The Cochrane Library has published a new review of the effects of homeopathy on cancer patients**. Its conclusion is that “there is limited evidence that homeopathic remedies ease the side effects of cancer treatments, but they at least seem to cause no serious adverse effects or drug interactions.” The Quackometer’s response is “No shit, Sherlock!”
Homeopathy is the application of nothing.* It is therefore rather likely that pills with nothing in them will have no effect. Why we need a review of the application of nothing from one of the most respected evidence-based medical organisations in the world is rather beyond me.
Medical science is quite a complicated thing. Testing treatments can be quite hard when humans and illnesses can be so variable. The convention is to accept we have a positive result when the chances of it being a false positive (i.e. just a chance result that looks like a positive result) is 1 in 20. If you do a hundred large and very well controlled trials then 5 of them on average will be giving you incorrect information. If you accept into the mix lots of lower quality trials that do not have all the checks and balances in them then many more of those 100 trials will be misleading. If researchers fail to publish their negative results and only the few ‘lucky’ positive results get through, then the evidence base can easily look like it supports a treatment when in fact it is ineffective.
The Cochrane reviews of various treatments take great care in taking these sources of error into account when examining evidence. However, I believe the approach it takes is prone to problems when it investigates highly implausible and pseudoscientific treatments like homeopathy. I have previously written that,
Firstly, and most importantly, to all intents and purposes, clinical trials of highly implausible treatments, such as homeopathy, can never be used as evidence of their efficacy. No matter how good the statistical result of a trial, or how much data is analysed in a meta-analysis, the probability will always be greater that we are just analysing flawed data rather than there being a real effect. Homeopaths complain that sceptics never accept that trial data is proof of the effectiveness of homeopathy. This approach shows that homeopaths are quite right in their fears, although sceptics ought to be careful to point out that it is not because there is no evidence, but rather than the available evidence falls far short of any meaningful threshold of acceptance. Without a degree of plausibility, homeopaths are asking scientists to believe in the daily occurrence of miracles, and that will not do.
The evidence-based medicine approach of Cochrane takes on the philosophical position that we should not worry about the mechanisms of a treatment. If we can show that a treatment works, then explaining how it works is secondary to successfully treating patients. I have strongly argued that this approach really only works when the treatment carries sufficient degrees of plausibility from prior knowledge that the conventional acceptance criteria of clinical trials will add to our understanding. If a positive result is obtained, but that result is more likely to be due to imperfect data, fraud or publication bias than being a genuine effect, then we will not have gained any new knowledge.
So, applying the standard principles of clinical evidence to highly implausible treatments can result in misleading information being generated. I fear that will happen with this review, for not all parts of it were entirely negative. Eight trials were examined – six were negative. But two showed an effect. Topical calendula appeared to lessen side effects from radiotherapy and an proprietary homeopathic mouthwash, Traumeel S, appeared to relief mouth inflammation during chemotherapy. Were these effects real? Given that homeopathic treatments contain no active ingredients, it would appear to be highly unlikely. The review authors are cautious and call for these trials to be replicated before any clinical recommendations can be made.
Despite this caution, we can predict two responses from the homeopathic industry:
1) Homeopathy has been ‘proven’ to be able ‘coexist’ with ‘conventional’ treatments without creating side effects.
2) Some treatments have been ‘shown by Cochrane’ to be effective for cancer patients.
Both these statements will be misleading, but the Integrative Health movement will be issuing statements to this effect without a shadow of a doubt. The Princes Trust for Integrated Health, headed by Dr Michael Dixon, advocate the use of nonsense treatments alongside real medicine. Reviews like this are grist to the mill, despite their overall negative conclusions. Selective quotation of evidence is bread and butter to ‘integrated health’. Traumeel S contains two of Prince Charles favourites, St John’s wort and Echinacea, but in homeopathic, non existent, quantities.
What is unfortunate is that it might well be quite legitimate to discuss the benefits and risks of offering inert treatments, like homeopathy, to cancer patients if the patient feels they get benefit through some sort of placebo effect. But this is not a debate about clinical evidence, it is a debate about clinical ethics – is it acceptable to tell a patient a pill may work for them when we know it is just a placebo? Indeed, the Princes Trust could hold a perfectly acceptable debate about this subject, but whilst they persist in their fantasies that magic sugar pills can cure where medicine cannot, they rule themselves out of being taken seriously.
The debate about homeopathy needs to be moved away from the serious arena of clinical trials in evidence based medicine and into the arena of medical ethics. That is the only route for homeopathy to survive, but I doubt that there is anyone mature enough in the world of alternative (or even integrative) medicine to take that step.
*************************************************************************************************************
* I have since found out that two of the homeopathic treatments reviewed were atypical homeopathic remedies and were undiluted. These were the two that have shown preliminary and tentative positive results. Do these treatments then really work? Well, at least they have some plausibility – but as the reviewers state, the positive results from these small studies need to be independently replicated. What is the betting that they will not but homeopaths will take this as all the evidence they need?
** Here is the Cochrane review…
http://www3.interscience.wiley.com/homepages/106568753/CD004845.pdf
PPS It has also just been pointed out to me that the paper on the topical cream dos not even mention the word homeopathy in it.
http://jco.ascopubs.org/cgi/content/full/22/8/1447
*******************************************************************************
Follow Up.
This is how the completely batshit site What Doctor’s Don’t Tell You said about the review:
Homeopathy relieves side effects of cancer therapies
15 April 2009
Homeopathy isn’t quite the quack medicine its critics claim. It can help relieve some of the side effects of cancer therapy, the prestigious Cochrane Collaboration has discovered.
Calendula ointment eases skin irritations after radiotherapy, and Traumeel, a combination of 14 homeopathic medicines, helps relieve mouth sores caused by chemotherapy.
The Cochrane researchers found eight studies, involving 664 participants, which produced valid results.
Don’t say I didn’t warn you.
On reading the title my first thought reflected your Sherlock reference. It went something along the lines of “hardly fucking surprising is it”.
I assume “Topical calendula” is some sort of cream or ointment. I also assume that “Traumeel S mouthwash” is some sort of standard mouthwash with some woo in it (or rather not in it I suppose given the circumstances).
Would not any cream or any mouthwash have had similar effects ? Was the pixy dust stuff compared against Nivea or Plax (other creams and mouthwashes are available).
I would have thought that virtually any mouthwash on the market
(especially the “medicated” ones like Corsodyl) would have some beneficial effect in relieving mouth inflammation (especially that lurid purple stuff the dentist gives you – and he has the added benefit of a proper surgery, white coat and face mask to improve the effect [not to mention an actual medical qualification]).
All we are seeing here is the benefit of a cream and a mouthwash, nothing else of any medical importance.
You are quite correct in thinking that the quacks will be all over this like a rash (but at least they will have a cream to cure it).
“I have since found out that two of the homeopathic treatments reviewed were atypical homeopathic remedies and were undiluted”
In what way can they then be considered homeopathic? Surely if they are not diluted, then they’re not homeopathic.
Good question Simon and not always easy to answer since homeopaths cannot agree on anything as they have no standards of evidence.
But, I guess most homeopaths would agree that the first principle of homeopathy is ‘like-cures’like’. That is, if a cream can give you horrible skin blisters then it can be used to cure horrible skin blisters (their rule, not mine). The problem is though, and it is an obvious one, that this is demonstrably wrong. If I have a blistered arm, I do not want more blisters, thankyou. To get around this homeopaths have developed various ‘delivery mechanisms’ of the cure. Dilutions to extremes will obviously remove all side effects (and effects) and is the most common ‘delivery mechanism’. Some maverick homeopaths use MP3 files or emails (don’t ask me how).
What is in these creams and mouthwashes is anyone’s guess really. Homeopaths are so deluded about about basic chemistry that it really is difficult to take their word for anything. I would argue that doing a clinical trial involving homeopaths is prima facie unethical for this very reason.
Homeopathy ‘eases cancer therapy’ according to the BBC
http://news.bbc.co.uk/1/hi/health/7998884.stm
Unsurprisingly it’s on the ‘most shared’ stories list
LCN, I entirely agree that the prior probability of homeopathy working is so low that no conceivable trial will convince me otherwise.
There is a tactical problem, though, with employing this argument rhetorically in that we can end up sounding very like our opponents. I have come across views from them that are explicit in asserting that because they axiomatically assume that homeopathy works, any negative trial data must be explained away as the consequence of a bad trial: not conforming to the ever-mobile rules of homeopathy; leading to quantum-entanglement of treated and untreated groups; or [insert further random feeble excuses ad infinitum].
I find this symmetry troubling and am not sure how best to deal with it. Essentially, we know we are right in that we have all of biology, chemistry and physics to support our estimation of prior probability, whereas they have the ravings of a 19th Century quack and the ramblings of subsequent generations of middle-class fools whose disposable wealth far exceeds their number of brain cells. But, it is a subtle problem to show that we are fighting fairly when we dismiss their evidence out of hand but demand that they accept ours.
If we present our argument as being based utterly in the implausibility of the mechanisms of homeopathy, it is too easy for them to reply that we don’t need to understand how a thing works to know that it does work, which turns us back to a demand for yet more trials. This is not quite fair, because we are arguing that our understanding of basic mechanisms tell us that homeopathy cannot work, which is not the same as allowing it to be an unexplained void in our knowledge.
One of the ways I have addressed this rhetorical problem over the years is to focus on the internal contradictions in homeopathy- take it seriously according to its own rules and then show that homeopaths hold mutually incompatible views about many aspects of their art. This can be a tricky argument to make stick and always leaves them a chink of light that somewhere, sometime there is a True Homeopathy that could be revealed if only the surface dross could be cleared away.
What this all boils down to is the gap between being right and showing that we are right in the forum of public debate. We also need to remember that we have several different audiences- the uncommitted public, our professional peers and regulators, the casual users of woo and the True Bleevers- and that different mixtures of our different arguments are appropriate depending on the audience
I no longer trust the Cochrane database as a source of reliable information after recently reading the details of two systematic reviews of acupuncture in the prophylaxis of headache and migraine and discovering that their positive conclusion was not justified by the results obtained, and that the authors all had vested interests in acupuncture.
(One of their recommendations was that, having established the value of acupuncture, is was now time to find the best protocol to maximise its effect. This of course meant that sham acupuncture need no longer be bothered with as a control!)
BillyJoe
All good points BSM. And yes, I am aware of the acuteness of the dilemma here. We either rule out experimental evidence and risk being accused of being closed minded, or accept trial evidence and risk shoddy or fraudulent work gaining more credence than it deserves.
One way out is for organisations like Cochrane to explicitly take on-board prior probability. Their reviews of things like homeopathy make it sound like we are only a few good trials away from the acceptance of quack remedies. The Integratives can then argue that we should not wait for expensive trials that may no be forthcoming when patients could well be benefiting. After all, the difficulty is that much of the evidence for homeopathy is not (in degree) too far removed from some conventional treatments. The difference is that we have lots of other reasons to suppose that real medicine might work.
Decision makers in health care do (or ought) to look at reviews from bodies like Cochrane. They have a duty to ensure that their scrupulously fair practices in themselves do not lead to misleading pictures. Without reviewing prior plausibility, I think that is hard.
The way forward it strikes me is for this cream to be given a full clinical trial.
If it is shown to provide benefits in the full large scale clinical trial that will be to the good we will be shown to be open minded and can absorb it into mainstream medicine. We can then show homeopathy that there are pathways by which individual treatments can gain mainstream acceptance.
If it is shown to have no benefits then we will have been shown to be correct in our view that homeopaaathy is woo.
If the full scale trial shows it to have no benefits, then it will suddenly turn out not to have been a proper test of homoeopathy.
Homeopathy can be viewed as being driven by Case Based Reasoning.
paulthomas – well, the efficacy of the topical cream should stand on its own – its success or failure is independent of the efficacy of homeopathy as this cream contains real ingredients and is more akin to a herbal ointment that a homeopathic treatment – as routinely practiced by homeopaths. The damage that this report could do is to suggest homeopathy works on the basis of something that is quite unlike what homeopathy means to most homeopaths.
LCN you are of course quite correct, which is why the efficacy of this treatment, which contains yer actual active ingredients, must be separated from homeopathy. I still think a full trial is the best way of doing this.
“Three of the eight included studies were of high quality and low risk of bias (Jacobs 2005; Pommier 2004; Thompson 2005) but only one of these studies was positive (Pommier 2004)”
And this is ‘homeopathy’ ‘working’? It’s a shame that the Cochrane people haven’t taken some important lessons from the much (unjustly) maligned Shang et al:
“smaller trials and those of lower quality showed more beneficial treatment effects than larger trials and those of higher quality […] most systematic reviews and meta-analyses are based on relatively few trials. Simulation studies have shown that detection of bias is difficult when meta-analyses are based on a small number of trials”
They seem to have been left here trying to pick the statistical bones out of solitary small trials, at least in some of the cases.
It is true that much money is spent for health information, but it is also quite true that so far no will find the cure for terrible diseases and quickly became generalized in our body, it calls on the authorities to better distribution of this money because it is spending so far in vain, I have friends who suffer from cancer, HIV, Alzheimer’s, and so far we can not find any solution to the disease, only the medicines in vicodin to control their pain, but until you take the same? actually there will be some day, the cure? Please have to be sensible and remember that nobody is free from disease and therefore it is important for everyone.
Not sure what your point is James. Not sure you know either.
I’ve also expressed my astonishment at this study here and I’m going to write to the authors of the review to see if they can explain why they included Pommier et. al. 2004, the Calendula cream study.
The study itself actually seems pretty decent, it was fairly big (254 patients), randomized, double-blind (as far as possible), etc. But the Calendula cream wasn’t described as being homeopathic – the word “homeopathy” appears nowhere in the Pommier et. al. paper!
Either Pommier et. al. mislead their readers by failing to disclose that their treatment was in fact homeopathic (very unlikely), or the Cochrane reviewers mislead their readers. I hope to find out which it is…
“The convention is to accept we have a positive result when the chances of it being a false positive (i.e. just a chance result that looks like a positive result) is 1 in 20.”
According to figures produced by the British Homeopathic Association and Faculty of Homeopathy they found 17 new trials in the years 2006-2008, one of which was positive. That’s not far off.
LCN – Just read your second postscript.
So that would have been a cream then !
Of course in theory it should have virtually every HY remedy in the world already in it as it is bound to have the odd molecule of all HY cures somewhere. In fact the hard thing would be to ensure that only a SPECIFIC given HY remedy was present and not all the others. Or am I confusing myself ? Or are they more confused than me ?
I have not read the follow on link but I am betting heavily with myself that this was probably some medicated or steroidal cream in the first place (with added oil of quack). I rest my case from the first posting above.
LCN – Just read your second postscript.
So that would have been a cream then !
Of course in theory it should have virtually every HY remedy in the world already in it as it is bound to have the odd molecule of all HY cures somewhere. In fact the hard thing would be to ensure that only a SPECIFIC given HY remedy was present and not all the others. Or am I confusing myself ?
I have not read the follow on link but I am betting heavily with myself that this was probably some medicated or steroidal cream in the first place. I rest my case from the first posting above.
Neuroskeptic: The only homeopathic allusion is in the manufacturer of the cream: “Pommade au Calendula par Digestion; calendula extract ointment; Boiron Ltd, Levallois-Perret, France”. Boiron, of course, being France’s premier BigQuacka homeopathy outfit.
According to their US website, where Calendula products are hawked as homeopathic, the Calendula content is 4% (or perhaps 1X of a 4% preparation, 0.4%?). Other Calendula products are in the 7% to 10% range (or 0.7% to 1%??). Either way, there’s some stuff in the stuff
Ooh, good use of Google there. I tried and failed to find Boiron’s English website.
Their “Calendula lotion” is described as being “Calendula officinalis MT HPUS 10%”. I presume that “MT” stands for “mother tincture”, meaning undiluted alcoholic solution in homeopathic terminology. Presumably that would make the 1x a 1/10 dilution of that… so still pretty strong.
Homeopaths rarely seem to concern themselves with the concentration of the MT – all that seems to matter is the number of times the preparation has been diluted and shaken.
The more I look at this Cochrane report, the more I think it stinks.
It was essentially written by the Royal London Homeopathic Hospitital – an institution under considerable threat and therefore, with a huge vested interest in generating good publicity. SO, pick a cochrane review that is bound to work – “homeopathy does not cause side effects in cancer patients” – no one would dispute this – and let the inevitable positive headlines ensue. Throw in some distinctly un-homeopathic like treatments and you can show some side benefits of ‘homeopathy’ maybe working.
But the point is, that these ‘working’ treatments are most unlike what goes on in the RLHH. Headlines like the BBC “Homeopathy ‘eases cancer therapy'” must be a godsend because it makes it look like the system of homeopathy actually works when this reviews quite clearly shows that there is no evidence for this.
Is this deliberate manipulation? I would have to conclude that Cochrane have been subverted here for vested interests.
We’ve covered this on the Cancer Research UK blog:
http://scienceblog.cancerresearchuk.org/2009/04/17/no-convincing-evidence-that-homeopathy-can-help-with-cancer-treatment-side-effects/
It’s certainly disappointing that the headlines have focused on the positive spin rather than the actual story, which is “it doesn’ work”.
Assuming the formula below (pasted from Neuroskeptics blog) is the one used in the trial.
Consider the most concentrated ‘actives’ Calendula and Hamamelis
in D1 potency. Note that only 1.326g of the D1 potency was added.
eg for 100g we in effect have only 0.1326g of both Calendula and Hamamelis plant added.
At 0.132% this is nearer to the D3 potency or more exactly around 1 part per 750.
Note also that the original Mother tincture is an Ethanol/Water extract that is unlikely to extract all ‘active ingredients’ in the plant.
Would we therefore expect clinical effects from Traumeel S at these concentrations bearing in mind that any Polysaccharides, glycosides ect would be expected in concentrations in the Traumeel S product in parts per million or 10s of parts per million?
At these levels the ‘actives’ are close to possible levels of impurities in the Ph Eur water used as an excipient.
100 g contains: Arnica montana, radix D3, 4.416g; Calendula officinalis D1, Hamamelis virginiana D1, 1.326g each; Bellis perennis D1, Hypericum perforatum D3, 0.295g; Millefolium D1, 0.265 g; Symphytum officinale D4, 0.295g; Aconitum napellus D3, Belladonna D3, 0.147g each; Hepar sulphuris calcareum D6, 0.074g; Chamomilla D1, Echinacea angustifolia D1, Echinacea purpurea D1, 0.442g each.
Dayoff: Certainly. When I make a cup of tea, I prepare a weak solution of about 2.5 g of tea in (roughly) 250 g of water, or 1% by weight. And I only give it a couple of mintutes to dissolve. But the result is still tea, not water. If I did the same thing with ethanol/water, I’d get even more of the active ingredients.
LCN – I have complained to Beeb about their presentation of this paper. I would also like to complain to Cochrane. Is there a way to do that as an insignificant member of public? Shouldn’t this have been peer reviewed and homeopathic spin removed before publication?
“It was essentially written by the Royal London Homeopathic Hospital – an institution under considerable threat and therefore, with a huge vested interest in generating good publicity.”
It can perhaps also be looked at in the context of the concept of “integrated medicine”. Responsible homoeopaths do not claim that homoeopathy can cure cancer or AIDS, but they like to suggest that it can be used as a palliative (of course, this contradicts their claim that homoeopathy addresses the causes of disease while medicine only treats symptoms) or to relieve the side-effects of chemotherapy (thus neatly raising the fact that effective treatments have side-effects). This enables homoeopaths to claim that homoeopathy “can help” with cancer/”be used to treat” cancer patients etc.
I emailed the authors of the review and they responded saying why they included the Pommier paper: see my post here.In a nutshell, they say that since the ointment was prepared in accordance with the German Homeopathic Pharmacopoeia, they considered it to be homeopathic.
i dotn belive it. my one of friend suffering from cancer got fully recovered only by homeopathy medicnes.