More of Louise Mcleans ‘facts’ about homeopathy , this time about scientific studies of homeopathy. This is an area where the homeopaths want to have their cake and eat it. On the one hand they say, as Mclean does in Fact 36, that homeopathy can never be properly tested in double blind randomised trials because each prescription is individualised to the patient and on the other hand they claim that there have been successful clinical trials.
In fact, individualised homeopathy can be trialled. The trial subjects all go to a homeopath who prescribes individualised homeopathic remedies for each. Half actually receive the prescribed remedy and half receive a placebo. The two groups can then be compared as in any other trial.
In Fact 32 she claims that in 2005 the World Health Organisation (WHO) brought out a draft report which showed that homeopathy was beneficial causing Big Pharma to panic and The Lancet to publish their “The End of Homeopathy” editorial. In fact the editorial was in response to a meta analysis of homeopathy trials (of which more anon) which concluded that homeopathy performed no better than placebo.
This report cannot be found on the WHO website which to Mclean is proof of suppression. It does not seem to occur to her that it was not published because its evidence did not withstand scrutiny. The evidence she puts forward here is that homeopathy is the second most popular type of medicine in the world and therefore it must work. I did not realise scientific facts were decided by popularity contests but even if they were, she should consider that the most popular medicine is evidence based medicine )which she denigrates as nasty poisonous Big Pharma ruled ‘allopathy’). By her own logic, this proves that EBM is totally wonderful. And better than homeopathy. Unfortunately, as already noted, consistancy is not the homeopaths’ strong suit.
In Fact 33 she gets onto this meta-analysis and claims it is fraudulent because 102 poor quality trials were excluded. She of course claims that they were excluded because they were positive. In fact they were excluded because of a number of methodological flaws such as:-
No or inadequate randomisation. By this I mean that the assignment of patients to the homeopathy arm or the placebo arm was not truly random and thus prone to bias such as making sure that the patients in the poorest health wind up in the placebo arm. They are less likely to recover unassisted than the healthier patients in the homeopathy arm so homeopathy is made to look good. (I should say that I am sure some Big Pharma researchers are not above pulling this stunt either.)
Inadequate blinding. This means that either the patients or researchers, or both, know who is receiving homeopathy and who is receiving placebo. This could influence the patients in their reporting of symptoms and allow the researchers to let their own (possibly subconcious) biases influence their interpretation of data.
High drop out rate. Patients who drop out of trials are much more likely to have suffered side effects or failed to improve. The dropouts should be followed up but usually are not because they would make the substance under test look less good. With a small drop out rate this might not matter so much but when it passes 40%, clearly it is going to have a major influence on the outcome.
Small sample size. The fewer patients in your trial, the more likely you are to get a seemingly good result by chance. For example, if you have five patients in the homeopathy arm and five in the placebo arm (believe me, some homeopathic trials are this bad) with three of the former and two of the latter recovering, you could trumpet this as “homeopathy 50% better than placebo!” when it is fairly obvious that such a result could come about by mere chance.
She continues with the notion that homeopathy has been clinically proven in Fact 34 by referring to “180 controlled and 118 randomised trials” meta-analysis of which seemingly proved that homeopathy performed better than placebo. No link or reference so we cannot check for ourselves. We certainly cannot check the sample size, which, as noted above, is pretty important.
There are two ways of interpreting the numbers she does give. Does she mean there were a total of 298 trials of which 180 were controlled but not randomised and 118 randomised but not controlled? If she does, they were all poor quality trials open to bias and meta-analysis of them is pretty meaningless. Being generous, perhaps she means there were a total of 180 controlled trials of which 118 were also randomised. This would mean that 62 (more than a third) were not randomised which could therefore produce false positives by the means outlined above. This would be sufficient to skew the result of any meta-analysis so they should have been excluded. They were not, rendering the analysis worthless.
In Fact 35 she references the Bristol Homeopathic Hospital ‘study’ which apparantly showed an overall improvement in the health of 70% of patients. In fact, this ‘study’ had no control group and was merely a satisfaction survey. Given that the homeopathic hospitals confine themselves to treating self-limiting illnesses, I am surprised the result was not nearer 100%.
In Fact 36, as previously noted, she adopts the classic line that if double blinded randomised trials don’t show that homeopathy works, then there must be something wrong with the approach. This attitude is not confined to CAM. Audiophiles reject “blind-listening” trials which show that their £1,000 gold-wired speaker cables do not improve sound quality.
In Fact 37 she tells us that homeopathic medicines are not tested on animals. Fair enough, I am not going to argue this point except to note that even if they were, the animals would not be harmed by them unless they were lactose intolerant.
In Fact 38 she claims that homeopathic medicines work even better on animals and babies than on adults. Since homeopathic medicines are not tested on animals I do not see how she can possibly know this.
In Fact 39 she notes that if homeopathy is ever proven to work it will turn established science on its head. This is true because our current undestanding cannot explain how the dilution and succusion process can make water ‘remember’ the remedy and not anything else with which it has been in contact; nor can it explain how, when a drop of this magic potentised water is placed on a sugar pill it can evaporate and leave its memory behind. However, our current understanding does explain many observed phenomena so scientists are reluctant to abandon it because it cannot explain phenomena that only exist in the minds of homeopaths.
If there is any credible evidence that homeopathy works better than placebo, it has yet to be published