Water on the brain?
by endlesspsych
Homeopaths don’t like the 1023 campaign.
They don’t like it so much that some homeopaths have been tweeting the “evidence” for homeopathy using the #ten23 hashtag.
Nigh on each and every one of the specific documents or claims made by the homeopaths has been addresses elsewhere but one particular theme among the homeopaths comments caught my interest.
The claim that homaopathy cannot be tested using conventional means and that some special sort of science is needed to “validate” homeopathys efficiacy. From what I can tell from the “evidence” frequently provided by homeopaths their special science amounts to sloppily designed randomised clinical trials (often only single blind) presumably so they can say “proven by RCTs” (well no not proven actually because the methodologies used are nearly always shoddy and the results are nearly never replicated in independent labs) or anecdote. The well worn phrase “the plural of anecdote is not data” applies here.
That homeopaths continually cite pseudoscience and anecdote to support their claim that ultrafilute substances actually do something other then nothing leads me towards one conclusion: Nobody has ever sat them down and explained the logic behind evidence based medicine.
Anecdotes are unreliable because people are unreliable. There are all sorts of reasons why people are unreliable judges of their own health. They don’t want to appear a burden, don’t want to admit to themselves that there is a problem – they could be afraid that a visit to the doctor will uncover something serious for instance?
When it comes to testing drugs people can be equally as bad at assessing their effects. This is down to a myriad of psychological effects – the expectation effect for one. Essentially people expect an effect from some sort of remedy and then over interpret or attribute any improvement experienced to the treatment given.
Clearly there needs to be a way of controlling for the unreliable nature of human beings…
…and there is the blind trial. To determine if a drug has a real effect patients are randomly assigned to one of two groups placebo (or control) and treatment (or experimental). The treatment group unknowingly recieved the remedy being tested and the placebo group are given a suitable alternative that it is known has no active ingredients or effect. If there is a greater effect size in the treatment group compared to the placebo group then you can be more sure that the remedy is effective…
Except, well there can be issues where conciously/unconciously the experimentors can bias the results of the study. To avoid this scenario studies can be double blinded.
This is where not only do the patients not know if they are getting a remedy or a placebo but those administering the treatment don’t know either. So there is no way they can bias the results!
Now this tells you if a remedy is better then placebo or more technically it tells you whether there is a significant difference between the two groups. To measure how much of a difference one should really employ an effect size measure of some sort. There is no point lauding a remedy based on statistical significance alone as the effect size it has over placebo could be tiny…
This is a brief explanation of double blind RCTs hopefully folks can see why they are the best way we have to assess medical treatments.
Homeopaths who wish to produce evidence that supports the efficacy of homeopathy should demand the same standards of evidence that all medical treatments are required to go through.
If they showed an effect then the skeptics would have to accept that it worked. However if they don’t the homeopaths would have to accept that it doesn’t.
Cheers
Excellent post.
And I agree – show me irrefutable, peer-reviewed evidence that homeopathy works at 30C – and I’ll believe it works.
But when you’ve got journals like “Journal of Alternative and Complementary Medicine” in existence, where papers are edited and reviewed by people with a vested (financial) interest in getting homoeopathy accepted (or at least generating pro-homeopathy literature), there isn’t any pressure on them to conform to the sort of rigour that you mention i.e. large scale, placebo controlled, randomised, blinded trials.
(for more – see Dr Aust’s great post : http://bit.ly/4qgMtO )
But then how do you create a placebo for a placebo which essentially is what homeopathy is.
‘A placebo performed no better than another type of placebo in this randomised double blind trial’ is the basic conclusion of any decent homeopathic trial!
Indeed.
You find me in complete agreement.
I just saw a worrying link on twitter regarding a single arm inblinded trial that homeopaths are claiming shows it treats malaria!
I mean bloody hell!
Sorry I meant to add that I think it is the general misunderstanding of the power of the placebo effect and what it is that allows these sorts of alternative treatments to prosper.
For example saying to someone that placebo acupuncture works just as well as normal acupuncture is tantamount to saying that it works to the general public. Hence the way a study that concluded that very thing was reported last year in the press as a positive acupuncture trial.
A placebo effect is still an effect afterall and so it can be hard to argue that a certain treatment has NO effect without being as disengenous as those that claim it works better than placebo.
Social comments and analytics for this post…
This post was mentioned on Twitter by endless_psych: http://j.mp/6jLKkf a little blog on EBM and homeopathy #ten23…
“The treatment group unknowingly recieved the remedy being tested and the placebo group are given a suitable alternative that it is known has no active ingredients or effect.”
We’ll give one group water and the other group water, whichever group see’s the most benefit We;ll assume that there is some homeopathic quality to. That Water must at some time since the forming of the universe come into contact with whatever we are trying to sell. Ergo We have proven Homeopathy works. As for the other sample, it was clearly elther too dilute or not dilute enough…