There is borderline evidence that acupuncture may be useful for pain and nausea, but no evidence of effectiveness for anything else. Trials here are complicated by the difficulty of finding good placebos for acupuncture treatment. This chapter introduces the concept of a placebo, the idea of meta-analyses combining multiple studies, and the Cochrane Collaboration, which is dedicated to high-quality systematic reviews of the effectiveness of different medical treatments.
The chapter on homeopathy spends a lot of time on its history — it was one of the many kooky endeavours favoured by Nazi science — and attempts, in the face of pretty much all of modern physics and chemistry, to find some kind of underlying mechanism for it. The actual evaluation of homeopathy in high quality trials has been done and is quite clear-cut: there is no evidence that any homeopathic treatment has any efficacy for anything. Homeopathic remedies are very expensive placebos, which may be dangerous where they displace treatments that actually work.
Chiropractic therapy has some efficacy for back pain, but not for anything else. It can also, however, be quite dangerous, and should only be used with care. Among other problems, chiropractic practitioners massively over-treat patients, with investigations showing that most will happily prescribe unnecessary X-rays and costly and sometimes dangerous courses of treatment, even for completely healthy patients with transient minor ailments.
There is good evidence for the efficacy of many herbal medicines. In comparison to conventional medical drugs, however — and perhaps almost by definition — they are poorly tested and regulated. Many are dangerous, either directly, through interactions with other medicines, or through contamination (high concentrations of heavy metals are common in Chinese and Ayurvedic remedies), and many are used for inappropriate conditions. Singh and Ernst provide a two page list of common remedies rating the evidence for their effectiveness as "good", "poor", or "none", as well as two pages summarising some of the risks of different herbal medicines; they also provide suggestions for finding good quality suppliers.
The chapter on herbal medicines also looks at some of the reasons people are attracted to alternative medicine ("natural", "traditional", "holistic"), along with the ways practitioners both attack science and pillage it for support, and some of the reasons both patients and practitioners can be misled into thinking they work (such as regression to the mean and confirmation bias). A final chapter considers a range of other general issues: the culprits responsible for the popularity of discredited and dangerous therapies, ways of improving conventional medicine, and an approach to regulating alternative therapies better.
An appendix provides one page summaries of the evidence for thirty six other alternative therapies, going alphabetically from Alexander Therapy to Traditional Chinese Medicine.
Trick or Treatment? works well read through, but the chapter structure is a little confusing. It doesn't really work to read just the introduction and the chapter on homeopathy, for example, since the key concept of a placebo is explained in the chapter on acupuncture. And the chapter on herbal medicine contains unrelated material on tests of prayer efficacy.
Trick or Treatment? will be a better book for most readers than Bausell's Snake Oil Science, which emphasizes epistemology — abstract problems of knowledge — more and history less, and which may be better suited to doctors or other medical practitioners. For those who fall somewhere between unthinking rejection and unthinking acceptance of alternative medicine, Ernst (a professor of complementary medicine) and Singh (a science writer) will be useful guides.