I recently wrote about Dr John Briffa, making reference to his approving comments about a study into arnica as a post-operative aid. I had an idea that homeopathic treatments like arnica relied on the placebo effect and was surprised to see that Briffa’s post described arnica as “effective”. There was some discussion of the actual paper on the badscience.net forum and Coracle wrote a very good blog post about it (which, unfortunately, seems to be down as I write this). It was here: but the blog has been removed, which is a shame. Superburger posted a comment in the Bad Science forum thread pointing out that more arnica pills than diclofenac pills were given [wondering if taking 30 ‘pillules’ a day of homeopathic arnica might result in a genuine dose of arnica getting through – perhaps not such a homeopathic remedy after all]. I thought about the placebo effect and remembered (OK, half-remembered) something about the colour of pills and the number of pills “modifying the strength of placebo effect” or somesuch thing.
Coracle referred in this post to the possibility that “if there is a placebo effect inherent in giving a pill, might it not be stronger in giving ten? Another issue not addressed by the authors” and I found a couple of papers that referred to the number of pills and the strength of a placebo effect. The summary of this paper begins: “The hypothesis that clinical improvement would be significantly correlated with number of daily placebo pills prescribed was supported for clinic and general practice patients but not for private psychiatric practice patients.” It goes on to state:
Our findings in this study have methodological as well as clinical implications. The fact that higher placebo intake levels resulted in reduced drug-placebo differences in improvement poses a practical problem for the clinical researcher, while the lack of placebo improvement observed at lower placebo intake levels indicates that 1 placebo pill per day is not a very effective agent in the symptomatic treatment of neurotic outpatients. Further research is needed to determine the optimal dosage for placebo therapy.
I also found this in another paper:
That placebo effects of greater magnitude are achieved by more believable and technically sophisticated agents seems to support this idea. For instance, placebo injections elicit greater placebo responses than placebo pills, and larger pills are associated with stronger placebo responses than smaller pills; the number of pills taken is also directly related to the magnitude of the placebo response
The reference for this is to a paper that does not, unfortunately, have an abstract (not sure if that should worry me or not, but if I’d published a paper I’d want people to be able to read a summary of it – I’d think they should be able to get an idea of the methods I’d used, for one thing). The reference is to: Ross M, Olson JM. An expectancy-attribution model of the effects of placebos. Psychol Rev 1981;88:408–37 [pubmed linky]. I reckon it’s quite possible (although I’m not sure how likely it actually is) that the reason that this arnica seemed to be positive is that there was a greater placebo response for the thirty pillules per day than for the three capsules per day. Which brings me onto another point (or five)…
Not from me, but from wilsontown: “As Superburger points out, the homeopathic group is taking 10 pillules of the remedy three times a day. From the abstract, it seems that the diclofenac group was getting one 50 mg pill three times a day. If so, then the blinding is potentially broken.” Oops. He also goes on to point out that “There is no placebo group. They claim that arnica is superior for patient mobility, but this is based on a “descriptive analysis” and is of marginal significance (p=0.045). The study is small-ish, with a total of 88 subjects.” This paper doesn’t seem to be strong evidence for arnica being “effective post operative aid” to me – in fact, the evidence is positively homeopathic. Funnily enough, and quite apart from anything else I’ve mentioned, diclofenac is a painkiller and it actually outperformed arnica in terms of pain relief.