Scientific article: "Evidence for Homoeopathy"

Robert Hahn 29/01/2015

News

Reading time: - minutes

by Prof. Dr Robert Hahn,
Researcher at Linköping University

The following article is by Robert Hahn, researcher and Professor of Anaesthesia and Intensive Care Medicine at Linköping University, Sweden. Hahn has published over 300 scientific articles in anaesthesia and intensive care medicine and has received several research awards. He had no prior involvement with homeopathy. However, he had noticed that an astonishingly unscientific debate about the evidence for homeopathy was taking place. He therefore wanted to form his own view and studied the current research. He arrived at surprising results.

The article is a summary by Robert Hahn of an article he published in October 2013 in the journal Forskende Komplementærmedizin.

About three years ago the organisation „Vetenskap och folkbildning (VoF)” (Science and Public Education) launched a summer campaign against homeopathy. During the political week in Almedalen VoF equipped a group of teenagers with T‑shirts bearing the slogan "Jag är skeptisk" - (I'm sceptical). Through this action the group ended up on television together with the astronaut Christer Fuglesang, where they were allowed to argue freely and undisturbed against homeopathy for an evening. Homeopathy was presented as a great fraud. The teenagers said one after another that there were no scientific studies proving that homeopathy worked.

I seriously wondered — how can boys of that age know this? Do they really have the necessary skills, knowledge and maturity to read and understand the relevant literature? My conclusion is — No, absolutely not. Reading scientific literature is complicated and requires (to some extent) that one works professionally in medical science. It was obvious that these boys had been told in advance by older members of VoF exactly what to say.

They served as a kind of advertising banner for something they neither understood nor could judge. They trusted people who present themselves as representatives of public education and science and then publicly voiced, in their own names, those people's version. They assumed that this version was correct and would be accepted as self‑evident in the scientific world.

Is there evidencefor homeopathy?

This disturbing fact prompted me to write about the existing evidence in favour of homeopathy. My three blogs on the subject attracted enormous attention in late summer 2011. The aim was to go through the scientific articles that addressed the question of whether homeopathy in medical conditions is, statistically speaking, more effective than a placebo (globules or dilutions).

The attention those blogs received had the effect that I was asked to summarise them and publish in a medical journal. This was published in October 2013 under the title: Homeopathy: Meta-analyses of Pooled Clinical Data in the journal Forskende Komplementærmedizin (2013; 20: 376–381).

If you want an overview of my other more than 300 scientific articles, you can view them at the following link http://www.ncbi.nlm.nih.gov/pubmed/?term=hahn+rg. However, this is my only article on homeopathy or alternative medicine, but I think it serves its purpose.

Systematic reviews and meta-analyses

The starting point for this article was the change in the approach to evaluating medical treatments that took place in the mid‑1990s. Previously this approach was academic. In other words, one wanted to understand the mechanism by which a treatment worked in order to accept it. This criterion leads to problems. Many treatments whose mechanism is still unknown are nevertheless used in hospitals. Nitrous oxide, commonly used during childbirth, is one example.

Medicine moved to the practical notion of "evidence‑based", which involves considering the entire literature in the field in the form of systematic reviews or meta‑analyses. The question asked was: does the treatment work or not? In a meta‑analysis all scientific studies are pooled in which patients are randomly assigned to receive the study treatment or a placebo (control group). If the study treatment (in this case homeopathy) achieves a statistically detectable better or faster recovery than the placebo, this is taken as evidence that it is more effective. How much better is given by an odds ratio, where a value greater than 1.0 indicates "better than placebo". The improvement being statistically significant is shown by a confidence interval (CI) that does not include 1.0 (the risk that the improvement can be explained by chance is less than 5%).

The reader should be aware at this point that it was no longer necessary to demonstrate the mechanism behind the treatment. This change created considerable problems for academic medicine (including VoF), which found it increasingly difficult to reject alternative treatments. It can now suddenly happen that these work even though the mechanism appears unclear, or the methods (in our time) are regarded as unreasonable.

Linde's paper

In Klaus Linde's meta‑analysis from 1997 (Reference 1) we have the first and, in my view, most honest paper in the field. Linde found an odds ratio of 2.45 and a CI of 2.05–2.93, a clear effect in favour of homeopathy compared with placebo. This result was based on 89 studies. One could follow how the selection of the individual studies was made and there is a reasonable balance between the number of included studies and the possibility of detecting effects statistically. Interestingly, Linde included an overview showing how effective homeopathy is for different diseases. It seems clear to me that homeopathy mainly exerts an effect in milder conditions related to the immune system (allergies, hay fever, etc.). Linde's study understandably caused a stir in the academic world. A countercurrent of articles followed that apparently aimed to refute the result. The first criticism concerned the quality of the studies, which can be categorised by the so‑called Jadad score. When this was applied to Linde's 89 studies, a trend emerged that studies with a low Jadad score showed a greater effect. Linde, however, showed that even for the highest Jadad score the positive effect of homeopathy actually increased (Reference 2).

Edzard Ernst

This argument was not accepted by the alternative‑medicine activist Edzard Ernst, who holds the view that one should draw a line between all Jadad scores except the highest and thus calculate the theoretical effect of the best studies (Reference 3). In my view this is completely crazy, because real data are replaced by virtual (theoretically calculated) data.

Ernst wrote in 2002 an alleged meta‑analysis of homeopathy which in truth is a systematic review (Reference 4). I discuss below some peculiarities of his review. Ernst, for example, swings between rejecting papers that show specific effects on diseases, while in other cases he excludes papers because they do not show specific effects. I have never seen a scientific writer as obviously biased as Edzard Ernst.

Cucherat's "type II error"

Cucherat (Reference 5) is basically an honest writer who made great efforts to finally reject homeopathy as a treatment method. Conventional meta‑analytic statistics were not used in this case; instead he chose, from five possible methods, the one most unfavourable to homeopathy. Cucherat also resorted to a technique that Edzard Ernst used in many studies, namely removing virtually all qualitatively good material. In the case of homeopathy it was not enough to remove 90% of the originally intended studies for analysis. Homeopathic treatment remained superior to placebo. When the author removed 95–98% of all studies, the statistics flipped. So it's not that the treatment is poor, but that statistics simply need more material to analyse. One can say that the author deliberately created a "type II error". I am extremely critical of several points Cucherat raises to reject the selected studies. Sometimes they are only marginal details that normally would not lead to exclusion.

First, Cucherat removed 101 of a total of 118 studies from his analysis. Homeopathy nevertheless remained highly effective. The risk that the difference between the effectiveness of homeopathy and placebo was due to chance was less than 3.6 in 100,000. Even when he removed further studies and only nine of the very best remained, homeopathy still showed a statistically significant effect. When he removed a further four studies, the statistics wavered and the risk that the effect of homeopathy compared with placebo could be explained by chance was 8.2 in 100. The conclusion? Cucherat thinks homeopathy is not effective.

I think Cucherat is a coward. At the risk of ruining his career and being at war with organisations like VoF, he did not dare to publish what his material actually showed. By ignoring more than 96% of the studies he had initially judged good enough for his meta‑analysis, he intentionally manufactured a "type II error".

Shang selected just as wildly

The next meta‑analysis was written by Shang and published in 2005 (Reference 6). Here, as in Cucherat, 96% of all studies were removed, but unfortunately without giving a reason. The odds ratio was twisted until the result — which despite all removals of study material showed homeopathy to be 13% more effective than placebo — was inverted and presented as if placebo were 13% more effective than homeopathy.

Shang also made use of the so‑called "funnel plot", which is completely inappropriate and unscientific when, as in this case, different diseases are mixed. The reason is that studies with high expected effect (here e.g. hay fever) always include fewer patients than studies where the expected effect is judged to be low. This is a generally recognised ethical norm. I must also note that the authors with whom Shang collaborated had published a very negative article on homeopathy some years earlier, which makes me doubt the group's objective standpoint.

With this I end my summary. Above are the six most important articles on the subject. My review in Forskende Komplementærmedizin contains a total of 22 references.

There is certainly more to say, but there is not enough space here.

Evidence and belief

So what does Robert Hahn believe? I believe nothing. It is not my task, nor of practical interest, since I do not work with homeopathy. What interests me is to expose the tendency and practical bias of scientists who are not nearly as objective as they would have you (and others) believe, waving their professorial titles to legitimise their personal beliefs. These are the people who mislead naive youngsters into following their ideological teachings. I don't like that, and my aim is to oppose it.

However, I must admit that I am critical of the starting point used in the studies cited here to answer the question of whether homeopathy is more effective than a placebo. That question is not very well posed, since the answers are based both on diseases where homeopathy is superior to placebo and on diseases where it is not.

To advance this question, it would be better to focus instead on the diseases that Linde in 1997 already mentioned as particularly suitable for studies. Rather than expanding knowledge about homeopathy's effects in different diseases, an ideological battle arose over whether homeopathy works at all. My summary should have made it clear that those who tried to prove that homeopathy does not work had to work very hard to do so.

Scientists are strongly influenced by their ideology

So who can you trust? We can begin by discounting Edzard Ernst. I have read several other studies he has published, and they are all dubious. His work should not be credited. Both Cucherat and Shang deliberately created a "type II error" by ignoring almost all actually published studies. The reason? Well, had they included just a few more studies, and been content to exclude "only" 90% of the records, it would have turned out that homeopathy was more effective than placebo. And you wouldn't want that, would you? So they simply excluded so much material until they constructed a "type II error".

I assume the authors had to try various tricks to find the result they wanted from the start — namely that homeopathy lacked effect. This mathematical game hardly appears serious. In this way one can preserve one's academic purity, remain welcome in the hallowed halls of science, and avoid the risk of being ridiculed or publicly mocked by organisations like VoF.

I am fascinated to see how much the scientific world is driven by its ideologies. In the case of homeopathy one should adhere to what the evidence reveals. And that says that the effectiveness of homeopathy can only be made non‑detectable if one removes 95–98% of all studies conducted in the field, studies that according to accepted principles could very well be analysed and which, among other criteria, include a placebo control group. Or one demands the use of a completely inappropriate method such as the funnel plot.

The reactions of science

I would like to give three examples of reactions from scientifically trained people to homeopathy in general and to my work and the evidence presented here.

A colleague posted a comment on Facebook. He said he was very surprised that I was interested in homeopathy since I am otherwise so scientific. Homeopathy — clearly a taboo subject. When it comes to homeopathy, you are not allowed to search for or evaluate evidence if you want to continue to be regarded as a good and respectable scientist. This kind of fear should not exist in science. But the fact that it exists is very striking.

Another example of how to suppress meaningful debate about homeopathy at the outset is the derogatory language frequently used. The Professor of Complementary Medicine at the Karolinska Institute supported VoF’s summer campaign against homeopathy in 2011 by declaring in the media that homeopaths are all charlatans who only peddle nonsense.

A third method is simply to lie. In my article in Forskende Komplementærmedizin I mention, among others, Dan Larhammar's article (which I did not name), which appeared in the newspaper Svenska Dagbladet during the summer campaign 2011. In it Larhammar argued that homeopathy is a scientific absurdity (which is correct). But he also wrote that science had found that homeopathy has no effect, referring to only two review studies summarising the state of research in the field. According to Larhammar, these two reviews were those by Shang and Ernst.

First, there are many more review reports on the subject. If we read all six reports mentioned above, it also becomes clear that those two are among the most negative towards homeopathy. Both men unfortunately chose very questionable methods to reach their conclusions. It seems Larhammar selected the literature so that its message matched his own.

A truly honest scientist would have assessed all the articles mentioned here. The picture would then be that it cannot be conclusively shown that homeopathy is a "sham medicine". The evidence analysis rather shows the opposite. Dan Larhammar succumbed to the temptation to lie to conform to the view of VoF (and his private view). He abuses his position as a professor and allows his personal worldview to take precedence over his scientific message.

Dan Larhammar is a specialist in everything

It is primarily Dan Larhammar, the former chairman of VoF, who attacks homeopathy. He has done so many times. I personally am convinced that Larhammar was the driving force behind VoF’s 2011 summer campaign. Within the VoF sphere one does not need special training or experience to be considered an expert in all areas. In the newspaper Dagens Nyheter (5/1 2014) Dan Larhammar appeared as an expert on primary education. He and others from the Academy of Sciences analysed and understood what is going wrong in schools. A few years ago he appeared in Dagens Nyheter as an expert on autism spectrum disorders. In the book Vetenskap eller villfarelse (Leopard Förlag 2005) he is named as an expert on neurocognition. In reality Dan Larhammar is a pharmacist and does basic research on fish. He lacks expert knowledge in all the fields in which he presents himself as an authority with his professorial title. He reminds me of Gus Goose from Donald Duck. He, too, was an expert on everything. Omnipotence is a heavy burden to bear.

Unfortunately, for copyright reasons I cannot place my complete article on homeopathy on the internet. You can find the abstract to my article at the following link (http://www.ncbi.nlm.nih.gov/pubmed/24200828).

If you want to read the full article you must unfortunately purchase it online (approx. €5). As mentioned above, this is due to the publisher's copyright. Unfortunately there is nothing I can do about that.

References

1. Linde K, Clausius N, Ramirez G, Melchart D, Eitel F, Hedges LV, Jonas WB. Are the clinical effects of homeopathy placebo effects? A meta‑analysis of placebo‑controlled trials. Lancet 1997;350:834–43.

2. Linde K, Scholz M, Ramirez G, Clausius N, Melchart D, Jonas WB. Impact of study quality on outcome in placebo‑controlled trials of homeopathy. J Clin Epidemiol 1999;52:631–6.

3. Ernst E, Pittler MH. Re‑analysis of previous meta‑analysis of clinical trials of homeopathy. J Clin Epidemiol 2000;53:1188.

4. Ernst E. A systematic review of systematic reviews of homeopathy. Br J Clin Pharmacol 2002;54:577–82.

5. Cucherat M, Haugh MV, Gooch M, Boissel J‑P. Evidence for clinical efficacy of homeopathy. A meta‑analysis of clinical trials. Eur J Clin Pharmacol 2000;56:27–33.

6. Shang A, Huwiler‑Münterer K, Nartey L, Jüni P, Dörig S, Sterne JA, Egger M. Are the clinical effects of homeopathy placebo effects? Comparative study of placebo‑controlled trials of homeopathy and allopathy. Lancet 2005;366:726–32.

- More at: http://roberthahn.nu/2014/01/05/min-vetenskapliga-artikel-om-homeopati/#sthash.GBpiJtin.dpuf

The author:

Prof. Dr Robert G. Hahn

Research Director, Södertälje Hospital

Professor of Anaesthesia, Linköping University

Associate Professor, Karolinska Institute

For the original article, follow this link

********************************************************************

Category: scientific article

Robert Hahn