Psychiatrist Patrick Lemoine believes that modern medicine still contains a large dose of magic - and it goes by the name of the placebo effect. According to Lemoine, the placebo effect can either swamp or enhance any proven pharmacological effect when a drug moves from the lab to real life. Human beings have exploited the placebo's formidable power since biblical times, yet Lemoine contends that western medicine is in denial about how heavily it still relies on the effect. The result? A lot of needless suffering. Laura Spinney asks him why the placebo effect remains medicine's dirty little secret.

How much do doctors use placebo therapy?

The most reliable estimates suggest that around 35 to 40 per cent of all official prescriptions are impure placebos, by which I mean pharmacologically inactive substances contaminated with a tiny amount of active ingredient - not enough to have a clinical effect, but enough for doctors to claim it does.

Why don't doctors admit they prescribe placebos?

The placebo effect - the difference between the predicted pharmacological effect of a treatment and what's observed clinically - rests in the relationship between the patient and the doctor. It is what remains of the craft of the witch doctor, because unless you really know what you are doing, it's unpredictable. Doctors hate not to be able to predict or control the outcome of a treatment, because it makes them feel like charlatans. As a result they tend not to prescribe pure placebos - that is, totally inactive substances - but impure placebos. This way they can fool themselves, at the same time as they fool their patients, that the treatment has predictable, scientifically tested effects.

Can you give an example of an impure placebo?

Magnesium is one. Rare conditions resulting from a deficiency of magnesium produce some symptoms very similar to those of anxiety. Magnesium is therefore often prescribed for anxiety in Europe, even though the cause, in most cases, is unlikely to be magnesium deficiency. The prescription of antibiotics for flu is another example. Antibiotics are ineffective against a virus, yet this practice was widespread until recently and is still endemic in France. The doctor tells the patient the drug will prevent secondary bacterial infection. It won't do the patient harm, but prescribing antibiotics for flu is irresponsible because it encourages the development of drug-resistant bacterial strains.

Are you guilty of prescribing placebos?

Impure placebos, yes. I confess that I prescribe magnesium for anxiety. My patients are generally satisfied and sometimes I even have the impression that not only do they show a remarkable improvement, but their relapse is almost immediate if the treatment is interrupted.

If I'm honest with myself, I don't really want to know if that improvement is a true pharmacological effect or a placebo effect. On the other hand, I would personally never prescribe homeopathy, because I consider that to be a pure placebo and my conscience wouldn't allow it.

Pure or impure, what harm can a placebo do?

When doctors prescribe antidepressants for mild depression, that can really be harmful. Numerous studies have shown that while severe depression responds well to antidepressants, mild depression responds no better to these drugs than to placebo. Because the drugs have unpleasant side effects their prescription for mild depression could be said to harm the patient. Still, doctors continue to prescribe them, because the act of prescription is such an integral part of the ritual of the medical consultation. They find it impossible to send a patient away empty-handed. The patient too, of course, expects to come away with a prescription.

How does the ritual of the consultation contribute to or reinforce the placebo effect?

A patient's susceptibility to the placebo effect depends on a number of factors, such as their level of anxiety and expectations, or the confidence the doctor expresses in the treatment. Since the placebo effect resides in the doctor-patient relationship, anything that boosts the patient's confidence in the doctor can enhance the effect. In our culture, that includes the wearing of a white coat, the ritual exchange of money and, last but not least, the incantation scribbled illegibly on a piece of paper, which is then handed over to another member of the priesthood at the pharmacy.

What kind of doctor-patient relationships minimise the placebo effect, or even produce a negative outcome?

It is not uncommon that a doctor's respect for his patient is eroded when that patient fails to respond to treatment. He may decide the patient is imagining or faking his symptoms. There is also a myth that is current among some doctors that only people of low intelligence respond to placebo treatment. If the doctor knowingly or unknowingly transmits his lack of respect to a patient, any treatment he then prescribes is less likely to be effective and may even produce unpleasant side effects. This is the nocebo effect and it can be quite noxious. Doctors should be taught how to avoid it but they aren't.

Can you cite an example of the nocebo effect?

A woman came into my clinic with major depression. For complicated reasons, she was wary of conventional medicine. When I prescribed her an antidepressant she developed a lot of side effects that were not explainable by the drug's pharmacology. This seemed to be a clear case of the nocebo effect. I eventually developed a technique whereby I started treating her with a pure placebo and gradually switched to the antidepressant. I told her I was going to do this, but I didn't tell her when she would receive the placebo and when the active drug, or how much of each. She made a full recovery.

Does the placebo effect influence the release of new drugs onto the market?

Let me tell you the story of Annie. Annie had spent more than 10 years in a psychiatric hospital in Lyon, suffering from severe depression. I discovered her curled up in an armchair in a corner of the ward. I built up a good relationship with her, and when I asked her to take part in a trial for a new antidepressant, she agreed. She responded extremely well to the drug. She left the ward, found an apartment and a boyfriend. Her recovery was so sensational that I'm sure it influenced the launching of the drug. Some time later, there was some refurbishment on the ward and, deep in Annie's armchair, we found the pills she was supposed to have taken. She had squirrelled them away - she hadn't taken a single one.

On balance, is it a good or a bad thing that the placebo occupies such an important position in the modern doctor's medicine cabinet?

Prescribing pure placebos can't be a good thing, because it involves a lie on the part of the doctor. But if you know what you're doing, giving an impure placebo can be justified. Even better, though, would be to persuade the patient that chemistry is not always the answer; that there are other ways to heal - through sport, for example, or love.