The British Homeopathic Association (BHA) is a registered charity founded in 1902. Their objectives are “to promote and develop the study and practice of homeopathy and to advance education and research in the theory and practice of homeopathy…” and their priority is “to ensure that homeopathy is available to all…” The BHA believes that “homeopathy should be fully integrated into the healthcare system and available as a treatment choice for everyone…”
This does not bode well, in my view. Specifically, it does not seem as though we can expect unbiased information from the BHA. Yet, from a charity we certainly do not expect a packet of outright lies – so, let’s have a look.
The BHA have a website (thank you Greg for reminding me of this source; I have long known about it and used it often for lectures when wanting to highlight the state of homeopathic thinking) where they provide “THE EVIDENCE FOR HOMEOPATHY“. I find the data presented there truly remarkable, so much so that I present a crucial section from it below:
START OF QUOTE
The widely accepted method of proving whether or not a medical intervention works is called a randomised controlled trial (RCT). One group of patients, the control group, receive placebo (a “dummy” pill) or standard treatment, and another group of patients receive the medicine being tested. The trial becomes double-blinded when neither the patient nor the practitioner knows which treatment the patient is getting. RCTs are often referred to as the “gold standard” of clinical research.
Up to the end of 2014, a total of 104 papers reporting good-quality placebo-controlled RCTs in homeopathy (on 61 different medical conditions) have been published in peer-reviewed journals. 41% of these RCTs have reported a balance of positive evidence, 5% a balance of negative evidence, and 54% have not been conclusively positive or negative. For full details of all these RCTs and more in-depth information on the research in general, visit the research section of the Faculty of Homeopathy’s website. Also, see 2-page evidence summary with full references.
END OF QUOTE
But is it true?
Let’s have a closer look at the percentage figures: according to the BHA
- 41% of all RCT are positive,
- 5% are negative,
- 54% are inconclusive.
These numbers are hugely important because they are being cited regularly across the globe as one of the most convincing bit of evidence to date in support of homeopathy. If they were true, many more RCT would be positive than negative. They would, in fact, constitute a strong indicator suggesting that homeopathic remedies are more than placebos.
One does not need to look far to find that these figures are clearly not correct! To disclose the ‘mistake’, we do not even need to study any of the 104 RCTs in question, we only need to straighten out the BHA’s ‘accounting error’ and ask: what on earth is an ‘inconclusive’ RCT?
A positive RCT obviously is one where homeopathy generated better outcomes than the placebo; similarly a negative RCT is one where the opposite was the case; in other words, where the placebo generated better outcomes than homeopathy. But what is an ‘inconclusive’ RCT? It turns out that, according to the BHA, it is one where there was no significant difference between the results obtained with placebo and homeopathy.
Yes, you understood correctly!
Outside homeopathy such RCTs are categorised as negative studies – they fail to show that homeopathy out-performs placebo and therefore confirm the null-hypothesis. An RCT is a test of the null-hypothesis (the experimental treatment is not better than the control) and can only confirm or reject this hypothesis. Certainly finding that the experimental treatment is not better than the control is not inconclusive bit a confirmation of the null-hypothesis. In other words it is a negative result.
So, let’s look at the little BHA – statistic again, and this time let’s do the accounting properly:
- 41% of all RCTs are positive,
- 59% are negative.
This means that, according to this very simplistic method, the majority of RCTs is negative. I say ‘very simplistic’ because, for a proper analysis of the trial evidence, we need to account, of course, for the quality of each trial. If the quality of the positive RCTs is, on average, less rigorous than that of the negative RCTs, the overall result would become yet more clearly negative. Most assessments of homeopathy that consider this essential factor do, in fact, confirm that this is the case.
Once all this has been analysed properly, we still have to account for factors like publication bias. Negative trials get often not published and therefore the overall picture gets easily distorted and generates a false-positive image. At the end of a sound evaluation along these lines, the result would fail to show that homeopathy differs from placebo.
Regardless of all these necessary and important considerations, the BHA website then tells us that the RCT method is problematic when it comes to testing homeopathy: “The RCT model of measuring efficacy of a drug poses some challenges for homeopathic research. In homeopathy, treatment is usually tailored to the individual. A homeopathic prescription is based not only on the symptoms of disease in the patient but also on a host of other factors that are particular to that patient, including lifestyle, emotional health, personality, eating habits and medical history. The “efficacy” of an individualised homeopathic intervention is thus a complex blend of the prescribed medicine together with the other facets of the in-depth consultation and integrated health advice provided by the practitioner; under these circumstances, the specific effect of the homeopathic medicine itself may be difficult to quantify with precision in RCTs.”
What are they trying to say here?
I am not sure.
Are they perhaps claiming that, even if an independent scientist disclosed their ‘accounting error’ and demonstrated that, in fact, the RCT evidence fails to support homeopathy, the BHA would still argue that homeopathy works?
I think so!
It looks to me that the BHA is engaged in the currently popular British past-time: THEY WANT THE CAKE AND EAT IT.
All this is more than a little disturbing, and I think it begs several questions:
- Is this type of behaviour in keeping with the charitable status of the BHA?
- Does it really ‘promote and develop the study and practice of homeopathy and to advance education and research’?
- Is it not rather unethical to mislead the public in such a gross and dishonest fashion?
- Is it not fraudulent to insist on false accounting?
I would be interested to get your views on this.
“How do I get to Dublin?”
“If I was you, I wouldn’t start from here.”
Homeopathy is not a philosophy or faith which is susceptible to rational analysis.
The BHA clearly state there is no plausible evidence that homeopathic remedies have any recognisable effects.
To sell, promote, or endorse their use might be fraudulent. Certainly, unethical. As is false accounting.
As I recall, Robert Mathie has claimed authoritative support for the BHA’s definition of positive, negative and inconclusive RCTs. I can’t remember the claimed source of that authority but I’ve never come across it anywhere else.
I always find their defence against scrutiny by RCT deeply irritating. Ooh, homeopathy is complex!! Maybe so, and they have big books of complicated and internally inconsistent rules, but it all boils down to a simple binary question, did the patient get better?. That can be by objective measurement or some subjective assessment, it doesn’t matter. Do more patients get more better with magic sugar than with placebo? A simple binary outcome is absolutely the locus for the RCT.
Precise, concise: “…simple binary outcome…” exposes homeopathy as a corpse claimed to be top-notch athlete by unscrupulous management.
Actually, the definition is on their website: http://www.britishhomeopathic.org/evidence/the-evidence-for-homeopathy/
Quote: “44% of the reviews concluded that the interventions studied were likely to be beneficial (positive); 7% concluded that the interventions were likely to be harmful (negative); and 49% reported that the evidence did not support either benefit or harm (non-conclusive). [El Dib RP, Atallah AN, Andriolo RB (2007). Mapping the Cochrane evidence for decision making in health care. Journal of Evaluation in Clinical Practice; 13:689–692.]”
I.o.W. 59% *are* negative. It would also be of interest to look at the p-values the positive assertion is based on. In working treatments, one has like 10% negative and 90% positive studies and these with low p-values. Homeopathic RCTs tend to be borderline positive. These numbers in itself represent a very clear hint that homeopathy is bogus.
Aside that, the BHA did apparently not (want to ?) understand what an RCT really compares. RCTs do NOT compare drugs. Strictly speaking they compare workflows, *regardless of their nature*. The workflow could encompass chosing an amerindian rain dance or writing a symphony a la Beethoven for the patient. It is the *design* of the workflow that allows a conclusion about the efficacy of a drug. That means, a workflow that encompasses the exact homeopathic procedure but omits the homeopathics in fact is very well suited to allow the conclusions about the effectiveness of homeopathics. There is no reason at all why an RCT would not be suited to test an individualized workflow, provided it is designed in a way that allows conclusion on a few aspects of the workflow.
Furthermore they do not seem to understand what homeopathy according to Hahnemann really is. For instance they falsely claim that homeopathy is also life style counseling. Strictly speaking, homeopathy is the treatment of diseases according to Hahnemann which is a treatment with drugs, nothing more. This further strengthens the position that RCTs are very well suited to test homeopathy.
Claims on life style counseling amount to poaching into fields homeopaths are not properly educated to practice.
My perception is that homeopaths’ education is primarily marketing and salesmanship
BTW, the BHA also missed one extremely important point. The classification according to El Dib et al. applies to *metastudies* and NOT original studies. Original studies are never inconclusive, metastudies can be (i.e. in the case when two equally good studies are one positive and one negative).
Oh, that’s interesting – I hadn’t appreciated that before. Reading the paper now…
Or, in pictorial format:
Of course, that doesn’t take into account the quality of those ‘positive’ trials.
I will use it for lectures, if I may.
Of course you can!
Some other trade bodies of homeopaths use slightly different numbers, but are essentially the same. If you would like something slightly different, let me know.
Alan, I know what you mean, however, the “Null Hypothesis” chart is incorrect: The 41% positive for homeopathy is the rejections of the null hypothesis; the 59% is the failures to reject the null hypothesis, which is the default hypothesis.
Yes, I see what you mean. I had used it previously in a different context where I’d given an explanation of what it meant to reject the null hypothesis, but I’ll look at trying to clarify what the chart means and update it.
There, that’s better, I think?
Thanks, Alan. Your updated version is very clear and it cannot be misinterpreted.
I only have the graphic posted at 12:03 on 11.02.17.
Has it now been updated?
Yes, that’s the updated one.
“Non-conclusive” is Alt-Science-speak for “Next time fish harder”.
I would like to thank Professor Ernst and his followers on this blog for reaching the point of annihilating the straw man model of homeopathy. All the studies of homeopathy that have been cited speak for themselves: if homeopathy cured those diseases it would be evident to see!
I have mentioned this previously: homeopathy is not based on the treatment of specific illnesses with a particular remedy. Although the homeopathic literature does state connections between a disease and a remedy, for example: Belladonna as almost ‘specific’ in scarlet fever; to my knowledge, Hahnemann did not get his band of followers together to ‘prove’ that a particular drug ‘worked’ for a particular illness.
The construction of the homeopathic repertories and materia medica is erroneous as remedies are listed in/cited for clinical ‘rubrics’ and the main homeopathic remedies are listed for virtually every ‘disease’. This is not ridiculous; it is absurd.
The infusion of non scientific thinking into homeopathy over the past 200 years has created a misunderstanding of what it is. The failure of homeopaths to establish a credible data base of evidence of its efficacy during this time has placed homeopathy in a difficult position in presenting its case to the scientific community and with the crumbling of strawman, the analysis needs to go back to the beginning of homeopathy. I am sure that the contributors to this blog have studied the original works of the founders of homeopathy- and if we can move to that domain, eventually the whole thing can get straightened out.
I know that some of you find great mirth in beating up strawman but watching you do it is comedy.
Well, it was the model you yourself presented.
Greg, I have read Hahnemann.The very founding of homeopathy is a badly conducted experiment done interpreted terribly wrong by a bad scientist. The famous cinchona bark experiment which lead Hahnemann to conclude like cures like is now well understood. Nothing, nothing whatsoever in Hahnemanns interpretation of the results is correct or comes even close to what really happened.
I recommend you to read Hahnemann’s recommendations how to develop a drug and compare them to James Linds studies on the discovery of lemon as a treatment against scurvy. The problem with Hahnemann is that he never ever checked the efficacy of his “discoveries” although RCTs where known and conducted in his time. The observational bias in Hahnemanns work is often evident in his own case reports where he observes the persistance of a disease yet claims a patient to be cured.
“Hahnemann did not get his band of followers together to ‘prove’ that a particular drug ‘worked’ for a particular illness.”
In the Organon, Hahnemann used the word ‘pruefung’, but that means ‘experiment or trial’ without any implication of ‘proof’ in the modern English sense of ‘establish validity’.
The German word for ‘to prove’ is ‘beweisen’ – and that is not the word Hahnemann used.
Hahnemann did not establish that any of his remedies, at any level of dilution, had any significant effect on any specific disease. He did not ‘prove’ anything.
His experiments and trial did establish that folks imaginations are strange places, and that the use of placebos can help some folks come to terms with their ailments – if only for a time.
Just as any TLC can.
And the methodology of RCT does not require that it does so. Nor has the discussion here required that homeopathy be trammelled into treating diagnoses defined by conventional medicine. I was very clear in the minimal requirements for an effective and fair RCT of homeopathy. Homeopaths claim to get patients better. They either do or they don’t.
It is, however, notable that when homeopaths play at science they also often seem to wish to play at conventional medicine as well by treating specific conventionally defined diseases and conditions. Time and again we see that homeopaths are not competent doctors, not competent scientists and not even competent homeopaths.
But if homeopathy really worked as claimed, its positive effects would be screamingly obvious in quite small-scale trials. They aren’t. Homeopathy doesn’t work.
That seems to nail it all pretty concisely.
Thanks guys, you make my day and I will be laughing in my dreams.
What is a ‘good homeopath’?
One who tells their patients that homeopathic remedies’ benefits are mediated by placebo effects, and gets fully informed consent on that basis.
What is the difference between a ‘doctor’ and a ‘competent doctor’?
Simon is right though, if homeopathy ‘worked’ in clinical trials, this would be measurable and noted (BBC Breaking News).
All doctors should indeed obtain fully informed consent.
Refer any who do not to the GMC (or me! I was an associate member of the GMC).
In particular, registered medical practitioners who also practice homeopathy should be reported if they do not explain the nature of their ministrations to their patients. Not to do so is unethical. Simple.
They should also point out that only about 1/3 of the population is even susceptible to the placebo effect and, if they’re luck enough to fall into that group it’ll only work 1/3rd of the time and you can’t determine if they will reliably fall into that group for any given treatment.
If you have evidence for that – that’s what patients should be told.
Very salient points. To which I shall add that, for those who do respond to a placebo treatment for their particular condition, their positive placebo reaction is very unlikely to be effective for longer than six months.
This is why the authors of New-Age self-help books, courses, and seminars, have to regularly author then sell modified versions of their previous medically-inefficacious works. Each new work is proffered as being the next step along the pathway to the eventual ‘true enlightenment’ of their desperate and/or gullible consumers.
“What is a ‘good homeopath’?”
Random Joe has a earache. Goes to doctor, who looks inside the ear and says there is an infection. Joe gets a prescription for an antibiotic and is told that the infection will go away in 2 or 3 weeks.
2 weeks later, Joe still has an earache. Goes to the same doctor, who looks in the ear again and sees the ear is not OK yet but is better than 2 weeks before.
Doctor says “HMMMMMMMMMMMM” “Interesting” and looks concerned.
Gives Joe a different set of white pills and says “take 1 of these along with the antibiotic. Double the amount of white pills if necessary. Your earaches should go away in 2 weeks approximately.”
Doctor thinks to himself “I hope this thing clears out quickly, otherwise this guy will be coming back to the clinic again and will catch something bad from the other people in the waiting room”.
Joe goes home, takes both the ATB and the sugar pills, feels fine in 1 week. Those pills from the 2nd visit were awesome.
So a good doctor is one who deliberately deceives his patient?! That’s a recipe for all kinds of abuse. Why can’t the doc equally well wave a crystal at the patient? After all, what he’s doing is assuming the patient is stupid or gullible.
Perhaps providing reassurance that sometimes things such as ear infections respond slowly and that the patient really needs to be patient (I know: that line goes back many years) is a more reasonable alternative? And doesn’t treat the patient as a fool.