Homeopaths are feeling under threat at the moment and are scrambling around wondering what to do about it. I think there are a number of things they could do: most importantly, they should better manage their own business by showing that they respect the boundaries of what they can reasonably assume is good practice, e.g. stop the dangerous nonsense of believing that they can do anything about dangerous conditions such as malaria and AIDS; they can be much more complementary and less alternative.
But there is something else that they can do too: start showing a desire to develop a base of data that can be relied upon, and respected, to support their methods. The focus to-date has been on clinical trials. Doubters say that trials show no evidence of efficacy. Supporters point to many positive trials. But it has been well documented that the many positive trials are most often poorly designed and reported, and are at best ambiguous in their results. There is not a compelling evidence base for homeopathy. If there was, there would be no argument.
So, let’s take a step back. What sort of evidence would be required to convince me that there might be something in it? Fundamentally, my problem with homeopathy is its total implausibility – it contradicts what we know about the behaviour of matter. How can a plain sugar pill have any significant therapeutic effect on health? So, why not test the basic plausibility of homeopathy – can homeopathy pills do more than sugar pills in predictable way? There are a number of discussions about this going on in bloggerland and I would like to pick up on these and set a challenge…
Here is a rough outline of the sort of test I would like to see done…
- A trained homeopath selects six homeopathic remedies of any type and strength.
- The remedies are posted to an independent third party who removes the labels and replaces them with a code letter, A, B, C…F, and posts them back.
- The homeopath takes each remedy in turn and notes the ‘totality of symptoms’.
- The homeopath writes down which remedy corresponds to which code letter.
- The third party ‘breaks the code’ and we note how many are right.
Pretty simple stuff. If the homeopath got all six right, then the odds of that being a fluke would be 1:720. (six factorial). This is far more significant than the typical outcome of a clinical trial, where the odds of a fluke result are more like 1 in 20. It would be pretty compelling if done fairly and a good start to building some real evidence.Now, admittedly, this is not a full trial of homeopathy. It does not test the ‘like-cures-like’ part of homeopathic ‘theory’ and so does not demonstrate that homeopathy can be used to treat illness. But it does somewhat get over the hurdle of total implausibility. What this trial is testing is similar to what is going on in homeopathic provings – the supposedly predictable effects of a remedy on a healthy individual.
Would a trial like this convince me? Well, no single scientific experiment should convince anyone of anything. (There is always the possibility of experimental error or fraud in any experiment.) But a test like this would certainly get my attention. Rarely do experiments start with a ‘big bang’ and all encompassing approach. Most often, preliminary tests are done, ‘proof of concept’ runs and so on. If this worked , then it could easily be replicated by other homeopaths. Larger versions done and properly written up for a journal. More stringent statistical tests could be set. Then, I think all sceptics would have to admit that the principle of homeopathic potentized remedies has merit.
This test is not totally fool proof. I could think of a few ways of cheating; some more devious than others. Do we think the odds of a homeopath cheating be more or less than 720:1? Nonetheless, I think it is a simple and good start that could be done with almost no money and would get the ball rolling. More rigorous tests along the same lines could take place afterwards. Conversely, should the test fail, then homeopaths would have a lot of explaining to do.The great thing about this test is that it could be done with very little money. The actual costs would be a few pounds for some remedies and postage, and some volunteers’ time. I doubt it would cost for than £50-60 (About $100). No need for the millions that ‘Big Pharma’ has. And, unlike a clinical trial, there are very few ethical issues – at least, no greater ethical issues than a homeopathic proving. This test is well within the means of a small group of homeopaths who wanted to show the world that they were not deluded. Homeopaths want to be taken seriously. Here is a good start. It’s the $100 Challenge – that is all it would cost.
What is surprising to me is that I can find no instance of a test like this being done before. I would have thought that this was pretty fundamental – can homeopaths determine the effects of a remedy under blinded conditions? One would have thought that this would have been a staple experiment done at homeopathy school. If any homeopaths can enlighten me as to why this has never been done, then please tell me.
So – the challenge: do any homeopaths want to give this a go? All I would ask is that you do this in the true spirit of enquiry and are open and honest about this. What I mean is that if you want to try this challenge, please follow a few simple guidelines:
- Tell the world in advance that you are going to do this. Post your intention on a blog or web site, tell the world what you are going to do, be open to suggestions about how to simplify and make it a fair test. The more detail you publish, the more trust you will have. Remember, sceptics have a problem with trust of homeopaths.
- State in advance what you think would be a successful result and any caveats you may have. Think of ways in which the trial may go wrong in advance, and make efforts to minimise those risks. None of us want excuses afterwards if it does not go well.
- Find a genuine independent third party – someone with no stake in the outcome. Publish who they are and ensure they are happy to field questions from people after the trial. (People will want to know that protocol was followed).
- Publish your results on the web before the code is broken to reveal how well you have done.
Feel free to jig around with the form of the trial. Add extra homeopaths or remedies if you like. Pick whatever remedies you think will maximise your chances of success. As long as the central rule of running the trial totally blinded (only the third party knows the code) then most variants ought to be fair. But publish what you intend to do so that others can judge the fairness of the test. Be open to comments and suggestions about how to make it a fair test. The most important thing you can do, if you want to impress the sceptics, is to convince people the test was properly blinded – that is, there was no way that the testers could know or guess which remedy they were taking.I think such a trial could be conducted in a week or two. The hardest part may be finding a third party. For the record, I am willing to act as that party. The sceptics will trust me – but the problem is that I suspect the homeopaths may think I will cheat and expose the remedies to moth balls or some other spoiler. I would suggest you could use a local newspaper editor, a GP (you do work with them and trust them, don’t you?), a priest or local politician. Basically, someone with no interest in the result and a reputation to lose if they cheated.
I see no reason why a trial like this could not be done. Instead of lots of homeopathic whining about how the sceptics are picking on them, this trial would be a big step forward in proving your case. I can see many homeopaths taking the line, “Why should I do this? I see proof in my practice every day”. If that voice is you, then rest assured the critics of homeopathy will not go away, because there is every reason to believe you are been fooled by the placebo effect, regression to the mean, and wishful thinking. They will see you as dangerously deluded.