 | | | 
 | 
Sponsored by the Wellcome Trust | |  |  |  |    | 
|  |  | | (This debate is closed and is a read-only archive.) | Failing the double-blind test
[31-May-2005]
| | 'Any perceived benefit of homoeopathy or reflexology arises purely from the placebo effect or the regressive fallacy.' |  | | | There are plenty of worthy contenders for the greatest achievement in modern medicine. Vaccination, antibiotics, chemotherapy and even aspirin would all have their defenders. I prefer the candidate proposed by Robert Park of the American Physical Society, one of the finest debunkers of pseudoscience around. It is the randomised, double-blind, placebo-controlled trial.
| This tool is not glamorous. But it is the best method yet devised for establishing whether or not a medical treatment works. Before this dispassionate judge of efficacy, all drugs and procedures are equal. It cares nothing for reputation or popularity: the concoctions of Nobel laureates and folk healers play by the same rules. All that matters is whether a therapy does more good than a placebo, when neither patient nor doctor knows which is which.
| It is particularly effective when combined with two other great institutions - peer-review and independent replication. These can account for deliberate fraud, badly interpreted results and simple chance. Together, these procedures lay the foundation of scientific medicine. Without them we easily become prey to the confusions of the placebo effect and the regressive fallacy. The first of these is well-known - belief in a medical procedure can have therapeutic benefits, even when the procedure itself is inert. The latter, however, is more insidious: it describes the common mistake of substituting post hoc with propter hoc.
| The human mind is not very good at evaluating cause and effect. We tend to assume that sequential events are linked, even when they are not. Sometimes, medical complaints get better on their own. If we take an Ayurvedic remedy and happen to improve, we attribute our recovery to the treatment, even though it may have done nothing at all. The only way to be certain is to use the statistical power of a proper trial. Anecdote tells us nothing.
| The great problem with so-called alternative remedies is that they have never been tested in a rigorous way - or, worse, they have been tried and found wanting. Any benefit perceived by the devotees of homeopathy or reflexology arises purely from the placebo effect or the regressive fallacy.
| Furthermore, most alternative therapies not only lack experimental support, but are also devoid of any plausible biological mechanism. Homeopathy, for example, asks us to believe in the power of solutions so dilute they cannot contain even a single molecule of active ingredient. Many Eastern medical traditions rely on concepts of mystical life forces and meridians unknown to pharmacology or anatomy. There is
neither a demonstrable effect, nor a conceptual pathway, worthy of further investigation.
| A few treatments that are often grouped under the complementary and alternative umbrella do, it is true, have some robust results behind them. Acupuncture, for example, seems to have some use in pain relief. But their advocates often take this far beyond the evidence, claiming wider benefits against all kinds of conditions.
| In truth, there is no such thing as alternative medicine: therapies either work or they do not. It is hard to find a better summary than that given by Richard Dawkins, in his forward to the late John Diamond's brilliant Snake Oil. 'If a healing technique is shown to have curative properties in properly controlled trials, it ceases to be an alternative: it simply becomes medicine.'
| Mark Henderson is science correspondent for The Times (London).
|
|
| | |
|
|
|
