Homeopathy Trial Flawed

Archived

Friday, 20 December 2002

A TV investigation apparently debunking homeopathy has sparked controversy following claims that its key experiment was fundamentally flawed.

Amid a fanfare of publicity, the flagship BBC science programme Horizon last month broadcast "Homeopathy - the Test". It revealed the results of the programme's own investigation into the claims for this highly popular form of alternative medicine. But critics insist the study was always too small to give definitive results either way.

Homeopathy claims to treat a host of ailments with extremely dilute preparations of certain compounds. The method is regarded with deep scepticism by mainstream scientists, as many of the medicines are unlikely to retain even a single molecule of the original compound.

Advocates argue that the water may retain a "memory" of the compound's original presence. In particular a French researcher, Jacques Benveniste, claimed homeopathic solutions could activate basophils, a type of blood cell involved in allergic reactions.

The idea was widely thought to have been debunked in 1988, following an investigation by Nature. But since then, other teams have claimed small positive results. Some studies used thousands of samples, but counted cells by eye, which critics claim is unreliable. So Madeleine Ennis of Queen's University of Belfast Medical School, who was involved in one of the larger trials, carried out a smaller study using automated counting and saw the same small effect.

The Horizon team set out to replicate her findings and settle the issue, employing state-of-the-art equipment and independent researchers. They compared 40 homeopathic samples with water controls, and found no evidence of any effect on the cells.

While sceptics seized on the result as proof that homeopathy is nonsense, the study itself has come under attack for being too small to have any hope of confirming the size of effect seen in previous studies. Even Ennis, who is an avowed sceptic, says the Horizon experiment was incapable of making any kind of definitive conclusion. "In our [original] study we had thousands of samples," she points out.

The programme's statistical adviser, Martin Bland of St George's Hospital Medical School in London, declined to comment on the issue of sample size. He did tell New Scientist, however, that "it is obviously true that because one system fails to detect 'memory' in one experiment we cannot infer that there is no 'memory'."

Nathan Williams, the programme's producer, stresses that each of the 40 samples was tested on several different sets of cells, boosting the statistical power of the study. "We weren't attempting to replicate the much larger studies, as there are problems with the counting methods used," he says.

Source: Robert Matthews New Scientist vol 176 issue 2372 - 07 December 2002, page 10