Memorandum submitted by Edzard Ernst HO 16


Many years ago, I have worked as a homeopath and therefore understand the concepts of homeopathy (e.g. like cures like and dilution increases effectiveness). These concepts are not supported by science and most homeopaths would probably admit that. However, they claim that several in vitro experiments suggest that the homeopathic dilution process does demonstrably alter the structure of water. The counter-arguments are that this only happens for nano-seconds; it also does not explain how such dilutions generate health effects in vivo; nor does it explain why (water-free) globuli used widely in homeopathy should be effective.


Homeopaths claim that, while we do not understand how their remedies work, clinical evidence shows that they work. The truth, however, is that systematic reviews or meta-analyses of the totality of the clinical data fail to show that homeopathic remedies generate clinical effects beyond those of placebo. Homeopaths counter by criticising the methodology of the latest Lancet meta-analyses by Shang et al. This, however, ignores the fact that over a dozen similar systematic evaluations have all come to the same conclusion.


Homeopaths furthermore claim that animal studies also show the effectiveness of homeopathic remedies. This implies that they must be more than placebos. Yet, if one reviews the totality of these data, a picture emerges which closely resembles the human trial data alluded to above.


Homeopaths also often refer to large observational studies, for instance, one from the Bristol Homeopathic Hospital, which imply that ~70% of patients improve after homeopathic treatment. Such "real life" studies and their years of experience, they claim, is more meaningful than clinical trials. Yet the discrepancy between the two sets of results is easy to explain: the patients in observational studies improved because of placebo-effects, regression towards the mean, concomitant treatments and many other confounders. In clinical trials, all of these factors are eliminated and therefore no differences are observed between homeopathic remedies and placebos.


Finally homeopaths claim that their approach is risk-free. This is, clearly not true. Highly dilute homeopathic remedies may well be free of side-effects. But forfeiting or delaying effective treatments, as homeopathy often does, can cause real harm. This issue is not well-researched, except for one particular area: many studies have confirmed that (lay) homeopaths tend to discourage parents from immunizing their children, often recommending using homeopathic vaccinations (which are ineffective) instead.


In conclusion, there is no good evidence to suggest that homeopathic remedies have any specific therapeutic effects and there is some evidence to show that homeopathy can cause harm. Thus its risk-benefit profile is negative.

E Ernst MD, PhD, F Med Sci, FSB, FRCP, FRCPEd


Complementary Medicine

Peninsula Medical School

Universities of Exeter & Plymouth


November 2009