Does Arnica Gel Work for Pain?

Arnica montana

Is it good for aches & pains when diluted down to a trace?

This is a detailed article about homeopathic (extremely diluted) creams for aches and pains. Traumeel (now T-Relief in the US and Canada1) is the best-known, but there are quite a few others these days. Most contain several herbs, but mostly they feature Arnica montana.2 Manufacturers claim the main benefits are “anti-inflammatory effects,” and they have a great reputation for being good for muscular pain, joint pain, bruising, and sports injuries.3

I have often heard high praise of arnica for both people and animals. It is spoken of as though it’s a kind of “secret” weapon against pain, under-rated and little known but powerful. But can testimonials be believed? If it works, how does it work? This article explores the hope and the science.

Other kinds of pain

No product can treat every kind of pain. Arnica creams are mainly intended and advertised for one kind: inflammatory pain.

Many causes of pain cannot be treated by any kind of ointment. For instance, by far the most common cause of stubborn pain is the humble “muscle knot,” technically known as a trigger point. Many visitors to this page are probably suffering from this kind of pain, and looking for solutions, hopefully a nice easy one in a tube — but trigger points are not inflamed and they are not an injury. They are a “sneaky” cause of pain that is often overlooked. People can suffer for years, when at any time they could have been helped by a 10-minute self-massage. Not every case is that easy, but some are.

People probably try arnica cream for quite a few problems that aren’t actually inflammatory in nature. Many stubborn, erratic, aching pains are quite likely to be caused by muscle and might be better treated simply with massage. Please be aware of this possibility — and other possible non-inflammatory causes of pain — as you continue reading about arnica.

Homeopathic or herbal?

Homeopaths treat with extremely diluted substances, and most homeopathic products are not “herbal” at all. Homeopaths believe that diluted ingredients are medicinally effective because the solution “remembers” something about the active ingredient — sort of like an echo. They also believe that this effect gets more potent as the original ingredient is further diluted — less is more, the homeopathic “Law of Infinitesimals”4 — even to the point that no molecules of the original substance remain, just their essence or vibrational imprint on the water. (“Vibrations” have always been a surprisingly huge theme in medicine and quackery: see Vibration Therapies, from Massage Guns to Jacuzzis.)

Arnica creams are ambiguous hybrids using weaker dilutions that make them both (1) normal herbal products as well as (2) homeopathic. And yet the flagship ingredient, arnica, is diluted to a truly homeopathic degree (very, very diluted). Thus arnica creams are usually mainly homeopathic, while still being a little bit herbal. In this article, I will look at both sides of this split personality.

Many people who use these products aren’t aware that they are homeopathic, because it is often not labelled that way. Even if they did, many people do not even really know what homeopathy is. However, many others buy them for the homeopathy.

Some products are not actually homeopathic, but use the marketing power of the word — its popularity — to sell products. For example, the cold-remedy Zicam is labelled as homeopathic, but there’s actually enough zinc in the product that it could damage your sense of smell, which is why the FDA banned in it in 2009.5 Some homeopaths might consider Zicam to be a “low potency” preparation, but such a mild dilution is usually considered too concentrated to be truly homeopathic.6

What’s in a homeopathic arnica cream? Trace amounts of the “main” ingredient, arnica

These are homeopathic remedies (albeit “low potency”) because the main ingredient is so diluted. If you just want some arnica, you won’t find much of it in a dab of one of these creams. On one popular brand, the packaging reads “Arnica montana 0.75g” and “3X” (or 3D), a homeopathic notation meaning “.75g of arnica diluted to 10% three times.”7

Although much less diluted than many other homeopathic products, which can go up to 30X or even 60X, 3X is still a lot of dilution. When all the math is done, that works out to 7.5 micrograms of actual arnica per (one gram) dosage or dab. Even in a whole tube, that’s only 0.75mg — not even one thousandth of a gram! It is difficult to overstate how little arnica we’re talking about here. The same amount of the most potent toxins known to man are not considered dangerous.8 This is what a scientist would call a “trace amount.” See footnote for all the math.9

Finally, it is well worth noting that because the dose is applied externally, most of it will never make it through the skin. The actual dose delivered to the tissues is probably much lower still. We know from a century of incredibly detailed microbiology and organic chemistry that drugs work through chemical effects on the human body. Is a “trace” amount of arnica enough for a chemical effect?

Could trace amounts of arnica have a chemical effect?

As they might say on MythBusters: “It’s plausible … but not likely.” And it is only just barely plausible, chemically speaking.

Dilution really takes the punch out of chemical compounds in a big way.10 Even if an infinitesimal concentration of a potent herb does have an effect, it would certainly not be a clinically significant effect … and that’s the key to this question.

Have you ever tangled with poison ivy? I have! On the edge of a natural hot spring in the Kootenay region of British Columbia. It was memorable, but not all that bad. If the poison ivy toxin (urushiol) had been diluted to 10% of full concentration — just 1X — I might not have even noticed it. Would I rub 3X urushiol in my eyes? No worries!

Indeed, dilution makes homeopathic treatments safe for exactly this reason. (Unless they are contaminated. Which is shockingly common.11) A microgram of an herb is simply not enough to have any significant effect — a microgram dose of anything is roughly the lower limit of chemical relevance. And a subtle effect is just not helpful for a terrible bruise or raging bursitis.

How about undiluted, pure herbal arnica? Would that work, at least?

Flower power?

Some people like their arnica diluted. Other people want to rub on the pure stuff. Is it 650,000 times more powerful?

Is the arnica plant itself medicinal? Perhaps. It’s worth considering the potential of full-strength arnica — even though you won’t find it in the most popular homeopathic arnica creams, there are plenty of other products that do deliver arnica flower extract as undiluted as possible. For instance, A Vogel’s “Atrogel” product offers “the liquid extract of fresh arnica flowers. 1g of gel contains 500mg of extract which is equivalent to an average of 160mg fresh arnica flowers.”12 That’s a whopping 65,000 times as much arnica as you’ll get from the same size dose of homeopathically diluted arnica.13

It is an oddity that people may might buy arnica in either of these concentrations expecting basically the same result! There’s probably something wrong with that picture.

There is almost no question that the undiluted flower has some anti-inflammatory effects. The problem is that it’s unlikely that the effects are strong enough to justify putting it in a bottle. There is no clear, strong, confirmed effect on inflammation, swelling, or bruising … according to the limited research done so far.14 That’s why medicinal anti-inflammatory creams like Voltaren Gel: Does It Work? use a drug like diclofenac — not arnica! Arnica has some potential, but no one’s rushing to confirm it. The research is limited mainly because thus far arnica just hasn’t shown enough promise. (Meanwhile, topical diclofenac actually works quite well.1516)

Arnica certainly has a well-established place in folk medicine as an anti-inflammatory. Such historical usage of an herb like arnica is a great starting place for a medicine. But an herbal remedy cannot be proven effective or safe based on that alone. Plants are like little chemistry sets! And most plants have a variety of minor biological effects. Only a few species have a strong effect of any kind … and many of those are toxic. A chemical effect can only be considered a medicine when it turns out to have a good effect on human physiology, with side effects we can live with. Since every plant consists of hundreds or even thousands of chemical compounds, it takes many years and lots of science to determine which compounds are significant, how significant they are, and how safe, and for what uses, and in what dosages. Decades after its debut, a plant-derived medicine can still turn out to have surprising “gotchas.” Arnica has not been subjected to such scrutiny for the simple reason that no study of arnica has ever revealed more than a few minor chemical effects. If there were some potent, useful compound in there — if it were anywhere close to as good as other famous plant-derived medicines — then pharmaceutical researchers (pharmacognosists) would probably have been all over it decades ago.

Arnica montana

It’s “Montana” because it grows in the mountains — mountain arnica. But would you climb a mountain for it? There is little doubt that this flower has some anti-inflammatory effects, but it’s not clear that they are strong enough to justify bottling the stuff. The research is incomplete … but partly because it doesn’t show enough potential.

Not just arnica! Can echinacea, belladonna, or calendula and friends have a chemical effect?

There are several other common herbal ingredients in many of the “arnica” creams, which makes them quite a bit more interesting than homeopathic arnica alone would be. This is where it really becomes clear that these are hybrid products: both homeopathic and herbal … at least a little bit. Analyzing one popular brand, eight of the herbs are just as diluted or more diluted than the arnica 3D, so they are chemically neutralized. But seven of them are less diluted (1X, or one tenth). These low dilutions are only just barely homeopathic, and they have some potential for conventional pharmacological effects — although probably only minor ones. (Remember, we’re still just considering the ingredients as chemicals right now, and putting off the question of whether or not homeopathic dilution makes them more “potent.”) Chemically speaking, these are still super low doses of substances that aren’t exactly seriously potent to begin with.

That doesn’t mean they have no effects. They may. Several slightly diluted herbs could have medicinal effects. A review of the research is coming later on in the article.

However, arnica is the star of the homeopathic pain cream show, and the brands that use it probably couldn’t thrive without it. There are also several other products that only contain homeopathic arnica, and there are many studies of homeopathic arnica alone — without its companion ingredients. Because the idea of homeopathic arnica exists independently of any brand, we really do need to know if it works without the other ingredients. That is why this article mostly focuses on arnica. And yet 3X arnica may be too diluted to have a “conventional” biochemical effect. If homeopathic arnica works for aches and pains, it has to do it homeopathically. And that is a bit of a concern …

It’s all about the homeopathy

Homeopathy is a super controversial subject. I will do my journalistic best to review the key issues without passing judgement myself, and leave it up to my readers to decide for themselves. I link to several information resources about homeopathy in an appendix to this article. There’s one pro-homeopathy resource that matters more than all the others combined, though: The National Center for Complementary and Integrative Health (NCCIH), an American organization.

NCCIH’s raison d’etre is to study and validate alternative health care practices such as homeopathy, and they have truly massive taxpayer funding, and they employ and fund hundreds of real scientists.17 If you can’t cite NCCIH in support of homeopathy, who can you cite? Unfortunately, even after many years of expensive research, NCCIH has still not produced evidence that homeopathy works. This is stated on their website:

Most analyses of the research on homeopathy have concluded that there is little evidence to support homeopathy as an effective treatment for any specific condition.

Homeopathy: An Introduction, National Center for Complementary Medicine (NCCIH.nih.gov)

So what’s the problem?

Three basic homeopathy controversies

Why is homeopathy controversial? Here are some of the major concerns that have been raised by critics:

1. The inability-to-end-the-argument problem. A truly effective medicine should be able to decisively put an end to controversy with benefits that are clearly impressive, not benefits that are subtle and perpetually debatable.
2. The zero concentration problem. It’s one thing to believe that low concentrations of active ingredients might be medicinal, but perhaps too much to believe that zero concentrations are (“Law of Infinitesimals”). Homeopathic dilutions are not just extreme, they’re routinely diluted to the point where literally none of the original substance remains at all.
3. The selective memory problem. Homeopaths have proposed that water “remembers” its active ingredients so they can work even after extreme dilution. However, even the most pure water contains countless trace contaminants, compounds of every description — how are those “forgotten”?18 And how is the homeopathic agent selectively remembered?19 There are, of course, no answers to these questions. But the jokes write themselves. eg

And two more homeopathic controversies need a little more attention.

Homeopathic arnica breaks homeopath’s own Law of Similars

The homeopathic “Law of Similars” is that “like cures like.” This idea is one of the two most important “laws” of homeopathy (along with the Law of Infinitesimals).

Guided by the Law of Similars, homeopaths choose ingredients whose normal effects are in some way like the problem they want to treat. For instance, an inflammatory condition is treated with an ingredient that would normally cause inflammation when undiluted.

Homeopathic arnica seems to break the Law of Similars! Arnica is believed to be an anti-inflammatory herb. Therefore, using arnica is not “like cures like,” and breaks homeopathy’s own rules … without any explanation I have ever heard.

The Law of SimiIars defines homeopathy. It is puzzling to me indeed that the most popular of all homeopathic products would break that law, and it seems like a major inconsistency. I may be ignorant of other homeopathic remedies that do not obey the Law of Similars, and the reasons for doing so. Homeopaths are invited to contact me with a proposed explanation — I will publish any explanation provided, in this spot. I have received plenty of email from homeopaths about this article, much of it “corrective” in tone — but not one has addressed this point.

The one-size-fits-all problems

Homeopathic arnica is often recommended for any kind of pain problem, even though many of them don’t really have much in common with each other biologically. For instance, consider tendinitis:20 acute tendinitis and chronic tendinitis are quite different conditions, with different chemistry.2122 It’s really unlikely that one medicine could be truly effective for both.

Another good example: delayed onset muscle soreness (DOMS), that nasty soreness for a day or two after exercise. It can be savage, and its physiology is quite mysterious.23 Whatever causes it, it’s quite different than any other common pain problem. It’s also widespread and deep in the tissues — not in one convenient spot to rub a cream into. It is simply impractical to treat DOMS with a cream, unless you could buy it in vats.

Anything is possible, but it does seem unlikely that any single medicine can be effective for many different kinds of pain.

The trouble with the anecdotal evidence

This stuff is popular! A great many people firmly believe that it works. But a lot of people thought drinking mercury was a great cure … and it was killing them.24 So it’s hard to trust what people believe.

The standard arnica story is, “I was amazed by how fast I healed.” Maybe so. I have experienced this myself on occasion, experimenting with the product. It may seem strange to many readers, but I actually doubt my own senses. There are many good reasons why, but this is the most important: how fast healing would have occurred without the treatment? How can we possibly know that?

Patients are generally in a poor position to judge how quickly healing normally occurs.25 It is impossible to accurately predict healing time from anything — every person and situation is different — and so patients can’t know that their recovery was accelerated. Faster than what? A guess? A pessimistic one? What about an optimistic one? Surprisingly quick recoveries happen, cream or no. They can easily create an illusion of success:

Some patients recover very quickly from surgery. If those taking arnica attribute their good recovery to the homeopathic remedy and this apparent association is widely reported, it is easy to see how the reputation can build. Since the experiences of patients who recover well without taking arnica and those who receive no benefit from arnica are less likely to be reported, the myth becomes reinforced.

Stevinson et al., 2003, Journal of the Royal Society of Medicine

Chronic pain is still common despite the popularity of homeopathic arnica

What if there really was a “wonder drug” for pain? Powerful and safe? Effective on nearly any kind of pain? What if it sold as well as homeopathic arnica does? What if that product was already being used by literally millions of people?

Shouldn’t the world be nearly free of pain?

If you really want to know how well a pain remedy works, ask someone with severe chronic pain. I get many notes from people telling me how diluted arnica helped their pain, but these stories are all about relatively minor injuries and pain problems — the kinds of things that tend to go away on their own sooner or later anyway.

I do not receive notes from people claiming to have treated severe pain problems with arnica. There are no emails from runners who fought for years with iliotibial band syndrome — a repetitive strain injury of the knee which can be quite nasty — only to finally beat it with a daily dab of a few molecules of arnica. I have never gotten a message from a formerly debilitated victim of a painful inflammatory disease who claimed to be liberated from their prison of agony by homeopathy.

It’s not for lack of trying. Severe pain drives patients to all kinds of cures. But personally, I have yet to hear a success story from someone who used diluted arnica to solve really serious pain.

The state of the research, according to proponents

It doesn’t matter how homeopathic arnica works if it works. But does it work? As they’d say in Missouri, the Show-Me State, “If it works, show me.” Better yet, impress me: if it works, there’s no reason why it shouldn’t be able to ace a fair test, showing obviously superior results to a placebo (or even competing products). What is the scientific perspective?

Obviously fans of these products believe that the research supports their position. A 2021 paper by three well-known homeopathic boosters is the first review of homeopathic Arnica in years, and it predictably concludes:26

Homeopathic Arnica has a small effect size over and against placebo in preventing excessive hematoma and other sequelae of surgeries.

But it is a seriously flawed review of even more seriously flawed research.27 What is it based on, in theory?

“No other homeopathic remedy has been subject to more controlled clinical trials” (Ernst et al.) … but that’s really not saying much. Only about a dozen noteworthy studies about the efficacy of homeopathic arnica or similar products have ever been undertaken, a couple dozen if you include some of the really poorly designed ones, and there’s definitely some of that. (Homeopathy research in general is objectively bad.28) Considering the popularity of these products, that’s almost no research at all — much less, say, than what is required for FDA approval of a new drug — and a lot of that research has been of generally poor quality.293031 And only the poor quality studies show that homeopathy works, while good quality studies have shown it doesn’t.32 Even when they sound positive, they actually aren’t.33

Also bear in mind that NCCIH — that very well-funded organization that exists to test alternative medicine — publicly, clearly admits that homeopathy doesn’t beat a placebo.

I analyzed ten scientific papers cited on a popular brand’s website. Some are well-known citations cited in many other places, while others are utterly obscure. In any case, one manufacturer believed that these citations supported their product:

• Zell J, Connert W, Mau J, Feuerstake G. Treatment of Acute Sprains of the Ankle: A Controlled Double-Blind Trial to Test the Effectiveness of a Homeopathic Ointment. BioLogical Therapy. 1989;7(1):1–6.
• Böhmer D, Ambrus P. Treatment of Sports Injuries with Traumeel Ointment: A Controlled Double-Blind Study. BioLogical Therapy. 1992 Oct;10(4):290.
• Metelmann H, Zenner S. Application Possibilities of Traumeel S Injection Solution: Results of a Multicentric Drug Monitoring Trial Conducted on 3,241 Patients. BioLogical Therapy. 1992;10(4):301.
• Thiel W, Borho B. The Treatment of Recent Traumatic Blood Effusions of the Knee Joint. BioLogical Therapy. 1994;12(4):242.
• Conforti A, Bertani S, Metelmann H, et al. Experimental Studies on the Anti-Inflammatory Activity of a Homeopathic Preparation. Biomedical Therapy. 1997;15(1):28–31.
• Zenner S, Weiser M. Oral Treatment of Traumatic, Inflammatory, and Degenerative Conditions with a Homeopathic Remedy. Biomedical Therapy. 1997;15(1):22–26.
• Heine H, Schmolz M. Induction of the Immunological Bystander Reaction by plant Extracts. Biomedical Therapy. 1998;16(3):224–226.
• Lussignoli S, Bertani S, Metelmann H, Bellavite P, Conforti A. Effect of Traumeel S, a homeopathic formulation, on blood-induced inflammation in rats. Complement Ther Med. 1999 Dec;7(4):225–30. PubMed 10709306 ❐
• Porozov S, Cahalon L, Weiser M, et al. Inhibition of IL-1beta and TNF-alpha secretion from resting and activated human immunocytes by the homeopathic medication Traumeel S. Clin Dev Immunol. 2004 Jun;11(2):143–9. PubMed 15330450 ❐
• Gebhardt R, Köhnke D. Lindan-suppressed wound healing re-established by Traumeel — studies with hepatocytes, granulocytes, and hepatic stellate cells in a newly developed in vitro model. Research in Complementary Medicine [Forschende Komplementärmedizin]. 2007;14(41).

Here are some concerns about all or most of these papers:

• Most are quite old. The newest are from 2007 and 2004 (test tube studies that can’t prove much). The rest are between 26 and 35 years old. Pretty stale stuff.
• In most cases, these experiments were done to prove that homeopathy works — not to find out if it does — so the risk of bias is generally off the charts. Seven (7!) of the ten are published in two journals owned by a major homeopathic manufacturer, Heel (BioLogical Therapy and Biomedical Therapy). To call these journals “obscure” would be an understatement.
• Only 2 of the 10 are human clinical trials of diluted arnica in an ointment form, and both are over 20 years old, published in BioLogical Therapy. Only one other is a clinical trial of arnica in another form. The remainder are test-tube studies, animal tests, physician surveys, and so on.
• The trials cited are “positive.” They all show arnica working. But how much? Across the board, the effects were modest. And the problem with minor results is that they are often an illusion, a fluke that disappears when other scientists try to do the same test.

A closer look at those references

Traumeel injected into freshly injured knees (Thiel & Borho, 1994). This experiment may be old, but it was designed well enough. Germans Thiel and Borho compared pain, swelling and range of motion in 73 freshly injured knees (hemarthrosis, bloody swelling), and Traumeel won. The lucky patients who got Traumeel all did better — but not a lot better. And the authors admit that the folks getting a placebo were more overweight on average — which would tend to slow down knee healing. Maybe they really didn’t do as well as the patient who got Traumeel. But it might have just been because more of them were overweight and slower to recover. That is just the sort of thing that can create the illusion of a difference. It’s a great example of how relatively minor effects can vanish in a puff of statistics.

More injured knees (Böhmer & Ambrus, 1992). A virtual clone of the previous study, this is another good experiment on freshly injured knees by a couple of Germans (homeopathy was created by a German, Samuel Hahnemann). The results were positive … but modest. This study tested Traumeel ointment (the last one was injections). They measured swelling, skin temperature, strength, pain, and recovery time. They report excellent, clinically significant superiority of Traumeel with a high statistical confidence.

Clinically significant? Really? Maybe by some technical measure. For instance, the pain difference was less than 16% on their scale, after two weeks. Sixteen percent is barely subjectively detectable — “taking the edge off” territory. No one’s going to say “no” to a real 16% improvement in pain, but it’s hardly impressive.

If they’re real, these are worthwhile differences — just really clearly not impressive. But their reality must be questioned as well, due to the overwhelming risk of (unacknowledged) bias.

A silly survey (Metlmann & Zener, 1992). This is one of the most amusingly poor quality items on the list of references: an ancient, poor quality, scientifically meaningless survey of how cases involving homeopathic arnica turned out, according to German doctors. Patients weren’t surveyed, and other medications were involved. Even the best surveys are highly unreliable and do not constitute evidence of efficacy. The only thing this one proved is that German doctors like homeopathy! (Remember, homeopathy was invented by a German.)

A test tube test (Porozov et al., 2004) allegedly showing some effect on the function of immune cells. Even if true, this is not the same thing as proving that a product works. Test tube studies are notoriously hard to interpret. What happens to cells in a petri dish often bears no resemblance whatsoever to what happens to cells in a body. Test tube studies have their place, but you just can’t make too much of them — indeed, nothing at all until you’ve actually proven clinical relevance. The history of science is full of interesting test tube effects that couldn’t be converted in “medicine.” Porozov and his team got it bass-ackwards, starting with the assumption that a product works, and then looking for the explanation in a test-tube. That is not a good way to do science.

These researchers set out to find something that would confirm their beliefs. Unsurprisingly, they found it: they concluded that the mixture reduced “pro-inflammatory mediators” by roughly half … and even more so when the ingredients were more diluted. If true, it would be almost as cool as proving that unicorns exist. It’s more likely that they made a mistake — or several of them — that just happened to confirm what they wanted to find.

As we know from John Ioannidis, most research results are wrong34… especially biased test tube studies with implausible results in obscure journals. To prove anything, we need a lot of confirmation from independent scientists. So: what does newer, better science in less obscure journals say about arnica and homeopathy?

The state of the research, according to others

If you make even a token effort to separate the good research from the bad, it’s the small, poorly-designed studies35 that claim to have proven that homeopathic arnica works on the basis of weak data, while better and larger studies and reviews find nothing. There is one interesting exception, discussed below.

The only significant review of the scientific research is from 1998, by Ernst and Pittler.36 Dr. Edzard Ernst is a critic of homeopathy today, but he used to actually be a homeopath, so he understands the issues better than most other scientists. Ernst looked only at studies that compared homeopathic arnica to a placebo. He chose only eight to pick apart,37 but found even those wanting:

Two studies yielded a statistically significant positive result (ie, arnica superior to placebo), 2 studies had a numerically positive result (ie, no formal test statistics were applied but an advantage of the arnica groups was apparent) and 4 studies showed a significantly negative result (ie, arnica not superior to placebo).

… the more rigorous studies tended to be the ones that yielded negative findings.

On balance, the trial data do not support the notion that [homeopathic] arnica is efficacious.

… the hypothesis claiming that homeopathic arnica is clinically effective beyond a placebo effect is not based on methodologically sound placebo-controlled trials.

In other words, homeopathic arnica doesn’t help people enough for it to be obvious in a fair test.

And what about the research since 1998? In 2001, there was a notable exception to the generally negative trend: the infamous Oberbaum et al. trial of a Traumeel mouthwash for kids with cancer had undeniably positive results,38 making it the only trial of homeopathy I am aware of that had a happy ending and no obvious flaws. However, that does not guarantee perfection! It was a small experiment, just fifteen kids, and a test of a mouthwash for young cancer patients is hardly a good test of an ointment for minor injuries. The results may well have been a fluke.

But … it was a positive study. One, small positive experiment.

Another good study, better-designed and a little larger than most, was conducted in 2003. Researchers tested homeopathic arnica to see if it would reduce swelling and pain after hand surgery for carpal tunnel syndrome. It did not: “The trial data do not support the notion that [homeopathic] arnica is efficacious.”39

In 2006, two more studies are noteworthy because their data seems clearly negative to me, but the authors wrote conclusions that sound positive.4041 This is a common pattern: twisting essentially negative results to make them sound like good news. When exposed, such examples are quite useful and almost entertaining, a kind of two-for-one deal: they show both a negative result and the intellectually dishonest way that biased researchers warp the presentation of their research. That’s why the next and final science section of the article is devoted to a detailed example.

How homeopathic arnica can keep you from going out to dinner

I was totally cured! After a few days, there was nothing left of that bruise but a little discolouration!

This study was conducted by surgeons at a cosmetic surgery clinic — not scientists. They wanted to see if a homeopathic arnica product (SINECCH) would reduce post-operative bruising after nose jobs.42 Their experiment is tiny and sloppy. Key data is omitted from the conclusion — for instance, patients actually had more pain with homeopathy! Instead, a trivial improvement in the size of bruising is emphasized — a difference so small that it could only be detected by instrumentation,43 and not by patients or doctors, and was statistically significant (barely) on only two of five measurements taken.

And yet this paper has been widely cited as a “thumbs up” study of homeopathy by homeopaths everywhere.

Does that sound like a positive result to you? What good is a reduction in bruising so small that the patient cannot detect it? With an actual worsening of pain? This is the ultimate in ho-hum “scientific” results — and that’s even if we accept the published conclusions as statistically meaningful.

Which we can’t. The study was just too small.44 A manufacturer describes the results as “highly statistically significant”! But does the word “highly” appear in the paper? Nope — not once. You can check it yourself: here’s the full-text of the paper.

The surgeons reported that patients who received the product “actually did worse than those in the control group at each time point” (and that is an accurate quote). It was not a statistically important difference, but that doesn’t matter: data showing a lack of benefit is still a thumbs down (evidence of absence). If the stuff works, it should cause a statistically significant reduction in pain, not any increase in pain! And who really cares what else goes on if pain isn’t improved? Who would buy it if they knew that it would only reduce bruising slightly, but have zero effect on pain?

This highly relevant data is blatantly dodged in the paper’s conclusion. The authors do say that there were “no subjective differences,” but they do not explain that those subjective differences include no difference in pain.

Amusingly, the surgeons also collected data on patient readiness for a night on the town — that is, they asked when patients felt that their bruising had faded enough that they were willing to go out to dinner! It sounds a bit silly, but it’s an interesting piece of data to collect: doubtless plastic surgery patients are acutely conscious of their appearance, and their dinner-readiness does seem to be one fair way to measure their recovery, and it might be even more emotionally important to them than pain.

So, how’d they do?

Amazingly, once again, patients getting arnica did worse — not much worse, not statistically significant, but still definitely not better. They took an average 11.2 days before they felt ready for a public appearance. If only they’d gotten the placebo, they would have been out schmoozing by 10.6 days!

Omitting that point is a lie by omission — blatant concealment of what anyone would actually find most interesting about the results. To try to justify the concealment by arguing that the differences weren’t statistically significant is an obvious evasion — what we care about is the absence of a statistically significant benefit.

An honest conclusion to this study, sans jargon, would have read like this:

15 patients treated with [arnica] had about the same pain and bruise colouration as 14 patients treated with a placebo, and took just as long to feel socially comfortable with the appearance of their faces. The size of their bruising was less, but the difference was so small that it was only detectable with instrumentation, and even then it was barely statistically significant on only 2 of the 5 days on which measurements were taken.

That’s the real conclusion. As evidence goes, the best you can say is that it’s unconvincing.

Wouldn’t it be amazing?

It would be amazing if homeopathic arnica creams worked. And, if they work well enough to matter, it should be easy to prove. So, where is that proof? Where are the results of the simple experiments that could end the argument? Without it, I’m not convinced.

Did you find this article useful? Interesting? Maybe notice how there’s not much content like this on the internet? That’s because it’s crazy hard to make it pay. Please support (very) independent science journalism with a donation. See the donation page for more information & options.

About Paul Ingraham

I am a science writer in Vancouver, Canada. I was a Registered Massage Therapist for a decade and the assistant editor of ScienceBasedMedicine.org for several years. I’ve had many injuries as a runner and ultimate player, and I’ve been a chronic pain patient myself since 2015. Full bio. See you on Facebook or Twitter., or subscribe:

SUBSCRIBE

Get PainSci updates in your inbox, mostly small, 0–3/week depending on coffee dosage. More info.

It’s been 168 years since Oliver Wendell Holmes wrote “Homeopathy and Its Kindred Delusions.”45 It’s sad that those delusions still persist this long after his masterful debunking. The homeopathy Energizer Bunny is still marching along, banging its drum and trying to drown out the voices of reason. It is still necessary to keep explaining why homeopathic theory is incompatible with known science and to point out that remedies like Traumeel have been shown not to work, as Paul Ingraham does eloquently in this well-researched article.

Dr. Harriet Hall, The SkepDoc, and author of Women aren’t supposed to fly: the memoirs of a female flight surgeon46

Homeopathy resources

Science-Based Medicine publishes a good collection of critical articles and resources about homeopathy. A great place to start is Dr. Steven Novella’s commentary on the FTC’s historic decision to finally crack down on homeopathy products.

Homeowatch is entirely devoted to debunking homeopathy.

EBM-First.com has a good collection of skeptical reading recommendations about homeopathy.

The Australian government’s “Statement on Homeopathy” is the result of an exhaustive review of homeopathy, concluding that “there are no health conditions for which there is reliable evidence that homeopathy is effective.” Several related documents go into much greater detail, especially: Evidence on the effectiveness of homeopathy for treating health conditions (PDF, 392KB).

The above are some of the most strongly anti-homeopathy sources available, created by physicians and scientists. It is not my impression that they are mindlessly dismissive of homeopathy: they seem to have put a great deal of thought into their criticisms, and the quality of writing is generally excellent. They impress me.

Wikipedia has a thorough article about homeopathy. It’s not neutral — homeopaths probably dislike it. Still, it seems well-written and it’s heavily referenced, and you can’t beat that as a starting point. It provides a lot of well-sourced information.

The National Center for Complementary and Integrative Health is a curious case, as described above. The organization has a mandate and huge budget to prove that homeopathy works, but hasn’t been able to do it. Their information page on homeopathy reads with a friendly-to-homeopathy tone, as though it’s a rather good idea … but openly acknowledges that there is still no evidence that it works (previously quoted), even after many years of well-funded attempts to prove it. An interesting and unusual source.

The National Center For Homeopathy disagrees with NCCIH, and claims that “there are literally hundreds of high quality, published basic science, pre-clinical and clinical studies showing that homeopathy works.” They publish a bibliography and a few articles. There are numerous other homeopathy associations with similar websites.

I have had trouble finding good quality sources of information promoting homeopathy. Information presented by homeopaths is often of such poor quality that citing it is often no favour to their cause. Consider the case of a recent YouTube video by homeopath Charlene Werner: she earnestly makes a case for homeopathy on the basis of a string of appalling misunderstandings of physics. If you know nothing about physics, trust me … neither does she. See Dr. David Gorski’s pained analysis.

I invite anyone reading this to contact me and supply examples of better quality resources about homeopathy. I will happily publish a link to any such source. Although I have received many comments, positive and negative, about this article, no one has yet suggested a single better information resource about homeopathy.

And still more!

•  Homeopathy: Cure or Con? Part 1 on YouTube.com. See also: Original CBC show page. Superb investigative journalism from Canada’s CBC television, from a consumer advocacy point of view. A massive homeopathy overdose demonstration; a homeopath is recorded promising to cure stage 1 breast cancer (such a claim would be condemned by homeopathic professional associations); a chemistry lab proves that active ingredients cannot be identified in homeopathic preparations; and a mom who vaccinates her kids with homeopathy admits that she is a surprised by the lack of active ingredients.
•  That Mitchell and Webb Look: Homeopathic A&E on YouTube.com. Totally hilarious.
• “Malaria advice ‘risks lives’: Some high street homeopaths claim they can prevent malaria, a Newsnight investigation has found,” Meirion Jones, News.BBC.co.uk. Secret filming revealed homeopaths were claiming their preparations could be used instead of anti-malarial drugs to protect travellers in high risk areas such as sub-saharan Africa.
• A 19-page comic strip about homeopathy.
• “Three men make a tiger: Why physics precludes homeopathy from having any real effect,” David Robert Grimes, 3menmakeatiger.blogspot.co.uk. One of the best one-stop-shop summaries I’ve found of why homeopathy’s central claims are scientifically untenable. (A shame about the eye-watering color scheme! But the writing is good.)
• Slipping through the Cracks: Health Canada, Traumeel, and Homeopathy, a more directly critical article by Canadian pharmacist Scott Gavura, and occupational therapist Kim Hébert.

Darryl Cunningham’s entertaining illustrated tour of homeopathy.

PainSci guides to some other kinds of pain medications:

• The Science of Pain-Killers — A user’s guide to over-the-counter analgesics like acetaminophen, ibuprofen, and more
• Voltaren Gel: Does It Work? — The science of the topical pain-killers, which can be effective without dosing your entire system
• Vitamins, Minerals & Supplements for Pain & Healing — Critical analysis of most popular “nutraceuticals” — food-like pseudo-medicines taken for medicinal purposes, especially glucosamine and creatine, mostly as they relate to pain, arthritis, and recovery from exercise and injury
• Does Epsom Salt Work? — The science and mythology of Epsom salt bathing for recovery from muscle pain, soreness, or injury
• Marijuana for Pain — The hype versus the science! What does the evidence actually show about cannabis and chronic pain?
• Vitamin D for Pain — Is it safe and reasonable for chronic pain patients to take higher doses of Vitamin D? And just how high is safe?

Acknowledgements

Quite a few of people have been directly and indirectly helpful in creating this article. Special thanks to Dr. Rob Tarzwell, an indefatigable source of medical expertise (I’m not sure I’ve ever asked Dr. Tarzwell for help with anything without getting ten times more than I expected, or thought was even possible). Thanks to Nienke van Houten, Ph.D., lecturer, Faculty of Health Sciences (Simon Fraser University), for invaluable assistance obtaining some obscure research materials; thanks to Scott Gavura, skeptical pharmacist, both for contributing expertise to this article and promoting it on his popular blog. Thanks to Dr. Harriet Hall for encouragement, moral support and many funny critical thinking lessons in our correspondence over the years.

Notes

1. In 2015, the “Traumeel” brand was retired in the United States and Canada after the sale of the manufacturer’s North American subsidiaries. The new owner, MediNatura, has replaced it with a similar product called “T-Relief.”
2. The rules for capitalizing and italicizing plants’ names are not exactly cut and dried. Having precisely introduced it as Arnica montana (“mountain arnica”) I’m going to stick to plain and simple “arnica” for the rest of the article.
3. Ernst E, Pittler MH. Efficacy of homeopathic arnica: a systematic review of placebo-controlled clinical trials. Arch Surg. 1998 Nov;133(11):1187–90. PubMed 9820349 ❐ PainSci Bibliography 55782 ❐ “The Homoeopathic Pharmacy lists as its first indication "trauma" and even recommends it as a first-aid treatment.”
4. The various “laws” of homeopathy are described by nearly every homeopath with an online presence. Here is a typical example. Exact definitions vary significantly, but there are strong common themes. The Law of Infinitesimals is one of several that is routinely mentioned. The hypothesis that less is more is discussed more below.
5. ConsumerReport.org [Internet]. FDA: Stop using three Zicam cold remedies; 2009 Jun 16 [cited 19 Feb 18]. PainSci Bibliography 55520 ❐

   Original FDA press release no longer published, but the official documents are still available.

6. Homeopathy-help.net [now defunct] explained that “The 6 potency is the lowest strength generally available and works very gently.” (Most arnica is prepared at a “3” potency, as described below.) Homeopath Singh Veet wrote (source defunct): “Experience has trained homeopathic practitioners that in most circumstances the smaller the dose, the more powerful it is.” Similar statements are easy to find, so it’s clear that many homeopaths equate potency with strength.

   However, other sources contradict this, explaining that “potency” does not actually mean “strength.” What it does mean is hard to be sure of. Some say that low potencies are for localized problems, while high potencies are for system-wide problems. Other sources describe low potencies as being appropriate for “minor” problems, though, which sounds more like the “strength” definition. Homeopath Gina Tyler describes 9 considerations in choosing the potency, including vague and subjective considerations like “the susceptibility of the person” and their constitution, temperament and habits! So, a more diluted preparation is always considered a more potent preparation, but the definition of “potency” seems to be rather flexible.

7. “X” is not “times” as in multiplication, but the Roman numeral for 10, indicating orders of magnitude of dilution: so every 1X is another 10% dilution. (Why use Roman numerals? Who knows. It’s arbitrary. 3D is an alternate, identical notation — it means exactly the same as 3X.) “C,” the Roman numeral 100, is used to denote a 1% dilution (equal to 2X). Homeopathic remedies are prepared by a series of a dilutions and mixings (succussions). The process, called “potentization,” goes like this: dilute the ingredient to 10% of full concentration, mix vigorously (succussion), and repeat. Each cycle is another 1X. Thus the three cycles of a 3X remedy result in an active ingredient 1000 (10³) times less concentrated than the original: a 10^th of a 10^th of a 10^th.
8. A microgram (µg) is a really, really small amount: one millionth of a gram! Let’s round up the dosage from 7.5µg to 10µg — how much is that? Let’s put it in perspective: even though lead is mind-bogglingly poisonous, one of the most poisonous things in existence, so poisonous that even a single molecule of the stuff is considered harmful in theory, you would not notice 10µg of lead. The measurable effects of lead poisoning don’t start until you’ve got at least 10–100µg/litre of blood, so that would be 50–500µg total in an average adult male … and at that level the effects would still be subtle. So 10µg of arnica (and that was rounding up about 30%) is an incredibly small amount, and arnica is definitely not as potent as lead, and even 10 times as much lead is considered safe.

9. The .75g of arnica on the tube is diluted to “3X,” a.k.a. 1000 times less actual arnica than .75g, so divide .75 by 1000 to get the amount of actual arnica. So that’s .75g diluted arnica ÷ 1000 = 0.00075g or 750µg (micrograms) of actual arnica per tube.

   How about per dose? Assuming that a dose is about 1 gram, there are about 100 doses in a tube, so the amount of arnica in a dose is 1% of the arnica in the tube. Therefore: .00075g actual arnica ÷ 100 = 0.0000075g or 7.5µg of actual arnica per dose.

10. Some extreme substances — plutonium, say — are certainly potent even when quite diluted. But most ordinary chemical compounds — chemicals that have no dramatic biological impact even at full concentration — will be rendered impotent even by modest dilution.
11. Rao N, Spiller HA, Hodges NL, et al. An Increase in Dietary Supplement Exposures Reported to US Poison Control Centers. J Med Toxicol. 2017 Jul. PubMed 28741126 ❐ Homeopathic remedies were the most guilty product category contributing to a 50% rise in supplement-related calls to poison centers from 2005 to 2012. This is caused by a huge and totally unregulated industry selling to legions of credulous people who are knee-jerk cynical about Big Pharma but never question the scruples of supplement sellers. What could possibly go wrong? Homeopathic arnica creams are probably less likely to be contaminated, or cause serious harm if they are, but it’s still food for thought.
12. A Vogel Atrogel Arnica Gel, as presented on https://bhma.info/atrogel (and many other places). Accessed 2023-05-02.
13. Above we calculated the amount of arnica in a 1g dose to be roughly 7.5µg — just shy of 10 millionths of a gram. That super tiny number divides into 500 milligrams rather a lot of times: 66,666.66 to be more exact.
14. Totonchi A, Guyuron B. A randomized, controlled comparison between arnica and steroids in the management of postrhinoplasty ecchymosis and edema. Plast Reconstr Surg. 2007 Jul;120(1):271–4. PubMed 17572575 ❐

    This is a test of arnica and corticosteroids for swelling and bruising after 48 nose jobs. None of them worked all that well. There was a little less swelling on the 2nd day with both arnica and corticosteroids, but it was a small difference that went away by the 8th day. Arnica had no effect on bruising at all.

15. Voltaren® is an anti-inflammatory cream (topical diclofenac), proven effective and safe for osteoarthritis, and it probably helps some other kinds of pain as well. Dosing is minimal, targeted, and far safer than the oral version. For more information, see Voltaren Gel: Does It Work? The science of the topical pain-killers, which can be effective without dosing your entire system.
16. Derry S, Moore RA, Gaskell H, McIntyre M, Wiffen PJ. Topical NSAIDs for acute musculoskeletal pain in adults. Cochrane Database Syst Rev. 2015;6:CD007402. PubMed 26068955 ❐
17. But many other scientists are deeply irritated by NCCIH. For instance, Dr. David Gorski, a cancer researcher, writes that NCCIH is “a grossly irresponsible use of taxpayer money not to prioritize funding for projects that have hypotheses behind them that have a reasonable chance of being true. Scarce NIH funds should not be for projects that have as their basis hypotheses that are outlandishly implausible from a scientific standpoint.” See The National Center for Complementary and Integrative Health (NCCIH): Your tax dollars hard at work.
18. The skeptical comedian Tim Minchin sassily explains this problem, “It somehow forgets all the poo it’s had in it!” See Storm 9:37. More technically stated by Three men make a tiger, “This is kind of disturbing when you think about it: all water on earth is a closed system and constantly gets recycled; this would mean that water could have an unduly strong recollection of all the human and animal waste it has come in contact with. However, homeopaths seem to think water has a somewhat more selective memory than this.” Grimes has an excellent explanation of the water-memory issue.
19. You can’t just say, as homeopaths often have, that the active ingredient was remembered “better” because it was recently in a high concentration. The dilution process doesn’t just replace the ingredient, but the water around it — after many dilutions, not only is there none of the original ingredient left, there’s also none of the original water left, either.

20. “Tendinitis” versus “tendonitis”: Both spellings are acceptable these days, but the first is the more legitimate, while the second is just an old misspelling that has become acceptable only through popular use, which is a thing that happens in English. The word is based on the Latin “tendo” which has a genitive singular form of tendinis, and a combining form that is therefore tendin. (Source: Stedmans Electronic Medical Dictionary.)

    “Tendinitis” vs “tendinopathy”: Both are acceptable labels for ticked off tendons. Tendinopathy (and tendinosis) are often used to avoid the implication of inflammation that is baked into the term tendinitis, because the condition involves no signs of gross, acute inflammation. However, recent research has shown that inflammation is actually there, it’s just not obvious. So tendinitis remains a fair label, and much more familiar to patients to boot.

21. Although straightforward enough in the acute stage, chronic tendinitis degenerates into something more complex and puzzling. Inflammation ceases to be a dominant feature of the condition. Instead of “tendinitis,” it is better called “tendinosis” — a less specific “disease” or “condition” of the tendon. The pathophysiology of tendinosis is not well understood. Although painful, the tendon is not inflamed.
22. Khan KM, Cook JL, Taunton JE, Bonar F. Overuse tendinosis, not tendinitis, part 1: a new paradigm for a difficult clinical problem (part 1). Phys Sportsmed. 2000;28(5):38–48. PubMed 20086639 ❐

    From the abstract: “If physicians acknowledge that overuse tendinopathies are due to tendinosis, as distinct from tendinitis, they must modify patient management … ”

23. The physiology of DOMS remains a good medical puzzle. One researcher described DOMS research as so deeply frustrating — difficult and poorly funded — that he abandoned it. See A Deep Dive into Delayed-Onset Muscle Soreness: The biology & treatment of “muscle fever,” the deep muscle soreness that surges 24-48 hours after an unfamiliar workout intensity.

24. The colorful history of medicine and quackery is overflowing with people who “swore by” treatments that were bizarre and dangerous. Bloodletting was popular almost until the 20th century, despite being relentlessly harmful. Some of the most lethal “cures” in history were inspired by the discovery of radiation. People happily drank metals like mercury and silver. Even drinking urine had near fad status for a while! They tried to purge disease with sulfuric acid, and stimulate their vitality (and virility) with powerful electric shocks. Women were sold Lysol as a douche … and women actually went along with it for a while. Voluntary lobotomy may be the craziest of them all: it was a popular treatment for all kinds of psychiatric disorders, and at least fifty thousand people volunteered to have their brains lanced.

    All of these terrible treatments, and many more obscure examples, had many fans and enthusiastic testimonials. People paid for them, believed in them, loved them, swore by them — that is how misleading testimonials can be. People believe what they want to believe.

25. One of the most common questions that patients ask health professionals is, “How long will this take to heal?” Most patients aren’t the least bit knowledgeable about what constitutes a normal healing time, and sensibly defer to clinicians who have seen hundreds or even thousands of examples — who all know that healing time varies wildly depending on countless known and unknown variables. We often recover faster or slower than we expected to for reasons no one can ever know. We also seem to recover faster or slower, depending on which psychological “goggles” we had on at the time.
26. Gaertner K, Baumgartner S, Walach H. Is Homeopathic Arnica Effective for Postoperative Recovery? A Meta-analysis of Placebo-Controlled and Active Comparator Trials. Front Surg. 2021;8:680930. PubMed 34977136 ❐ PainSci Bibliography 52003 ❐

27. Dr. Edzard Ernst points out several flaws beyond the obvious bias, including but not limited to the inclusion of unpublished trials, trials published in junk journals, trials of Arnica in non-homeopathic concentrations, trials that weren’t actually trials, and trials with endpoints that weren’t mentioned in the review. Remember that small effect sizes extracted from junky little studies are probably all illusory. I doubt there has ever been such a result that was eventually replicated by a high quality trial.

    This paper a classic example of garbage-in-garbage-out junk science, and the conclusion is weak sauce even with intense bias-powered p-hacking.

28. Gartlehner G, Emprechtinger R, Hackl M, et al. Assessing the magnitude of reporting bias in trials of homeopathy: a cross-sectional study and meta-analysis. BMJ Evid Based Med. 2022 Mar. PubMed 35292534 ❐

    This is some meta-science (science about science) confirming that homeopathy research kinda sucks. Researchers studying homeopathy don’t register their trials often enough, many registered trials are never published (probably because the results were negative), and even when they are registered and published they often change the primary outcomes (moving the goalposts to get a “score”). “This likely affects the validity of the body of evidence of homeopathic literature and may overestimate the true treatment effect of homeopathic remedies.”

29. Ernst E, Pittler MH. Efficacy of homeopathic arnica: a systematic review of placebo-controlled clinical trials. Arch Surg. 1998 Nov;133(11):1187–90. PubMed 9820349 ❐ PainSci Bibliography 55782 ❐ “ … burdened with a multitude of methodological limitations. Small sample size and lack of test statistics are frequent and obvious ones.”
30. “ … many of these trials have methodological limitations that make the findings unreliable.” Stevinson C, Devaraj VS, Fountain-Barber A, Hawkins S, Ernst E. Homeopathic arnica for prevention of pain and bruising: randomized placebo-controlled trial in hand surgery. Journal of the Royal Society of Medicine. 2003 Feb;96(2):60–5. PubMed 12562974 ❐ PainSci Bibliography 55776 ❐
31. And even from NCCIH (remember, supposedly sort of pro-homeopathy): “ … many of the studies have been flawed.”
32. Shang A, Huwiler-Müntener K, Nartey L, et al. Are the clinical effects of homeopathy placebo effects? Comparative study of placebo-controlled trials of homeopathy and allopathy. Lancet. 2005;366(9487):726–32. PubMed 16125589 ❐

    Homeopathy is a widely used treatment modality based on highly implausible, pre-scientific ideas. Some studies seem to support its use, but may be the result of researcher bias, which can skew the results of smaller and less rigorous studies in many ways, producing the answers that researchers are hoping for. The purpose of this review was to identify and review homeopathy studies with a lower risk of this. And how they did that was rather clever: they identified 110 randomized, controlled trials of homeopathy, and then compared them to 110 studies of medical treatments for the same conditions.

    As expected, lower quality studies of both types tended to produced more “positive” results. In other words, there was probably bias on both sides of the alternative/mainstream fence. But when the analysis of results was restricted to only the higher quality results on each side, something important was revealed: there was only “weak evidence for a specific effect of homoeopathic remedies, but strong evidence for specific effects of conventional interventions.”

    In other words, only crappy studies show that homeopathy works. Good quality studies show that it doesn’t … but modern medicine does.

33. Robertson A, Suryanarayanan R, Banerjee A. Homeopathic Arnica montana for post-tonsillectomy analgesia: a randomised placebo control trial. Homeopathy. 2007 Jan;96(1):17–21. PubMed 17227743 ❐

    This test of homeopathic (extremely dilute) Arnica montana was conducted by homeopaths and published in a journal for homeopaths, so the risk of the results being bent by bias is as high as it gets. But … it is the right kind of study, on paper at least: a randomized double blind trial of 190 patients over the age of 18 undergoing a tonsillectomy. Subject received either Arnica 30c or a placebo for two weeks.

    The Arnica-getters had a slightly larger drop in pain score (28.3 points, on a 100-scale) compared to the placebo group (23.8), so the researchers concluded that this was “a small, but statistically significant, decrease in pain scores compared to placebo.” Sounds okay, right? The right kind of test with modest but clearly positive results?

    The effect size is all that matters, and it’s trivial. Emphasizing statistically significance is a red herring. They neglected to mention the well-known principle that statistical significance in tests of highly implausible claims is meaningless (see Pandolfi). So “small but statistically significant” in this context basically means they found nothing at all.

34. Ioannidis J. Why Most Published Research Findings Are False. PLoS Medicine. 2005 08;2(8):e124. PainSci Bibliography 55463 ❐

    This intensely intellectual paper — it’s hopelessly nerdy — became one of the most downloaded articles in the history of the Public Library of Science and was described by the Boston Globe as an instant classic. Despite the title, the paper does not, in fact, say that “science is wrong,” but something much less sinister: that it should take rather a lot of good quality and convergent scientific evidence before we can be reasonably sure of something, and he presents good evidence that a lot of so-called conclusions are premature, not as “ready for prime time” as we would hope. This is not the least bit surprising to good scientists, who never claimed in the first place that their results are infallible or that their conclusions are “true.”

    I go into much more detail here: Ioannidis: Making Medical Science Look Bad Since 2005.

35. And in journals that are probably strongly biased, like Homeopathy and Complementary Therapies in Medicine … and especially Heel’s own journals, BioLogical Therapy and Biomedical Therapy! Many modern journals exist essential for the purpose of promoting alternative medicine, their reviewers are all alternative medicine practitioners, and they would never publish a negative study.
36. Ernst E, Pittler MH. Efficacy of homeopathic arnica: a systematic review of placebo-controlled clinical trials. Arch Surg. 1998 Nov;133(11):1187–90. PubMed 9820349 ❐ PainSci Bibliography 55782 ❐
37. Here are seven of the eight: Campbell, Savage, Hildebrandt, Pinsent, Kaziro, Tveiten, Gibson.
38. Oberbaum M, Yaniv I, Ben-Gal Y, et al. A randomized, controlled clinical trial of the homeopathic medication TRAUMEEL S in the treatment of chemotherapy-induced stomatitis in children undergoing stem cell transplantation. Cancer. 2001 Aug;92(3):684–90. PubMed 11505416 ❐ This is a widely cited study of homeopathy, because its good design and clearly positive results are a rarity. It was a randomized, controlled and blinded trial of Traumeel S for kids with cancer, who get painfully inflamed mouths during stem cell transplants. The results seem happy — statistically significant and clinically meaningful. However, with a small sample size of only 15, even a small mistake in data entry, or a tough choice about how to represent some tricky data, could queer the statistical significance. Science is full of such opportunities to be wrong. It happens even to the best scientists. In any event, “statistically significant” only means that the results probably weren’t a fluke. And this is why it always takes lots of good quality, bigger trials to come anywhere close to “proving” anything.
39. Stevinson C, Devaraj VS, Fountain-Barber A, Hawkins S, Ernst E. Homeopathic arnica for prevention of pain and bruising: randomized placebo-controlled trial in hand surgery. Journal of the Royal Society of Medicine. 2003 Feb;96(2):60–5. PubMed 12562974 ❐ PainSci Bibliography 55776 ❐
40. Brinkhaus B, Wilkens JM, Lüdtke R, et al. Homeopathic arnica therapy in patients receiving knee surgery: results of three randomised double-blind trials. Complement Ther Med. 2006 Dec;14(4):237–46. PubMed 17105693 ❐ In this study, surgeons found a small positive effect in 57 patients, and emphasized that in their conclusions … instead of emphasizing their best data, which showed no benefit in 262 other patients. Averaged out, it would have been a strongly negative study — good evidence of absence. But they spun it as a positive study. See the full bibliographic record for details.
41. In the other, the authors failed to emphasize that patients actually had more pain with homeopathic arnica, and instead painted the results as positive because of a barely measurable improvement in bruising. This study (Seeley) will be dissected in detail below as a classic example of “positive” homeopathic research.
42. Seeley BM, Denton AB, Ahn MS, Maas CS. Effect of homeopathic Arnica montana on bruising in face-lifts: results of a randomized, double-blind, placebo-controlled clinical trial. Archives of Facial Plastic Surgery. 2006;8(1):54–9. PubMed 16415448 ❐ PainSci Bibliography 55779 ❐
43. Using custom computer software, “ … the exact density of pixels per square centimetre for that image was calculated. This density was then extrapolated to the study area, which had a known number of pixels, and thus an exact area of ecchymosis was calculated, despite its markedly irregular borders.”
44. In science, such a tiny study is the sound of one hand clapping. This experiment was as small as they come — 14 patients getting SINECCH compared to 15 getting a placebo. These are such small groups that an abnormality in a single patient could skew the stats. Even the most dramatic results in such a study could easily be a fluke, and would have to be confirmed by multiple larger studies. That’s how science works.
45. Ebooks.adelaide.edu.au [Internet]. Holmes OW. Homeopathy and Its Kindred Delusions: Two lectures delivered before the Boston Society for the Diffusion of Useful Knowledge; 1842 [cited 12 Feb 19]. PainSci Bibliography 55672 ❐

    This is a nicely done web-reprint of Holmes’ classic 1842 essay on homeopathy — a critical analysis of the subject that has easily stood the test of time, its reasoning validated in countless ways by experimental evidence ever since.

46. Hall H. Women aren’t supposed to fly: the memoirs of a female flight surgeon. IUniverse; 2008.