Randomised controlled trials for homoeopathyBMJ 2002; 324 doi: http://dx.doi.org/10.1136/bmj.324.7336.498 (Published 02 March 2002) Cite this as: BMJ 2002;324:498
Who wants to know the results?
- Gene Feder, professor of primary care research and development,
- Tessa Katz, general practitioner
- Department of General Practice and Primary Care, Queen Mary's School of Medicine and Dentistry, London E1 4NS
- Lower Clapton Group Practice, London E5 0PD
Papers p 520
Why should you read about a trial comparing homoeopathic treatment to placebo? If you prescribe homoeopathic medicines a trial will not influence your prescribing decisions because most trials of homoeopathic medicines do not individualise treatment, the hallmark of homoeopathic practice. If they do1 it is difficult to apply the results to individual treatment decisions in practice. Moreover randomisation and blinding of participants substantially distorts the context of homoeopathic prescribing, potentially weakening its effect. If you do not prescribe homoeopathic medicines you will not use the results directly in your practice, so why take any interest in such trials? One reason could be that every year 8.5% of adults in the United Kingdom and 4% in the United States use a homoeopathic medicine.2 It is also possible to refer patients to homoeopathic specialists in the NHS or refer to general practitioners who prescribe homoeopathically within a practice or primary care trust. The number of such referrals is growing.
The study by Lewith and colleagues (p 520) in this issue joins the pool of good quality placebo controlled trials and no doubt will take its place in the next meta-analysis.3 It is a negative trial in patients with asthma, showing no difference in lung function or their asthma-specific quality of life between those treated with placebo and those who received ultradiluted allergen. It is a test of isopathy (the use of homoeopathically prepared allergens to treat allergies), not a test of homoeopathy as such. The study was designed to replicate a previous trial by Reilly et al using the same intervention.4 The main differences between this and previous trials are the outcome measures and duration of treatment, which may account for the different result, although chance is another explanation.
Most trials of homoeopathy have a different function from those in orthodox medicine: their underlying rationale is to test whether homoeopathic medicines have any clinical effect greater than placebo. Without evidence of such an effect, it is difficult for orthodox clinicians to justify referral to homoeopathic services. The use of randomised controlled trials to test the legitimacy of homoeopathic treatments is the latest chapter in an ideological and scientific struggle between homoeopathy and orthodox medicine going back to the 19th century.5 The fervour of this struggle is reflected in the 58 electronic responses to another trial of homoeopathy reported in the BMJ.6
Are the results of placebo controlled trials in homoeopathy convincing? Linde et al's meta-analysis of 89 trials suggests an effect of homoeopathic medicines greater than placebo.7 The aggregated effect size of homoeopathic treatments, when possible publication bias is taken into account or only high quality trials are included, is modest.8 How seriously clinicians take these findings depends on their prior beliefs.9 If you cannot conceive of highly diluted solutions with undetectable drug concentrations having a biological effect, then no matter how well designed the trial or robust the meta-analysis, a positive result will not change your view. If you are less concerned about the integrity of our model of the universe or are intrigued by controversial laboratory work showing the activity of highly diluted histamine solutions10 than the overall positive result of the trials makes it easier to take homoeopathy seriously.
Despite homoeopathy's popularity with patients, orthodox medicine has had the upper hand in terms of institutional support, research funding, and strong evidence of effectiveness. Nevertheless, the flurry of trials in the past 20 years has changed the terms of the debate. At the very least, those who consider homoeopathy to be absurd have had to muster different philosophical and methodological arguments to defend their position. Randomised controlled trials may be efficient arbiters of clinical effectiveness, but they are not particularly good for resolving philosophical disputes.
Current trials are of a high methodological standard and, if positive, may sway agnostics. Opponents of homoeopathy have made it clear that no number of well designed trials showing an effect greater than placebo will overcome their prior belief that homoeopathy cannot work. Research funding is a scarce resource. Unlike other commentators in this journal,11 we believe that new trials of homoeopathic medicines against placebo are no longer a research priority. The question whether ultramolecular dilutions can have any measurable physical effect, a scientific rather than philosophical question, is best tackled with laboratory methods. However, there is still a role for pragmatic trials comparing the effect and cost effectiveness of orthodox and homoeopathic treatments. Within the homoeopathic medical community and other groups that use homoeopathy, such as anthroposophical physicians,12 there is a call for outcome studies to evaluate the individualised treatment decisions that are at the heart of their clinical method and compare outcomes to orthodox treatment.13
GF was paid a consultancy fee by Weleda, a pharmaceutical company that manufactures homoeopathic medicines, to develop prescribing guidelines. TK received a grant from the Blackie Foundation to pilot a randomised controlled trial of homoeopathic treatment.