A Brief Investigation of Homeopathic Research

In the post “Why I Oppose Homeopathy” I referred to the research that the Maun Homeopathy Project cited to support their assertions. I noted that hey did not give internet links and that I did not subscibe to the journals in question so I was unable to comment in detail on the claims.

Since then, a reader has kindly emailed me copies of two of the papers.  One (Are the clinical effects of homeopathy placebo effects? A meta-analysis of placebo-controlled trials (The Lancet Vol 350, 20 September 1997 pp 834 – 843))appears to be the one that the homeopaths cite as evidence that homeopathy is two and a half times more effective than placebo.

It is true that the findings are that the combined odds ratio for the 89 studies entered into the meta-analysis is 2.45 in favour of homeopathy. However, when the 26 good quality studies are considered this drops to 1.66. Having said that, the authors’ interpretation is that the results are not compatible with the hypothesis that the clinical effects of homeopathy are completely due to placebo.

Unfortunately many of these trials involved self-limiting conditions, ie conditions where the patients will recover even if nothing is done. People recover at different rates so in a trial it is possible that one group happens to contain more fast recoveres than the other. When the trial also has a small number of participants, as many of these do, statistical artifice cannot be ruled out.

The authors conclude their interpretation by saying:-

“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.”

This was ten years ago.

The second paper is Homeopathy in HIV infection: A trial report of double-blind placebo controlled study (British Homeopathic Journal (1999) 88, pp 49 – 57). This studied 100 HIV+ cases, divided into the asymtomatic strata and the PGL strata. In each strata, half were placed in the placebo group and half in the verum group.

During the course of the trial, a number of participants dropped out so by the end the asymptomatic strata had 23 in the placebo group and 19 in the verum group. The PGL strata ended with 18 in the placebo group and 20 in the verum group. Again these small numbers mean the results are subject to statistical artifice.

Futhermore, one of the PGL placebo group dropped out because of the need for active therapeutic intervention for non responding febrile illness. Two of the PGL verum group dropped out because they developed hepatitis during the course of the study and wisely attended for other treatment. In other words, three of the participants were dropped from the analysis because they had got worse. Given that two of the three were receiving the homeopathic treatment, clearly their removal from already small groups distorted the results in favour of homeopathy.

This is some way from the rigorous and systematic research called for by the authors of the Lancet paper. Homeopaths always demand that someone else do the big and thus expensive trials that would be needed. But why shouldn’t homeopaths finance them. Pharmaceutical companies have to pay for their own trials, and homeopathy is big business these days, not the cottage industry it likes to pretend to be. Furthermore, Big Pharma has to register its trials in advance to avoid publication bias (the cherry picking of positive results). No such requirement is placed on homeopaths. Perhaps the research has been done, proved to be negative and thus “lost”? How would we know?

Tags: homeopathy, Maun Homeopathy Project, research