Then, in 2000, the European Commission ordered a review from four eminent French clinical pharmacologists, asking them to report on homeopathy to its Science, Research and Development directorate. These investigators employed even more rigorous quality selection criteria than their predecessors, choosing to analyze only 17 out of 118 clinical trials. Again, they concluded that the balance of evidence was in favour of homeopathy, citing a probability figure of 0.000036 (i.e., the likelihood of this being a fluke result is very low indeed), and thus "extremely significant", according to medical statisticians (Eur J Clin Pharmacol, 2000, 56: 27-33).

Apart from the heavily criticized 2005 Lancet review, the most recent "critical overview" was carried out three years ago by a small international team of experts, including one from the Harvard Medical School (Ann Intern Med, 2003; 138: 393-9). They presented a meta-review of all the reviews to date, plus-for the first time-a meta-analysis of homeopathic trials on specific conditions.

Overall, they concluded that while "the quality of clinical research in homeopathy is low. . .when only high quality studies have been selected for analysis. . . a surprising number show positive results". Turning to specific conditions, they again criticized the quality of the basic research as "scant, of uneven quality", but still were able to tease out some pretty clear indications of what conditions homeopathy works best for (see box above).

Power of the Placebo
Apart from the quality of the research, one problem that has bedevilled homeopathic clinical trials is the size of the placebo effect. Homeopaths acknowledge that patients who go through the elaborate consultation process with a homeopathic doctor are likely to have a huge clinical response, even if given dummy placebo pills.

The importance of the doctor-patient relationship in the curative process has been known for millennia, but in the drug-dominated 20th century it has tended to be dismissed and derided as a "mere" placebo effect. Recently, however, the sheer power of the placebo effect has become recognized, as scientists have been able to map the pathways between the brain and the immune system. What the new research shows is that someone's belief in a medicine may boost the body's self-repair system sufficiently to produce a cure-all by itself. This means that any medicine has a mountain to climb in order to show that it is better than placebo-and, of course, the stronger the placebo effect, the higher the mountain.

"I prefer to call it a non-specific effect rather than placebo, but it undoubtedly does form a large part of homeopathy," says Dr Peter Fisher. "It's to do with the art of medicine-the doctor-patient relationship. But homeopathy clearly has specific effects over and above the non-specific ones. Besides, from the patients' point of view, what do they care how the curative effect is achieved? All they're really interested in is getting better".

Assessing just how much patients actually benefit is the latest aspect of homeopathic research. In addition to testing homeopathic medicines as if they were drugs, researchers are now looking at the so-called "outcome". In short, do people really get better with homeopathy (as opposed to a placebo effect), and how does that compare with other types of medicine?

The largest outcome survey has been done by doctors at the Bristol Homeopathic Hospital. They analyzed over 23,000 outpatient consultations from 1997 to 2003, and found that over 70 per cent of their patients reported "clinical improvement". Particularly striking is the fact that many of their patients had chosen homeopathy only after mainstream medical treatment had failed them.