function have also been linked to placebo responses.

It therefore appears that expectation can influence physiology. Mind over matter is a real phenomenon. The placebo effect may lie at the heart of complementary medicine - a 30 per cent success rate is reasonable (though not something to rely on for serious conditions or when good medicines are available).

There have even been suggestions that doctors should deliberately use the placebo effect for tricky patients - giving them a dummy medicine that they know won't work but the patient believes will. Many doctors will admit privately to doing this. But it does raise issues about trust - as doctors have to deliberately deceive patients in order to help them get better.