Recently an interesting article caught my eye. It was published in the official journal of the ‘Deutscher Zentralverein Homoeopathischer Aerzte’ (the professional body of German doctor homeopath which mostly acts as a lobby group). Unfortunately it is in German – but I will try to take you through what I believe to be the most important issue.
The article seems to have the aim to defame Natalie Grams, the homeopath who had the courage to change her mind about homeopathy and to even write a book about her transformation. This book impressed me so much that I wrote a post about it when it was first published. The book did, however, not impress her ex-colleagues. Consequently the book review by the German lobbyists is full of personal attacks and almost devoid of credible facts.
A central claim of the defamatory piece is that, contrary to what she claims in her book, homeopathy is supported by sound evidence. Here is the crucial quote: Meta-Analysen von Kleijnen (1991), Linde (1997), Cucherat (2000) und Mathie (2014) [liefern] allesamt positive Ergebnisse zur Wirksamkeit der Homöopathie… This translates as follows: meta-analyses of Kleijnen, Linde, Cucherat and Matie all provide positive results regarding the effectiveness of homeopathy. As this is a claim, we hear ad nauseam whenever we discuss the issue with homeopathy (in the UK, most homeopathic bodies and even the Queen’s homeopath, P Fisher, have issued very similar statements), it may be worth addressing it once and for all.
This paper was the result of an EU-funded project in which I was involved as well; I therefore know about it first hand. The meta-analysis itself is quite odd in that it simply averages the p-values of all the included studies and thus provides a new overall p-value across all trials. As far as I know, this is not an accepted meta-analytic method and seems rather a lazy way of doing the job. The man on our EU committee was its senior author, professor Boissel, who did certainly not present it to us as a positive result for homeopathy (even Peter Fisher who also was a panel member should be able to confirm this). What is more, the published conclusions are not nearly as positive as out lobbyists seem to think: ‘There is some evidence that homeopathic treatments are more effective than placebo; however, the strength of this evidence is low because of the low methodological quality of the trials. Studies of high methodological quality were more likely to be negative than the lower quality studies. Further high quality studies are needed to confirm these results.’
The meta-analysis by Linde et al seems to be the flag-ship in the homeopathic fleet. For those who don’t know it, here is its abstract in full:
BACKGROUND: Homeopathy seems scientifically implausible, but has widespread use. We aimed to assess whether the clinical effect reported in randomised controlled trials of homeopathic remedies is equivalent to that reported for placebo.
METHODS: We sought studies from computerised bibliographies and contracts with researchers, institutions, manufacturers, individual collectors, homeopathic conference proceedings, and books. We included all languages. Double-blind and/or randomised placebo-controlled trials of clinical conditions were considered. Our review of 185 trials identified 119 that met the inclusion criteria. 89 had adequate data for meta-analysis, and two sets of trial were used to assess reproducibility. Two reviewers assessed study quality with two scales and extracted data for information on clinical condition, homeopathy type, dilution, “remedy”, population, and outcomes.
FINDINGS: The combined odds ratio for the 89 studies entered into the main meta-analysis was 2.45 (95% CI 2.05, 2.93) in favour of homeopathy. The odds ratio for the 26 good-quality studies was 1.66 (1.33, 2.08), and that corrected for publication bias was 1.78 (1.03, 3.10). Four studies on the effects of a single remedy on seasonal allergies had a pooled odds ratio for ocular symptoms at 4 weeks of 2.03 (1.51, 2.74). Five studies on postoperative ileus had a pooled mean effect-size-difference of -0.22 standard deviations (95% CI -0.36, -0.09) for flatus, and -0.18 SDs (-0.33, -0.03) for stool (both p < 0.05).
INTERPRETATION: The results of our meta-analysis are not compatible with the hypothesis that the clinical effects of homeopathy are completely due to placebo. However, we found insufficient evidence from these studies that homeopathy is clearly efficacious for any single clinical condition. Further research on homeopathy is warranted provided it is rigorous and systematic.
Again, the conclusions are not nearly as strongly in favour of homeopaths as the German lobby group assumes. Moreover, this paper has been extensively criticised for a wide range of reasons which I shall not have to repeat here. However, one point is often over-looked: this is not an assessment of RCTs, it is an analysis of studies which were double-blind and/or randomised and placebo-controlled. This means that it includes trials that were not randomised and studies that were not double-blind.
But this is just by the way. What seems much more important is the fact that, in response to the plethora of criticism to their article, the same authors published a re-analysis of exactly the same data-set two years later. Having considered the caveats and limitations more carefully, they now concluded that ‘in the study set investigated, there was clear evidence that studies with better methodological quality tended to yield less positive results.’
It is most intriguing to see how homeopaths cite their ‘flagship’ on virtually every possible occasion, while forgetting that a quasi correction has been published which puts the prior conclusions in a very different light !
The much-cited article by Kleijnen is now far too old to be truly relevant. It includes not even half of the trials available today. But, for what it’s worth, here are Kleijnen’s conclusions: At the moment the evidence of clinical trials is positive but not sufficient to draw definitive conclusions because most trials are of low methodological quality and because of the unknown role of publication bias. This indicates that there is a legitimate case for further evaluation of homoeopathy, but only by means of well performed trials.
If the homeopathy lobby today proclaims that this paper constitutes proof of efficacy, they are in my view deliberately misleading the public.
The Mathie meta-analysis has been extensively discussed on this blog (see here and here). It is not an overall meta-analysis but merely evaluates the subset of those trials that employed individualised homeopathy. Crucially, it omits the two most rigorous studies which happen to be negative. Its conclusions are as follows: ‘Medicines prescribed in individualised homeopathy may have small, specific treatment effects. Findings are consistent with sub-group data available in a previous ‘global’ systematic review. The low or unclear overall quality of the evidence prompts caution in interpreting the findings. New high-quality RCT research is necessary to enable more decisive interpretation.’
Again, I would suggest that anyone who interprets this as stating that this provides ‘positive results regarding the effectiveness of homeopathy’ is not telling the truth.
MY CONCLUSIONS FROM ALL THIS:
- Some systematic reviews and meta-analyses do indeed suggest that the trial data are positive. However, they all caution that such a result might be false-positive.
- None of these papers provide anything near a proof for the effectiveness of homeopathy.
- Homeopathy has not been shown to be more than a placebo therapy.
- To issue statements to the contrary is dishonest.