EBM and homoeopathy

This letter was published in the Br Med J (8 Sept, 2000) as a “rapid response” to a paper1 on homoeopathy.


The medical community has an increasing concern about the quality2 and fairness3 of reporting of clinical research, and experience as well as repeated stories suggest that, whatever their source, overstatements regarding the benefits or risks of therapeutic options may have serious consequences in terms of cost, safety or community health. In this context, one may be surprised by the BMJ commenting the work by Taylor et al. as “evidence” favouring homoeopathy over placebo (p. 518) and, more generally, by the presentation of this paper as a significant one.

Actually, the first “evidence” is that the study in question was a negative one as its results were clearly comparable for homoeopathy and placebo regarding the primary efficacy parameter. Additional evidence includes a relative weakness in experimental design, since the use of concomitant drugs was left to the patient’s discretion : greater use in vasoconstrictors and antihistamine is all the more likely to account for the reported improvement in nasal inspiratory peak flow that initial aggravation of rhinitis symptoms in the homoeopathy group is likely to have worsened the baseline imbalance regarding the use of these agents (see the authors’ Table 1). Concerning the experimental design on the other hand, one may remark that in an indication such as perennial rhinitis, the usual 4 efficacy parameters are generally the consumption of anti-allergic drugs and symptoms scores, and not visual analogue scales (VAS); likewise, one may wonder whether a total duration of 4 weeks might have the slightest clinical relevance in such a chronic disease.

Finally, this is an elementary notion of the meta-analytic methodology that pooling data from some clinical trials (i.e. without any systematic method of selection) can never be seen as “evidence” of anything – especially when this pooling includes heterogeneous diseases, some of them acute (hay fever) and some chronic (asthma, perennial rhinitis); incidentally, it is uncertain whether VAS is robust parameter to assess efficacy in a disease such as asthma…

It may be of interest to note that similar questions on selection biases as well as clinical relevance were raised about a previous review published in the Journal and also presented as suggesting an efficacy of homeopathy5. In contrast to physics, chemistry or biology, most of clinical research cannot be reproduced (due to obvious problems of recruitment, duration of treatment, organisation, cost, etc.) and this is why we have a professional obsession of fraud, misleading reports or misstatements: this is one of the greatest challenge and responsibility for experts to develop, as a sixth sense, their ability to detect every indirect sign of fallacy. In the present case, the issue is by no means any suspicion of fraud or misconduct: but when published evidence itself does not support the authors’ discussion, it is not expected that Editors will amplify overstatement by inappropriate comments.

We need far more solid results prior to rejecting “the hypothesis that homoeopathy in a placebo response” i.e. before challenging our current understanding of physical world: for the time being, “evidence” from modern clinical research is still that homoeopathy results from an anachronistic persistence of outdated medical misconceptions.

  1. Taylor MA, Reilly D, Llewellyn-Jones RH, McSharry C, Aitchison TC. Randomised controlled trial of homoeopathy versus placebo in perennial allergic rhinitis with overview of four trial series. BMJ 2000; 321: 471- 476.
  2. Altman DG. Better reporting of randomised controlled trials: the CONSORT statement. BMJ 1996; 313: 570-1.
  3. Rennie D. Fari conduct and fair reporting of clinical trials. JAMA 1999; 282: 1766-9.
  4. Malling HJ. Immunotherapy as an effective tool in allergy treatment. Allergy 1998; 53: 461-472.
  5. Girard M. Trials of homoeopathy. BMJ 1991; 302: 727.