“For 2,400 years,” wrote the historian of medicine David Wootton, “patients believed doctors were doing them good; for 2,300 years they were wrong.” Only in the past 100 years have treatments in the mainstream of medicine been consistently subject to clinical trial, to discover what works and what doesn't. Much medicine, though, still stands defiantly outside this mainstream. Can these alternative therapies really claim to be medically effective judged by today's standards, or are they no better than the blood-letting and snake oil of darker centuries?

Simon Singh, a science writer, and Edzard Ernst, a doctor, have set out to reveal the truth about “the potions, lotions, pills, needles, pummelling and energising that lie beyond the realms of conventional medicine”. Their conclusions are damning. “Most forms of alternative medicine,” say the authors, “for most conditions remain either unproven or are demonstrably ineffective, and several alternative therapies put patients at risk of harm.” The book is dedicated, in ironic homage, to HRH the Prince of Wales, that famous champion of complementary medicine, especially homeopathy (said to be used both for the prince's own body politic and for the cows at Highgrove).

Fearless, intelligent and remorselessly rational, the authors exemplify the same Enlightenment spirit of criticism that animated The Lancet in its early days. One by one, they go through the most influential alternative therapies (acupuncture, homeopathy, chiropractic and herbal medicines) and subject them to scientific scrutiny. In each case, they ask what is the evidence base for saying that a given therapy “works”? Acupuncture, homeopathy and chiropractic all come out badly. Singh and Ernst build a compelling case that these therapies are at worst positively dangerous - chiropractic neck manipulation can result in injury or death - and at best, are more or less useless. For example, tests done in Germany have shown that “real” acupuncture works no better in easing migraines than sham acupuncture, a random application of wrongly positioned needles, working as a placebo.

Singh and Ernst do not deny that placebos are powerful things. This being so, does it matter if homeopathy really “works” in scientific terms? If it makes me feel better to rub arnica cream into a bruise (notwithstanding the fact that the active ingredient in the cream is so dilute as to be nonexistent), what harm is done? The authors argue that it does matter, for three reasons. First, if, as the evidence base indicates, homeopathy is merely a placebo, then the price tag is a rip-off. They give the example of a single French duck, whose endlessly diluted heart and liver were used to generate $20m worth of homeopathic flu remedies. A second problem lies in the ethics of the doctor-patient relationship. In order to make the placebo effect work, doctors would have to suppress their knowledge that homeopathy was bogus. “In fact, the best way to exploit the placebo effect is to lie excessively to make the pill seem extra-special, by using statements such as ‘this remedy has been imported from Timbuktu' etc.” Third, and most worrying, by putting his or her faith in homeopathy, a patient may fail to seek out more effective conventional treatment. In the case of a minor bruise, this doesn't matter. It's altogether more serious when it comes to asthma, say, or cancer. The authors mention a Devon-based homeopath who used her own medicines to treat a malignant melanoma on her arm and, as they bluntly put it, “condemned herself to an inevitably early death”.

Does this mean that all alternative therapies are to be dismissed? In the case of herbal medicines, Singh and Ernst admit that some are effective, such as devil's claw for musculoskeletal pain, or hawthorn for congestive heart failure, but even here they argue that that, once an alternative treatment passes proper tests, it will be accepted into the mainstream and cease to be alternative. The examples they give are St John's Wort for the treatment of mild depression as well as osteopathy (a gentler alternative to chiropractic) and fish oils for preventing heart disease. They would like to see all alternative medicines jump through the same expensive hoops as mainstream drugs. Until they have passed such tests, they should come with cautions (“Warning: this product is a placebo”), though of course any such warning would work against the placebo effect.

Their case against the folly, vanity and damage of HRH et al is hard to argue with. But it would have been even stronger had they been more honest about the reasons people run so eagerly into the arms of quacks. They admit that, in the 19th century, patients were sometimes better off with homeopathy (ie no treatment at all) than with the mainstream practices of “bloodletting, intestinal purging, vomiting, sweating and blistering, which generally stressed an already weakened body”.

Today's medicine is, of course, infinitely more effective in the treatment of disease. But it is hopelessly primitive when it comes to preventing it. The past 100 years have not been an unmixed triumph for benevolent science. The “evidence-based” medicine that Singh and Ernst are so fond of does not look so great when we consider the profiteering of big pharmaceuticals, who would rather sell us drugs to manage our illnesses than to help us stay well. Alternative medicine flourishes in the space that conventional medicine neglects by focusing so relentlessly on cure rather than prevention. Is it any wonder that some people - against all the evidence - prefer the warm lies of the crystal merchants to the cold drugs of the men in white coats?

Trick or Treatment? Alternative Medicine on Trial by Simon Singh and Edzard Ernst
Bantam Press £16.99 pp352

Times Recommends

• The 100 Best Books of the Decade
• The Times picks the 50 best paperbacks of 2009
• The conversation: James Ellroy