After our debate on November 26, 2012, “Homeopathy: Mere Placebo or Great Medicine”, homeopath Andre Saine and colleagues asked me to answer some further questions, which I am pleased to do. I should point out that I have already addressed many of these issues in a series of columns I wrote for the Montreal Gazette and in an answer to a letter from the President of the Syndicat professionnel des homeopathes du Quebec who attacked me in a rather unprofessional fashion.
Andre also queried how much I get for consulting fees with companies. The answer is zero. My salary comes from the university. He also asked why I would favour a lawsuit against a company that markets homeopathic products given that I had stated that placebos could indeed serve as effective medicine. The lawsuit is being funded totally by a group of lawyers who initiated it. The motive behind it is simple. Homeopathic medications are claimed to have active ingredients without any proof that they contain such. Therefore they are mislabeled. I had nothing to do with the suit but I do support it because I think all products that make medical claims should be held to the same standards of safety and efficacy. In comparison with the demands that regulatory agencies make for other drugs, homeopathic products are getting a free ride.
Part of the problem is that most people do not actually know what homeopathy is. They believe it is just some sort of general term for “natural medicine.” It is of course not that at all. Consumers are often shocked to learn that homeopathy is based on the principle of non-existent molecules having an effect. For example, Boiron’s flu remedy Oscillococcinum, is made by the repeated dilution of a single duck’s liver 200 times, equivalent to taking the liver and diluting it in a sphere of water far bigger than the observable universe. I have given a more detailed description of the folly of this remedy here.
Little wonder that Boiron has had to settle a class action lawsuit in California that alleged false advertising on behalf of Oscillococcinum. The company agreed to refund consumers a total of $5 million and has agreed to pay legal expenses and to add a disclaimer that their health claims have not been evaluated by the U.S. Food and Drug Administration. It also agreed to add an explanation of how the “active” ingredients have been diluted. Hopefully there will be a similar settlement of the Canadian lawsuit.
I believe that after digesting my previous remarks on the subject and perusing my answers below, Andre’s summary comment that “the future of medicine is homeopathy” will be seen in its true light.
Before tackling the questions, let me begin with a few preliminary observations.
It was back in 1865 that French physiologist Claude Bernard opined that “It is better to know nothing than to keep in mind fixed ideas based on theories whose confirmation we constantly seek.” He didn’t have homeopathy in mind but he certainly could have.
Embracing a fixed idea and sticking to it in the face of emerging contrary findings is a real stumbling block in the evolution of any scientific discipline. In true scientific methodology, errors are uncovered, and science rights itself. But in rare cases, an entire discipline can be based on a fundamental flaw, which I believe is the case for homeopathy. Its very underpinnings that relate to the effects of non-existent molecules inducing some sort of memory effect on water are scientifically implausible in the light of knowledge we have gained since the introduction of homeopathy by Samuel Hahnemann over two hundred years ago. Yet homeopaths today still embrace Hahnemann’s archaic ideas. No other scientific discipline has shown such a lack of evolution.
The argument that there has been no need for change because Hahnemann’s brilliant breakthrough cannot be superseded smacks of juvenile thinking and a lack of familiarity with the astounding amount of biological and chemical knowledge that has been acquired over the last two centuries. But it is unlikely that the cleft between homeopaths and their critics will ever be breeched because one side takes its nourishment from science while the other is mired in faith. Science versus faith arguments are not resolvable, but clearly both have merits and detriments.
Most visits to a physician are prompted by the patient experiencing a symptom or set of symptoms for which they hope to receive effective treatment. If they are fortunate, some sort of treatment will be devised and a cure will follow. The natural assumption is that the cure was due to the treatment. And it may well be. But there are a couple of other possibilities. Many illnesses are self limiting and resolve even in the absence of any intervention. If, however, there is some sort of intervention, it usually ends up getting credit. There is yet another possibility. That’s the placebo effect. Basically this is a “cure” that is not related to the pharmacological or physical effects of whatever intervention is undertaken, but is the result of the patient’s faith in the treatment and the physician’s belief that the treatment will be effective.
Patients are keen to believe that they are receiving appropriate therapy, and in response to patients’ claims of improvement, physicians may be readily deluded into thinking that they have offered effective treatment. While placebos do not affect the progress or outcome of disease, there is no doubt they can have a powerful effect on subjective phenomena such as pain or distress. And placebos do not necessarily have to be pills, potions or physical treatments, although there is evidence that the more invasive a treatment is, the stronger the placebo effect. Injections, for example, work better than pills. But a physician’s words and attitude can also serve as effective placebos. This was demonstrated in a fascinating fashion by a British family practitioner who designed a clever study as reported in the British Medical Journal in 1987.
Dr. K.B. Thomas studied 200 patients who presented with symptoms such as cough, sore throat, abdominal pain, back pain, leg pain, headache, fatigue, earache, neck pain and chest pain in whom no definite diagnosis could be made. Half the patients were dealt with in a so-called “positive manner,” in which they were given a firm diagnosis and were told with confidence that they would be better in a few days. The other half had a “negative” consultation in which no firm assurance was given, with the doctor saying, “I cannot be certain what is the matter with you.” Furthermore, each group was divided into two, with half the patients in each group being given a medication, the other half not. In the positive group the patients were told that they would benefit from the medication, in the negative group they were told that the drug may or may not help.
The results were stunning. Two weeks after the initial visit, 64% of those receiving a positive consultation got better compared with 39% of those who received a negative consultation. Even more interestingly, whether the patients were given a medication or not made no difference. So what are we to make of this? It seems the effect of a doctor as a placebo may be more powerful than the placebo effect of a medicine! It should come as no surprise then that homeopaths enjoy success even though their medications, which do not contain any pharmacologically active agents, are totally bogus. They are usually caring people who devote a great deal of time to their patients and confidently offer them positive outcomes. That’s the formula for an effective placebo.
Here then are the questions and my answers:
You said in the debate that a well-conducted randomized controlled trial (RCT) will be convincing. What was not convincing about the ADHD or the sepsis RCT presented by Andre Saine?
Essentially all of the questions posed here are very familiar to the critics of homeopathy and have been extensively answered in the scientific literature. Indeed, the cornerstone of science is the randomized controlled trial (RCT), or, more correctly stated, the PROPERLY CONDUCTED randomized controlled trial. Unfortunately, the literature is replete with RCTs that are extensively flawed. Either the blinding is faulty, the allocation of subjects isn’t truly random, there is a lack of proper accounting for drop outs, the sample size is inadequate or there are methodological problems with the statistics. Given that statistically significant results can occur by chance alone, a single outcome, even if the study is properly conducted, does not carry much weight. By convention, researchers abide by the P=0.05 factor in scientific studies, which means that once in twenty times the result can be due to chance alone. To give a specific example, if a hundred coins are tossed in the air, upon landing one might expect a result like 45 heads-55 tails, or perhaps 48-52, or 46-54. But if this is repeated twenty times, one might very well once get 40 heads and 60 tails. If this were the only trial reported, one would get the impression that the coins were biased.
For all these reasons, more than one RCT that deals with the same question in the same fashion and arrives at the same result is required before one raises the possibility of legitimate evidence. And only when a sufficient number of RCTs show consensus in a meta analysis do we consider that the point has been made. There is yet another issue. While peer-reviewed journals are regarded as the gold standard for scientific evidence, it should be recognized that not all such journals are equal. There are some 6500 peer-reviewed journals in circulation in the world, many with very weak criteria for accepting papers. A study about homeopathy published in a homeopathic journal cannot be regarded the same way as if it were published in The Lancet, JAMA or the New England Journal of Medicine. Yes, the paper may be peer-reviewed, but the “peers” asked to referee a paper destined for publication in a homeopathic journal are not likely to be totally objective. Indeed, it might be quite difficult to induce a main-line scientist to review the article. There are also journals that are commonly referred to by homeopaths that are not peer reviewed at all. For example, papers by Rustum Roy, commonly referenced by homeopaths actually appear in Materials Research Innovations, which is Roy’s own journal.
There is no argument that there have been individual published papers alleging a positive effect for homeopathy. But almost all of these are to be found in journals that cater to “alternative” modalities. It is curious to note that in such journals there are no reports of negative results. This is quite unlike mainstream scientific publications which commonly feature negative findings. It seems somewhat odd that when homeopaths carry out a study they never encounter a failure, while this is common in conventional science.
In summary, the studies that are claimed to provide proof of homeopathy just don’t stand up to the rigours of science. They either have not been repeated, are methodologically flawed, or report results that can be due to random chance. All of these issues have been discussed in much greater detail than is possible here and the so-called positive studies have been dissected by numerous experts. I would urge interested readers to take the time and digest the following before accepting the notion that homeopathy is supported by properly carried out randomized trials. The first is by surgical oncologist Dr. David Gorski who systematically dissects some recent attempts by homeopaths to scientify their discipline. This is followed by comments from physician Dr. Ben Goldacre, the U.K’s premier science commentator and physicist and engineer A. P. Gaylard.
If homeopathy turns out to be a great medicine as homeopaths have claimed for more than 200 years what are the implications for our society as a whole and each individual?
If homeopathy turns out to be the great medicine that homeopaths have claimed, we will have to rewrite all of our chemistry, biology and physics texts and assume that basically everything we know about the workings of the world is wrong. If this turns out to be the case I will eat my hat or anything else that is offered.
Your claim from the beginning is that homeopathy is ‘unscientific’, that the pills have nothing in them and are therefore a placebo. Dr. Saine came up with study after study, all good SCIENCE, to refute your claims, but your brushed them away as if they meant nothing. If good science won’t convince you that homeopathy is scientific, what will?
I did not brush the placebo effect aside at all. It is perhaps the most important effect in medicine, conventional or alternative. As far as the studies go, Andre Saine’s interpretation of what is good science differs from mine and from that of the majority of the scientific community. I’ve already laid out the major points above and the question of the validity of studies relating to homeopathy has been in great detail in the references I provided. It must be remembered that “scientific” studies can be found to support almost any claim, but the nature of the studies has to be carefully scrutinized before jumping to conclusions.
If homeopathy is quackery as you say, can you define what is a quack and which aspects of it apply to Hahnemann?
A quack is either a fraudulent or scientifically ignorant promoter of ideas or practices that do not conform with knowledge that has been acquired through the proper practice of the scientific method. I would not call Hahnemann a quack. Medical knowledge at the time was virtually non-existent and within his empirical framework homeopathy made sense. His patients did better than those who were brutalized by conventional medicine at the time, and it is easy to see how he came to the wrong conclusion based on correct observations. Hahnemann could not have been aware of the fact that his solutions were so dilute that they didn’t even contain a single molecule of the original substance. Avogadro’s number, namely the number of molecules in a mole of a substance, was not determined accurately until Millikan’s famous oil drop experiment a hundred years after Hahnemann. The number was named after Avogadro, not by him. It is a myth propagated by homeopaths that Hahnemann was a great chemist who knew all about molecules.
An appropriate analogy here would be the theory of “phlogiston” which made sense to noted scientists like Robert Boyle and Joseph Priestly because it explained their observations about combustion. But the discovery of oxygen provided a much better explanation. Similarly, in the context of the times, before a proper understanding of molecules was developed, and before the full benefits of placebos were appreciated, Hahnemann’s ideas were not outlandish. He was without doubt a good man who did his best to help patients with methods that in all likelihood were superior to the “heroic” medicine practiced by other physicians at the time. But given what we know today about molecules, pathology, disease processes, drug mechanisms and mind-body reactions, the notion of “like cures like” and Hahnemann’s proposed inverse dose-response relationship has to be relegated to the junk heap of science. But junk can be dangerous. See following:
When evaluating published RCTs, what criteria are you using to know if the researchers are actually testing genuine homeopathy?
I’m not sure what this question really means. What is genuine homeopathy? The RCTs that are examined in meta analyses are ones that are designed by homeopaths so one would assume they are testing “genuine homeopathy.” Basically such studies come down to testing one placebo against another. I know that homeopaths often claim that their practice is so individualized that it is not amenable to randomized trials, but that is just a convenient “out.” I would be interested in hearing what would constitute a trial of “genuine homeopathy.”
You have high regards for the Shang meta-analysis. Can you comment on what Andre Saine said about the 6-8 studies being “non-homeopathic” studies.
It is not only I who has a high regard for this major meta analysis that was published in The Lancet, but everyone else who is well versed in scientific methodology. It was predictable that homeopaths and particularly manufacturers of homeopathic medications like Boiron Laboratories would attempt to skewer the study. They did try that with the following claim:
“In August 2005, the weekly journal The Lancet published a new study on the effectiveness of homeopathy. In its editorial the journal drew some surprisingly controversial and unfavourable conclusions on homeopathy. It concerns an analysis which like the 3 previous meta-analyses concludes that homeopathic medicine is effective. However, to arrive at the opposite conclusion, the authors implicitly removed series of trials afterwards, retaining just 14 (8 on homeopathy) of the 220 initial trials (110 of which were on homeopathic medicine).”
This criticism is invalid. Shang et al examined all the trials and relied only on eight which were of the highest quality. The rest just did not meet the scientific criteria. They certainly did not “implicitly remove” a series of trials afterwards. The surviving 8 homeopathic trials were simply the ones that made the cut. Shang’s conclusion was that there is “weak evidence for a specific effect of homoeopathic remedies, but strong evidence for specific effects of conventional interventions. This finding is compatible with the notion that the clinical effects of homoeopathy are placebo effects.”
It is noteworthy that Klaus Linde, the author of a previous meta-analysis that has been used by homeopaths to support their arguments sent the following letter to The Lancet: “We congratulate Aijing Shang and colleagues on their meta-analysis examining the clinical effects of homoeopathy. Their methods largely reproduce those of our meta-analysis on the same topic published in The Lancet 8 years ago. We agree that homoeopathy is highly implausible and that the evidence from placebo-controlled trials is not robust. Our 1997 meta-analysis has unfortunately been misused by homoeopaths as evidence that their therapy is proven.”
Andre Saine just reiterates the criticisms that have been aimed at the meta analysis apparently without being aware that these have been dealt with extensively by many, including Shang himself. It is fruitful to read through Shang’s reply which I’ve reproduced here:
Are the clinical effects of homoeopathy placebo effects? — Authors’ reply:
Homoeopathy dates back to the late 1700s when Samuel Hahnemann first formulated the principles of similars and potentisation. At that time, homoeopathy might well have been superior to conventional medicine, considering that bloodletting was the standard treatment for many disorders. Contrary to what Harald Walach and colleagues suggest, homoeopathy and allopathy thus both started from a poor evidence base. Our study showed that more than 200 years later, based on more than 200 placebo-controlled trials, it has become clear that the clinical effects of homoeopathy are compatible with the placebo hypothesis and probably due to the non-specific effects of complementary and alternative medicine discussed by Walach and colleagues. By contrast, with identical methods, we found that the benefits of conventional medicine are unlikely to be explained by unspecific effects.
We followed the approach of Klaus Linde and colleagues1 in assuming that, if the clinical effects of homoeopathy are placebo effects, the positive results seen in placebo-controlled trials of homoeopathy must be explained by bias in the conduct and reporting of trials. When these biases were taken into account in the analysis, little evidence remained for a specific effect of homoeopathy. Linde and Jonas now seem to discover major limitations with the approach they pioneered: they argue that, if homoeopathy works for some but not all conditions, our study could produce a false-negative result. Their argument is not supported by the evidence. The homoeopathy trials analysed in our study, including the large trials, and large trials of higher quality, covered diverse conditions ranging from childhood diarrhoea, warts, the common cold, to chronic headache. There was little evidence that treatment effects varied according to clinical topic (p=0·66). An important strength of our study is the inclusion of a carefully matched sample of conventional medicine trials. If the fact that a therapeutic system may work in some but not other conditions leads to false-negative results, why did this not happen in the case of conventional medicine?
We agree that the larger trials of higher methodological quality (references 46, 55, 71, 80, 84, 94, 96, 97 in web appendix 1 and 23, 25, 45, 53, 66, 72 in web appendix 2) should have been identified, and are grateful for the opportunity to rectify this oversight. Our study was not a standard meta-analysis of Cochrane reviews, which are typically based on fewer than 10 trials,2 but a large comparative study of 110trials of placebo-controlled trials of homoeopathy and 110 carefully matched trials of conventional medicine. In this situation it is difficult to report on every variable that may be of interest. This is also illustrated by Linde and colleagues’ 1997 article,1 which included 89 homoeopathy trials. Linde and colleagues failed to identify the trials that were adequately concealed, the trials that reported predefined outcome measures, or the five trials that they included in a “worst case” scenario (high methodological quality, indexed in MEDLINE, higher dilution).
Contrary to the claims of Peter Fisher and colleagues, we clearly stated the matching criteria and made all the references available in web appendices and provided additional information on outcomes in a webtable. A list of excluded studies, with reasons for exclusions, is now also available from http://www.ispm.ch.
We strongly reject Fisher and colleagues’ notion that our conclusions were based on “eight anonymous trials”. Following the same logic, Linde and colleagues’ study,1 which they and others have repeatedly quoted as proving the efficacy of homoeopathy, would be based on the five “anonymous trials” included in the worst case scenario. We estimated treatment effects in trials as large as the largest trials identified, based on an analysis of 110 trials. As for the analysis restricted to eight large trials of higher quality, we found no convincing evidence that homoeopathy was superior to placebo.
Neither of the two studies mentioned by Fisher and colleagues were regarded as large and of high quality. The influenza trial did not meet our prespecified quality criteria and the asthma trial was available as an abstract only and excluded. In response to Walach and colleagues, two of the larger homoeopathy trials of higher methodological quality assessed individualised classic homoeopathy, including the 1997 trial by Walach and colleagues (reference 96 in web appendix 1).
In response to Flávio Dantas and to Walach and colleagues, the pattern seen for trials of classic, individualised homoeopathy was closely similar to that seen for all trials of homoeopathy (figure). Our study affirms the poor reporting of clinical trials highlighted previously.3 The assessment of their methodological quality is therefore compromised, but it is clear that large, multicentre trials tend to be of higher quality than small trials.4 Dantas argues that homoeopathic laboratories “do not spend nearly as much money on marketing and product promotion as pharmaceutical companies”. Interestingly, we have been pursued by public relations firms paid by the manufacturers after our publications on rofecoxib (Vioxx)5 and homoeopathy. We agree with Dantas that we need to compare homoeopathy and allopathy: this was the aim of our study. We also need to be prepared to accept the results of well designed studies, even if they challenge our own fervently held beliefs.
1 Linde K, Clausius N, Ramirez G, et al. Are the clinical effects of homoeopathy placebo effects? A meta-analysis of placebo-controlled trials. Lancet 1997; 350: 834-843. Summary | Full Text | PDF(136KB) | CrossRef | PubMed
2 Sterne JAC, Gavaghan DJ, Egger M. Publication and related bias in meta-analysis: power of statistical tests and prevalence in the literature. J Clin Epidemiol 2000; 53: 1119-1129. CrossRef | PubMed
3 McNamee D, Horton R. Lies, damn lies, and reports of RCTs. Lancet 1996; 348: 562. Full Text | PDF(19KB) | CrossRef | PubMed
4 Kjaergard LL, Villumsen J, Gluud C. Reported methodological quality and discrepancies between large and small randomized trials in meta-analyses. Ann Intern Med 2001; 135: 982-989. PubMed
5 Juni P, Nartey L, Reichenbach S, Sterchi R, Dieppe PA, Egger M. Risk of cardiovascular events and rofecoxib: cumulative meta-analysis. Lancet 2004; 364: 2021-2029. Summary | Full Text | PDF(193KB) | CrossRef | PubMed
How can you explain that a homeopathic substance, which, according to you, is supposed to be placebo, can work effectively on animals and plants?
The straight forward answer here is that homeopathic substances do not work on animals or plants. This is a myth that has been created by repetition. There are no quality studies that show homeopathy working in animals, but humans do have an astounding ability to see what they want to see and believe what they want to believe. Animals can recognize when they are being cared for and can respond to interventions even if these are physiologically inactive. Again, Andre seems to be unaware of the massive amount of literature that exists on this issue. To simplify, a vet who is using homeopathic remedies, is using his authoritative position to convince the animal owner that the animal being treated with homeopathy is getting better. This can lead to a placebo effect by proxy where the animal’s owner is assured that the treatment will work and therefore becomes less anxious. An animal can sense this and responds to its owner’s more positive attitude. This can provide the illusion of efficacy but the animal remains medically untreated. Here are a couple of interesting analyses:
What is your opinion on homeopathic and naturopathic schools being represented at McGill Graduate Schools and Professional Fairs?
These schools are legal educational institutions so there is no reason not to make students aware of their existence, especially given that admission requirements are much lower than for medical school. My experience has been that the few students who have gone in this direction have done so after failing to get into a medical school. There are, however, undoubtedly some students who find the naturopathic approach attractive and make this a first career choice. They should, however, have full knowledge of what they are getting into as well explained in the December 14 entry of science based medicine:
In many countries such as Switzerland, homeopathy is part of medicare, therefore considered effective enough for public health benefit. How can it be called placebo-based?
There is no requirement of efficacy to include homeopathy in a health care system. In a democratic society governments will cater to public desires whether these are scientifically legitimate or not. Since homeopathy is cheaper than conventional medicine, it can alleviate government spending on health care. Most people visit homeopaths for minor conditions that generally resolve by themselves or for chronic conditions that have stymied conventional doctors but which would not be helped by further visits. But there is always the concern about including a treatment that has not been properly validated in national health care systems which are publicly funded. This issue is coming to the fore in the United Kingdom where homeopathy has been on the National Health Service but a House of Commons Science and Technology Committee that has examined the situation has recommended that homeopathy be removed from the National Health Service.
Can you comment on the Swiss study?
This was not a study but an opinion rendered by a group who judging by their affiliations and publications have a favourable view of alternative therapies. The report summary was written by Gudrun Bornhöft and Peter Matthiessen: Chair in Medical Theory and Complementary Medicine, University of Witten/Herdecke, Germany, Ursula Wolf, Institute for Complementary Medicine (KIKOM), Inselspital, University of Berne, Klaus von Ammon, Stephan Baumgartner and André Thurneysen, Institute for Complementary Medicine (KIKOM), Inselspital, University of Berne, Marco Righetti, Medical Practice, Zurich, Stefanie Maxion-Bergemann: PanMedion Foundation, Zurich. Dr. Righetti, who is listed as “Medical Practice” actually practices homeopathy and has written voluminously on the subject. There were many foxes in charge of this henhouse. Again, Andre appears to be totally unaware of just how much astute scientific criticism has been piled on the Swiss ‘”study.” Here is some interesting reading:
Isn’t it true that the only non-war statue in Washington DC is of Samuel Hahnemann?
There is an impressive statue of Hahnemann in Washington but it certainly is not the only non-war statue in DC. There are numerous statues, even in Congress, that have nothing to do with war. Indeed, outside the National Academy of Science building, there is an interesting statue of Albert Einstein who of course opposed war.
Is there any value in observational, animal and in vitro studies?
Absolutely. These all contribute to evidence-based science.
Have you ever been to a homeopath?
Only incognito. A remarkable experience. Not a positive one.
Someone in the audience wanted to know if you would submit a proving on your own?
No problem. I would be willing to take any homeopathic product.
As is evidenced by the foregoing, it is my view that homeopathy is a form of unscientific, fanciful thinking that is pursued by generally well-meaning individuals who have put critical thinking on the back burner and appear to be oblivious of the established fact that the placebo response occurs consistently in 30-40% of cases likely due to expectation of relief by the patient.