Homeopathy does work

[Garrette]

Interesting that this:

Originally posted by Dirk Gently:

David Taylor Reilly's groundbreaking paper describing his double blind randomised crossover trial demonstrating the physiological effect of potentised pollen on human hay fever sufferers. This is a solid and extremely well constructed trial showing the effect of remedies on the body. The mechanisms are still unclear, but the effect is very shown here to be very real.

Is taken word for word from here .

A quick google reveals that the Reilly paper is oft-mentioned but never linked.

 

[Dirk Gently]

sorry for the missing link!

 

[Ashles]

Is homeopathy a placebo response? Controlled trial of homeopathic potency, with pollen in hay fever as a model. Reilly DT et al., Lancet 1986, Oct18;2(8512):881-6

David Taylor Reilly's groundbreaking paper describing his double blind randomised crossover trial demonstrating the physiological effect of potentised pollen on human hay fever sufferers. This is a solid and extremely well constructed trial showing the effect of remedies on the body. The mechanisms are still unclear, but the effect is shown here to be very real.

Not the first time this has been discussed:
Discussed here

 

[Dirk Gently]

I'm referring to the fairly well documented phenomenon of animals responding to very subtle cures from owners.

Yes, it's called homeopathy.

 

[Rolfe]

That's too anecdotal for me.

OK, so long as you undertake never to support homoeopathy with equally one-off anecdotes.

So, the literature.

Taylor et al. (1989), Veterinary Record. Not available online, however a summary states

Taylor et al. (1989) described seven calves treated with a commercially purchased homoeopathic nosode for Dictyocaulus viviparus and seven with a sham solution, both groups then being challenged with virulent larvae. No antibody could be demonstrated in either group, mortality in both groups was high, and there was no difference in the number or morphology of the worms recovered from the groups post mortem. The conclusion was that "There were no discernable differences between the treated and the control groups in their manifestations of resistance to D. viviparus or their clinical responses to the diseases produced."

de Verdier Klingenberg et al., 2003, Acta Vet Scand

No clinically or statistically significant difference between the 2 groups was demonstrated. Calves treated with Podophyllum had an average of 3.1 days of diarrhoea compared with 2.9 days for the placebo group. Depression, inappetence and fever were presented equally in the 2 groups. These results support the widely held opinion that scientific proof for the efficacy of veterinary homeopathy is lacking. In the European Union this implies a considerable risk for animal welfare, since in some countries priority is given to homeopathic treatments in organic farming.

Scott et al., Canadian Veterinary Journal

A commercial homeopathic remedy and a placebo were administered orally as individual agents to 18 dogs with atopic dermatitis. The pruritus was reduced by less than 50% in only 2/18 dogs; 1 of these dogs was receiving the homeopathic remedy, the other was receiving the placebo. One dog vomited after administration of the homeopathic remedy.

Holmes et al, 2004, Veterinary Record

There were no significant differences between the SCC of the two groups on any sample day.

That's your lot. The only controlled trials of veterinary homoeopathy ever published. Resoundingly negative. Anything positive you might produce is going to be as exactly as anecdotal as the story of mine you just rejected. Deal with it.

Rolfe.

 

[Ashles]

No, I really don't.

I wasn't referring so much to actual symptom based treatment, as to when homeopathy is 'prescribed' for more behavioural-type animal problems.
Nervousness, agitation etc.

But obviously I'll trust yours and BSM's knowledge in these areas.

 

[Ashles]

Yes, it's called homeopathy.

I meant 'cues'.

 

[delphi_ote]

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=2876326&dopt=Abstract
Info on the author:
http://www.minimum.com/p7/engine/auth.asp?n=1041
His title: "homeopathic consultant, Glasgow Homoeopathic Hospital" Most certainly an unbiased scientist...

Hmm... also found these, for what they're worth.

http://bmj.bmjjournals.com/cgi/content/full/321/7259/471
http://www.liebertonline.com/doi/abs/10.1089/107555301750164208?cookieSet=1
http://www.theannals.com/cgi/content/full/39/4/617

 

[Rolfe]

Not the first time this has been discussed:
Discussed here

Oh, not fair to anyone to reference a Xanta thread!

Reilly is one of the most fervent homoeopathy believers on the planet. However, what he's doing isn't actually homoeopathy, it's isopathy. Does it matter? Well, yes, because every time anyone tries to repeat this sort of test and gets a null-effect, then this is dismissed because "it's not homoeopathy".

But basically his statistics are so wrong that a less generous person than I would call them dishonest. The trials are of very poor quality indeed, the initial dagnosis of some of the subjects has been seriously questioned, and the magnitude of the effect he reported, even if true, was clinically insignificant.

Just to let you know how much we've already studied the literature (before you start posting all about Kleijnen, Boissel and Linde as if you think we've never read that stuff), here's a link to a post of mine further on in that thread where I include a summary of all the major scientific studies on homoeopathy, with relevant quotes.

So perhaps you'd like to start from where we are, rather than from where you are?

Rolfe.

 

[Dirk Gently]

Quadruple – Blind (Editorial), The Lancet, April 22, 1989, p. 914.4

Can blind discussion remove bias from the reader? Take a trial in which 149 general practitioners entered 487 patients with an influenza-like syndrome into a randomized double-blind comparison of active treatment and matching placebo, both given sublingually. The first dose was supervised, the other four doses were taken on the following mornings and evenings. 478 of the entered patients (98.2%) met the admission criteria (5 out of 242 patients in the active treatment group and 4 out of 245 placebo patients were ineligible). At admission the groups were similar in age and proportion with severe illness. The patients recorded their rectal temperature morning and evening and whether they still had any or all of five cardinal symptoms within forty-eight hours of the start of treatment. The recovery rates were 39/228 (17.1%) in the active treatment group and 24/234 (10.3%) in the placebo group (P=0.03,X2). The relative risk of recovery was 1.67 (95% confidence interval [CI] 1.1-2.7). The difference in the proportion of patients who recovered was 6.8% (95% CI 0.6-13.0%). Logistic regression showed that several potential confounders did not substantially alter the effect of active treatment (odds ratio 1.9, 95% CI 1.1-3.4; P=0.02). Age and severity at admission were significantly associated with recovery: younger patients and those with mild or moderate illness recovered better, as might be expected. All the patients were asked about the effectiveness of their therapy, and more expressed favourable judgments about the active treatment (61% vs 49%, P=0.02). Use of other symptom-relieving drugs for pain, fever, cough, or coryza and use of antibiotics were not confounders; in fact, more patients in the placebo group used compounds to relieve pain or fever. Can the trial be criticized more than the authors do already? There might have been imbalances between the general practitioners in their recruitment of patients: every participating doctor should have entered 4-6 patients, to give a total of at least 596 cases. Also, data on 16 eligible patients were not analyzed for efficacy. There were only four unsupervised doses, but compliance was not reported. Finally side-effects in both groups were not recorded or reported. The authors are restrained in their discussion "The effect was modest ... but nevertheless is of interest". A 7% difference in efficacy as defined would be a respectable proportion in most drug trials. Now let the code be broken–the active treatment was a homeopathic preparation.

Ferley, J.P., A Controlled Evaluation of Homeopathic Preparation in the Treatment of Influenza-like Syndromes, British Journal of Clinical Pharmacology, 1989, 27, pp. 329-335.

Day, C.E.I., Control of Stillbirths in Pigs Using Homeopathy, International Journal for Veterinary Homeopathy, Vol. 1, No. 2, October 1986, pp. 26-28.

Fisher, P., Effect of Homeopathic Treatment on Fibrositis (Primary Fibromyalgia), British Medical Journal, 1989, 229, pp. 365-6.

Kleijnen, J., Clinical Trials of Homeopathy, British Medical Journal, 1991, 302, 216-23.

Gibson, R.G., Homeopathic Therapy in Rheumatoid Arthritis: Evaluation by Double-Blind Clinical Therapeutic Trial, British Journal of Clinical Pharmacology, 1980, 9, pp. 453-459.

Jacobs, J. et al, Treatment of Acute Childhood Diarrhea with Homeopathic Medicine: A Randomized Clinical Trial in Nicaragua, Pediatrics, Vol. 93, No. 5, May 1994, pp. 719-725.

Reilly, D., et. al, Is Evidence for Homeopathy Reproducible?, The Lancet, 1994; 344: pp. 1601-06


http://www.homeopathicmedicalclinic.com/homeopathy/book_english/Apendix.htm

 

[Garrette]

Let's see:

1. Subjects were allowed to self-administer undisclosed over-the-counter medications, too.

2. It says this:

Age and severity at admission were significantly associated with recovery: younger patients and those with mild or moderate illness recovered better, as might be expected.

But doesn't discuss proportions of age and severity for the control group vis-a-vis the homeopathic group.

---

It appears, Rolfe, that dear Mr. Gently intends to simply continue pasting what he considers supporting links while declining to discuss them and declining to read what you suggested he read.

 

[Rolfe]

Just for the heck of it, I'll re-post my summary of the main journal articles here, minus the snipping at Xanta.

Maybe we need to look at the actual clinical evidence regarding homoeopathy. After all, if the remedies don't have any effect on the likelihood of patients recovering from disease, there's not much point speculating about mechanisms of action.

There have been quite a few meta-analyses and literature reviews over the years. The first one was totally negative.

Hill, C. and F. Doyon (1990) Review of Randomized Trials of Homoeopathy. Review D'Epidemilogie et de Sante Publique 38: 139-147.

In our opinion, the results do not provide acceptable evidence that homoeopathic treatments are effective.

The next believed that there might be something there, but couldn't really say because of the poor quality of most of the papers they'd reviewed. Remember, this is an entire system of medicine, but they couldn't be sure whether there was any effect there at all.

Kleijnen, J., P. Knipschild and G. ter Riet (1991) Clinical Trials of Homoeopathy. British Medical Journal 302: 316-23.

At the moment the evidence of clinical trials is positive but not sufficient to draw definitive conclusions because most trials are of low methodological quality and because of the unknown role of publication bias. This indicates that there is a legitimate case for further evaluation of homoeopathy, but only by means of well performed trials.

The homoeopathic community has three papers it regularly trots out in support of its case. This Kleijnen paper is the first, and the interesting point about that is that they usually claim that the authors state that the better the quality of the study design, the higher the chance of a positive result. In fact the paper says nothing of the sort. Homoeopathy proponents are simply copying and pasting this page here, including the lies. (Acknowledgement to Geni for finding this.)

The next paper to appear is never mentioned by the homoeopaths, and perhaps one can see why.

PRESCRIRE INTERNATIONAL (unattributed articles) (1996) Homeopathy update. Prescrire International 5:21.

Despite the large number of comparative trials carried out to date there is no evidence that homeopathy is any more effective than placebo therapy given in identical conditions.

The next paper is the one the homoeopaths like to boast about most. It's certainly the most positive mainstream publication they've ever had in their favour. But even this can't say for sure that there's any effect at all!

Linde, Klaus, Nicola Clausius, Gilbert Ramirez, Dieter Melchart, Florian Eitel et al. (1997) Are the Clinical Effects of Homoeopathy Placebo Effects? A Meta-Analysis of Placebo-Controlled Trials. Lancet 350: 834-43.

The results of our meta-analysis are not compatible with the hypothesis that the clinical effects of homeopathy are completely due to placebo. However, we found insufficient evidence from these studies that homeopathy is clearly efficacious for any single clinical condition. Further research on homeopathy is warranted provided it is rigorous and systematic.

However, several people found the first part of that statement a bit surprising, and had a closer look at the data, with the Bandolier article presenting a detailed analysis. The main point at issue was whether indeed there was no decline in the probability of a positive result with increasing quality of study design. Finally, goaded by the volume of critical comment, the original authors published what virtually amounted to a retraction.

Linde et al. (1999) Impact of Study Quality on Outcome in Placebo-Controlled Trials of Homeopathy. J. Clin. Epidemiol. 52: 7, 631-636.

We conclude that in the study set investigated, there was clear evidence that studies with better methodological quality tended to yield less positive results.

It's striking that when homoeopaths boast about the earlier Linde paper, they never get round to mentioning this one!

The third paper they present refers to a 1996 report to the European Commission on homoeopathy. This is usually represented as positive, however the 1996 document is not accessible online so I haven't been able to verify the quotes. They may be no more accurate than the representation of the Kleijnen study, for all I know. What is available is the peer-reviewed version which was published in 2000.

Cucherat M., Haugh, M. C., Gooch, M. & Boissel, J. P. (2000) Evidence of clinical efficacy of homeopathy. A meta-analysis of clinical trials. Eur. J. Clin. Pharmacol. 56(1): 27-33.

There is some evidence that homeopathic treatments are more effective than placebo; however, the strength of this evidence is low because of the low methodological quality of the trials. Studies of high methodological quality were more likely to be negative than the lower quality studies. Further high quality studies are needed to confirm these results.

A couple of years later another meta-analysis specifically collated the findings of these earlier papers, with paticular reference to the plethora of re-analyses of Linde's original data. Although the author of this review trained as a homoeopath, his opinion is not respected by the homoeopathic community!

ERNST, E. (2002) A systematic review of systematic reviews of homeopathy. Br. J. Clin. Pharmacol. 54, 577-582.

Collectively [the reviews] failed to provide strong evidence in favour of homeopathy. In particular, there was no condition which responds convincingly better to homeopathic treatment than to placebo or other control interventions. Similarly, there was no homeopathic remedy that was demonstrated to yield clinical effects that are convincingly different from placebo.

A sixth meta-analysis was published more recently.

ALMEIDA, R. M. (2003) A critical review of the possible benefits associated with homeopathic medicine. Rev. Hosp. Clin. Fac. Med. Sao Paulo, 58, 324-331.

This is very up-to-date, very penetrating, and the homoeopaths don't really seem to want to add it to their favourites list either.

As a result of the recent scientific research on homeopathy, it can be concluded that ample evidence exists to show that the homeopathic therapy is not scientifically justifiable.

And after I posted this the second time, Badly Shaved Monkey kindly pointed out an omission.

O'Meara, S., Wilson, P., Bridle, C., Kleijnen, J. & Wright, K. (2002) Homeopathy. Effective Healthcare 7: 3, 1-12.

This is one of a series of reviews published by staff at the NHS Centre for Reviews and Dissemination, University of York. Note that the authors include at least one known homoeopathy proponent.

There are currently insufficient data either to recommend homeopathy as a treatment for any specific condition, or to warrant significant changes in the provision of homeopathy. .... Any future research evaluating homeopathy should address the methodological inadequacies of the existing evidence base

Or, to put it more simply, the trials the homoeopaths like to boast about are so full of glaring holes that their assertions cannot be supported.

It's worth repeating that these are reviews not of a single drug, or even a single class of drugs or treatment approach. Homoeopathy is represented as a complete system of medicine, and the best it can manage in the entire canon of scientific literature is the opinion (much disputed) that it's possible there might be something there, but nobody can say for sure, and better quality research is needed.

Rolfe.

 

[Rolfe]

Day, C.E.I., Control of Stillbirths in Pigs Using Homeopathy, International Journal for Veterinary Homeopathy, Vol. 1, No. 2, October 1986, pp. 26-28.

Sorry, that's fan-mail in a fanzine. In fact Mr. Day published his preliminary trial for this study in the Veterinary Record, as a short communication, which is not required to meet particularly stringent standards of methodology. It was, by the way, completely unblinded, and at least half of the extremely short text was pure speculation.

There's a very good reason the "definitive" publication is in a fanzine. The Veterinary Record wouldn't publish a final piece of work so shoddily constructed, and claiming effects from such slim "evidence".

As for the rest, I think we've kind of covered most of it....

Rolfe.

PS. For the uninitiated, I should perhaps mention that not only is Mr. C. E. I. Day one of the vets we've referred to who knows how much profit there is in magic sugar pills, he's actually the very person being spoken of in the conversation I relayed in the other thread, where a very see-no-evil hear-no-evil colleague who knows him personally simply extended her hand and scratched her palm when asked whether she thought he was sincere or simply in it for the money.

PPS. If you followed the link BSM gave above to an earlier post of his about blatant dishonesty he had personally witnessed in veterinary homoeopaths, you may rightly conclude that it's no coincidence that the initials of the homoeopath in the first anecdote are "CD".

 

[Rolfe]

Well that could account for some, for sure, but they don't account for every positive result.

Horses, dogs and other animals don't read skeptic forums, but for some reason, they still respond to homeopathic treatments better than placebo. Try it on your own pets next time they ail for anything.

Getting back to the basics of your argument here Dirk. Your response to a suggestion that wishful thinking and coincidental recovery might explain the alleged "effects" of homoeopathy was that (to paraphrase), "it works on animals".

You have also rejected a story I told to illustrate a point I was making (to Ashles actually, not to you) as "too anecdotal for me".

So, your main argument so far to refute the allegation that homoeopathy is entirely subjective, is that it works on animals. And you don't like anecdotal evidence.

I've just posted the details of the 4 (four) existing properly-conducted blind trials of homoeopathy on animals, all resoundingly negative. Your reference to Chris Day and his aborting pigs is irrelevant because it was rejected from a quality journal for bad design, it's completely unblinded, and it's only published in an almost home-made homoeopathy fanzine.

So, where is you evidence to support your seminal assertion that

Horses, dogs and other animals don't read skeptic forums, but for some reason, they still respond to homeopathic treatments better than placebo.

Looking for something better than anecdote (and better than Chris Day and Mark Elliott too, please).

Rolfe.

 

[Rolfe]

Taylor et al. (1989), Veterinary Record.
de Verdier Klingenberg et al., 2003, Acta Vet Scand
Scott et al., Canadian Veterinary Journal
Holmes et al, 2004, Veterinary Record
That's your lot.

Sorry, I checked this too late to edit it. The Scott et al. paper is 2002, while the Holmes et al. paper is actually 2005 - just last month actually.

Rolfe.

PS. I should also have mentioned that the Taylor et al. study wasn't actually blinded because of the husbandry of the calves. However, it doesn't make any real difference there because the conclusions are all based on precise measurements and pathology reports, nothing in the least subjective about it.

 

[Mojo]

Sorry, I checked this too late to edit it. The Scott et al. paper is also 2003, while the Holmes et al. paper is actually 2005 - just last month actually.

This highlights something I've noticed about the studies that tend to get cited as supporting homeopathy, illustrated by Dirk's posts in this thread. Nearly all of the references Dirk has cited are from the 1980s, with even the most recent being over ten years old. In view of the fact that this area is still somewhat contentious (to say the least...) , shouldn't the homeopathic community have been falling over themselves to repeat these sort of results?

 

[Rolfe]

This highlights something I've noticed about the studies that tend to get cited as supporting homeopathy, illustrated by Dirk's posts in this thread. Nearly all of the references Dirk has cited are from the 1980s, with even the most recent being over ten years old. In view of the fact that this area is still somewhat contentious (to say the least...) , shouldn't the homeopathic community have been falling over themselves to repeat these sort of results?

Oops, sorry Mojo, I was still wrong. The Scott et al. paper is 2002, not 2003 at all. The truth was always apparent from the link to the abstract though! (For some reason the necessity of changing the de Verdier citation from 2001, its original Swedish-language publication which is now a broken link, to the refereed paper which appeared two years later, threw me right off.)

Quadruple – Blind (Editorial), The Lancet, April 22, 1989, p. 914.4

Now that editorial is quite famous, indeed very famous for its "lets stir it up!" tone. (Note, editorials are not peer-reviewed!) It's not easy to see whether there really are flaws in that study from the information given, but as Donks has pointed out, the date is a dead giveaway. Sixteen (16) years ago.

Now we all know that statistical flukes turn up from time to time. And the selection for statistical flukes is called publication bias, or data mining. But even well-intentioned single studies can turn up a statistical fluke. This is why repeatability is such an important point of any scientific discovery. David Reilly says that he can make a clinically tiny but statistically significant difference to patients with hay fever. Pons and Fleischmann say they can make deuterium atoms fuse at room temperature. Madeleine Ennis says she can tell 30C histamine from the stock solvent by the way it degranulates basophils. But then other people try it and say they can't.

OK, it's not quite so simple, certainly in the latter two cases we have some people who say they can replicate it, arguing with others who say they can't. But in the end, the argument progresses. Either more and more people get the hang of it and realise it's all on the level, or the proponents become isolated and discredited. Guess what, Reilly has been in the outer darkness for years, Pons and Fleischmann went to ground with red faces in about 1991, and Madeleine Ennis is looking as if she's in the same boat.

So, what about this 16-year-old blind trial? If there was a real effect there, it would have been proved beyond doubt ages ago.

This entire field of so-called medicine, claiming miracle cures by the truckload, and nobody has yet been able to establish, clearly and repeatedly, that there's anything there at all! For any remedy, for any disease or symptom pattern.

Pity, really.

Rolfe.

 

[Chris Haynes]

...PPS. If you followed the link BSM gave above to an earlier post of his about blatant dishonesty he had personally witnessed in veterinary homoeopaths, you may rightly conclude that it's no coincidence that the initials of the homoeopath in the first anecdote are "CD".

Earlier this year I remember you mentioning what happened when a vetinarian purchased a homepathic vet practice... he found out after checking the records that there was actual fraud going on.

 

[Rolfe]

Earlier this year I remember you mentioning what happened when a vetinarian purchased a homepathic vet practice... he found out after checking the records that there was actual fraud going on.

Yes. This wasn't the same person though (just the other guy I hoped DirkGently could do better than).

Rolfe.

 

[Rolfe]

Jacobs, J. et al, Treatment of Acute Childhood Diarrhea with Homeopathic Medicine: A Randomized Clinical Trial in Nicaragua, Pediatrics, Vol. 93, No. 5, May 1994, pp. 719-725.

Just for the hell of it, before I retire for the night.

This is another one I recognise. Now I'm doing this from memory so I may not be absolutely accurate, but the gist of it is that this study was done in Nicaragua because it wouldn't have gained ethical approval in this country (allowing sick infants to go untreated). Even so, only mild and moderate illness was included, severe cases were excluded.

It was in effect a data mining exercise. There was one statistically significant result, that the treated group had on average one less loose stool on day 4. However there was no difference between the groups on either dat 3 or day 5. Put another way, this translated to a shorted duration of the diarrhoea of no more than a few hours, on average, in the treated group. "If that's the best they can do, why bother?"

Note that the de Verdier Klingenberg paper I cited was an attempt to do the same thing but with calves. Unfortunately, in that case the treated group actually took longer to get better than the placebo group.

This all highlights a very real problem with testing homoeopathic treatments on actual patients. If there is already an established, effective treatment for a condition, then it may be considered unethical to leave patients untreated. For real medicine these days, the solution would be to give one group the established treatment and the other the new stuff. But that's not much use when you're trying to establish whether there is actually any effect there at all, rather than whether this stuff is an improvement on what is already available.

In fact, the big omission from the Jacobs et al. paper was a group treated with the usual stuff. Of course in that case they wouldn't have been arguing about one significant figure out of the data-mining, or a couple of hours difference in the ending of the diarrhoea, but a very quick clear-up with the usual stuff, compared to the poor kids on the homoeopathy lingering on sick for days.

Rolfe.