The acupuncture paradox

The scientific consensus appears to be that, to the extent that acupuncture makes people feel better, it is through relaxing the patient, also the acupuncturist might help in other ways, encouraging the patient to focus on his or her lifestyle.

A friend recommended an acupuncturist to me awhile ago and I told her the above line, to which she replied: No, I don’t feel at all relaxed when I go to the acupuncturist. Those needles really hurt!

I don’t know anything about this, but one thing I do know is that whenever I discuss the topic with a Chinese friend, they assure me that acupuncture is real. Real real. Not “yeah, it works by calming people” real or “patients respond to a doctor who actually cares about them” real. Real real. The needles, the special places to put the needles, the whole thing. I haven’t had a long discussion on this, but my impression is that Chinese people think of acupuncture as working in the same way that we think of TV’s or cars or refrigerators: even if we don’t know the details, we trust the basic idea.

Anyway, I don’t know what to make of this. The scientific studies finding no effect of acupuncture needles are plausible to me—but if they’re so plausible, how come none of my Chinese friends seem to be convinced?

Maybe I just haven’t asked enough people, and now they’ll come up to me and say they don’t think that acupuncture works. But I don’t think so.

P.S. Just to clarify: my question here is not whether acupuncture could work (possibly through some backdoor mechanism like the needles stimulating your body in some useful way, or whatever) but on the evidence of how much it does work. As noted, I think the overwhelming impression among my Chinese friends—statisticians included—is that it does work, and not merely through some vague calming effect. But this would seem to contradict the research, so I don’t know what to think.

56 thoughts on “The acupuncture paradox

  1. We humans are great at compartmentalization. I guess the Chinese people you’ve asked have grown up with acupuncture as a natural, self-evident part of their lives, and everybody around them take it as given it works, so they never think to apply their scientific mindset to it, and may even strongly resist it. It’s exactly the same mechanism that lets somebody be a scientist and religious at the same time.

  2. Wow.. Do you really think scientific experiments and random opinions of chinese people should be given the same epistemological weight? Obviously the simple answer is that the ancient east asian wisdom is just wrong, in this case as well as that of dragons, fan death, reincarnation, and kung fu stopping bullets.

    • Darwinspinning:

      I’m not talking about random opinions. I’m talking about the opinions of trained statisticians. I’m not saying that a statistician has to be right about such things, I just find it interesting that they seem to have no doubt about the efficacy of accupuncture (and not merely that it works through some calming effect, but that it really matters where they put the needles etc.

      No, I don’t think my friends believe in dragons etc., any more than I as a white person believe in the Loch Ness monster, bigfoot, etc.

      • This is no surprise. Humans have a great capacity for holding multiple contradictory beliefs. Confirmation bias plays a role here and it has been well documented that stat and research expertise has little to no impact on recognizing the bias in our own thoughts and beliefs.

      • Whatever the opinions of these trained statisticians is, it does not appear (from what you say) to be based on a statistical analysis of experimental data (treatment vs. placebo). So their status as trained statisticians seems to me to be irrelevant to their opinions.

        • Bill:

          That’s right. The key point is that they’re my friends and I respect their opinions on science in general, not that they have particular expertise in evaluating pain-relief treatments.

        • Andrew, good comment. We all have our blind spots, even if we are scientists/statisticians. This is indeed a warning that all of us have to examine our personal biases carefully, and not mix them up with our analysis. And, we should not use our special standing as professionals to promote causes that we have no special standing to promote. (There is all too much of this today, e.g., engineers and other scientists outside the field posing as climate scientists, to give one egregious example)

          I don’t know what to do about the friendship angle, except to respectfully point out this problem.

        • Bill:

          It’s not just that one or two of my Chinese friends think acupuncture works. It’s that all of them seem to think so, including people who I’d think would respect statistical evidence.

        • Andrew,

          If in the past you have found that your Chinese statistician friends have provided reliable opinions on science in general, then of course you are right to seek their opinion on the efficacy of acupuncture. But if the relevant scientific data overwhelming support the thesis that acupuncture is not efficacious (beyond a minimal placebo effect, of course), and if your Chinese statistician friends are aware of the data and yet maintain in their support for the efficacy of acupuncture, then you now have some evidence against the reliability of their opinions on science in general, and good evidence against the reliability of their opinions on scientific pain management.

  3. I think the reason intelligent people believe that Acupuncture works hinges on the possibility that there are some processes in the body that are still not measurable. Some people feel pain in places that cannot be confirmed as a pathology site. Perhaps there is a way that tension is stored in a location, which cannot be confirmed by independent observation, but that is relieved when stimulated with acupuncture needles.

    I suppose we could also bring in the growing strength of the placebo effect in medical trials. What is the process that is making placebos more effective? Measurement error? Or some thing the body is capable of doing that science has not yet been able to measure/understand?

  4. As someone who’s trained a bit in Chinese martial systems and who’s received acupuncture and tui na treatments to seemingly good effect, I’d like to add my compartmentalized, pre-scientific voice to those who find the benefits of acupuncture plausible. For a less anecdotal voice, I recommend that of Ted Kaptchuk amongst others (e.g.,

    (Also, as impressed as the Science-Based Medicine writers are with their own intelligence, I don’t think the case they make is nearly as strong as they seem to believe; the actual analyses they link to are old and thin, included as afterthoughts to prop up a case they apparently think hardly needs making.)

  5. I don’t know which studies you are thinking about so they may address the concern that follows:

    A few years back I was looking casually at various alternative medicine studies and was concerned about how they go about selecting the practitioner they compare the placebo with (e.g., which acupuncturist to include in the treatment condition).

    Alternative medicine has low barriers of entry and hence there must be many fakes out there. Say 20% of acupuncturist (or whatever you are testing) know what they are doing (compare to, say, 75% of Rheumatologists?). Customers may shop around until they find a member of the 20%, paying once to a bunch of 80%ers, and then finally repeatedly to the 20%er they found (note that this makes the fakes survive in the market).

    The question is, if you want to study the benefits of Acupunture, you need to identify that 20%, how do you do that as a researcher?

    I guess one easy to answer research question is: “if you choose one acupuncturist at random, how likely are you to get well”, a more interesting one is “if you find a “good” acupuncturist, how likely are you to get well”

    I don’t have a horse in the race, just an observation.

    • If you’re testing whether acupuncture *could* work, you need to find the best acupuncturists and test them, but if you’re testing whether acupuncture *does* work (i.e. whether it’s likely to work if you walk into an acupuncturist and say “help me”), then it’s okay to take a random sample of acupuncturists who haven’t yet gone out of business.

      Putting aside questions of statistical test design, I think a field where the frauds aren’t obviously performing at an inferior level to the non-frauds is a field where the non-frauds aren’t doing anything special.

  6. I should point out that people responding to someone who actually cares about them is as real as it gets. Western medicine treats human emotions as airy irrelevance rather than the primitive animal system that they are. I’m aware that there are increasingly papers showing ‘remarkable’ results showing the idiocy of dualist thinking, how people who are loved live longer, etc, but it’ll be decades before Buddhism is routinely taught to our children.

    Acupuncture as a Western-style hack on a presumed robotic system? I don’t know. Reflexology impresses me though – and that’s just feet!

  7. Surely the fact that your Chinese friends are trained statisticians is irrelevant: unless they have run double blind trials themselves they are still offering anecdotal evidence, and are open to conscious and unconscious bias. Believe the studies. To extend a favourite phrase of mine: the plural of “anecdote” is not “data,” even when the anecdotes are reported by experts.

  8. Excuse me but which studies are you thinking about? How to design control group for such studies? I’m really curious….BTW, in my personal opinion it does work – and I did not grow up around this or any other oriental medical tradition; it was something a (luckily) found later in my life.

  9. I don’t know to what extent the problem only concerns alt-med (acupuncture only being one example I take it, we could probably find other alt-med supported by reasonable people, even statisticians).

    As you are probably aware of, some researchers (Ioannidis in particular) have claimed that most research findings in medicine are likely to be wrong. It also seems that as late as the 70s, only a very small percentage of prescribed treatments had no proven efficacy (I don’t know how high that % woud be today). And clearly (I think), most of us have no idea of how whatever treatment we receive works.

    So maybe Chinese people have simply learned that they could trust acupuncture and that it made sense in the same way as we’ve learned to trust Western medicine. We don’t trust medicine because we’ve made specific enquiries or know how it works (again, in the vast majority of cases). We just do.

  10. I think it’s a combination of things:
    Number 1, “science-based” medicine is terrible in dealing with chronic pain and illnesses. I have seen people go to various specialists with chronic pain and get widely differing treatments, most of them w/o discernible effects. So people are rightly skeptical of Western medicine’s claim to understand much of what’s actually going on in the body wrt chronic pain.
    The other thing is that acupuncture does work in the sense that placebo acupuncture is _hugely_ effective – see e.g. – almost 50% of participants get pain relieve as opposed to 25% with traditional treatment. So yes, the special places may not matter much, but if you have chronic pain you’d be foolish not to undergo acupuncture.

  11. Like Keith brings up, the comedy is the many ways western evidence based science has shown that western medicine could be practically improved, but isn’t implemented for organizational, institutional, and larger social reasons.

    I agree chinese consensus should fold to best practices research -but our conceit of what best practices research is should also fold to actual best practices research. I suspect accupuncture is bullshit, but I won’t have my worldview shattered if it turns out Western arrogance skewed supposed empirical inquiry into accupuncture.

  12. I’m not aware of studies that attribute accupuncture directly to relaxation. But, while the needles might be painful there might be some relaxation involved in having someone take your concerns seriously, spend time with you, and carefully and painstakingly actively attempt to cure you. Furthermore, there would be a degree of self generated relaxation in order to deal with the pain. In fact, that’s what the accupuncturist tells you to do. It’s just like getting an injection or blood samples taken at the doctor’s office. There’s a little anxiety with the anticipation but in order to deal with it you need to make yourself relax.

    I do remember seeing a talk a few years ago at a psychiatric conference and they mentioned that if medicine tastes bad it can enhance the placebo effect. Some people attribute this to associations with real bad tasting medicines that had actual effects. I think it’s probably more likely that the fundamental belief in the effectiveness of the medication is enhanced by the bad flavour and that belief is well established to be able to enhance the placebo effect. It’s entirely plausible that the pain from the needles has effects very similar to making medicine taste bad. (BTW, market research from the Buckley’s found that people though it more effective when it tasted bad)

    As a side note, it seems many dismiss placebo affects as merely something imagined. It’s not that simple. For example, a pain treatment may have no effect on the cause of pain or the perception of it. However, if you believe it does your brain actually generates chemicals that actually reduce the feeling of the pain. It’s not just an imagined effect. (which is one of the many reasons lots of physicians are looking into using placebos to treat people given it’s effectiveness is often very high, sometimes in the 30% range, especially for things like migraines)

  13. You fail to mention the biggest stumblingblock is the fact that acupuncture, by it’s very definition, traditional Chinese medical approaches include SHEN and JING, not just QI (read: Mind/ Emotions and Constitutional/ genetic tendencies not just the physical body) and acupuncture, like surgery and physical therapy, cannot be successfully studied and validated using the randomized double-blind placebo-controlled clinical trial model…it WILL NOT FIT! NO study to date has shown “Verum” acupuncture truly superior to “Sham” acupuncture because the Sham/ control groups are not physiologically OR psychologically inactive, nor blindable. Do MORE homework before you post. Real, real. ;-)

    • Boyd – that makes no sense. Randomization takes care of different physiological and psychological behavior patters (which are going to be equally distributed in treatment and control groups) and at least the large German study I link to does test for blinding and finds that it did very well on that.
      And if you can’t test it with a trial and there aren’t any measurable causal paths, how would you know that verum acupuncture works better than sham? At that point it just becomes a religious believe in the wisdom of the elders.

  14. I have been receiving acupuncture treatments for nearly two years. I was initially skeptical being fully aware of the placebo effect. My acupuncturist who is American, Jewish and an MD has gone over the research. He says the placebo effect is a component of the relief, but not all of it. They have objective measurements by way of brain scans that show something objective happens. Moreover vets use acupuncture to treat pain in animals. It’s hard to imagine that animals have a placebo effect.

    My own experience.

    1. Very little pain. Once in a while a needle will hurt going in. Once in a while an inserted needle will start to hurt but that dies out. For the most part the experience is pain free and much less stressful than a trip to the dentist.

    2. I don’t find it particularly relaxing although I do fall asleep, but I would anyway after lying down in a quiet dark room.

    3. Sometimes significant relief of nerve root pressure pain and migraines. Let’s put the before treatment pain at 10. Then after treatment, I will go to a 5 or 6. In other words, some relief, but not spectacular. But worth doing.

    There are wild claims made for acupuncture that can’t be true. But for muscular-skeptical pain the evidence for efficacy seems good.

    • Of course brain scans show something is happening – you stick a needle in someone, it hurts, and so something changes in your brain. I was a participant in a study that did this – and the needle really hurt. (Apparently it was ‘high dose’ acupuncture, to maximize effect sizes), so the bit of my brain that said “Someone’s sticking something in my hand and it really ****ing hurts!” probably lit up like a Christmas tree, but the acupuncturists take this as evidence that “Something’s happening”.

  15. The linked-to article is written in first person and makes strong editorial claims, but has no visible (to my quick scan) author or credit. Who wrote the article? And / or, have you come to the same conclusions from reading the primary literature?

  16. Wikipedia has a pretty good write up on acupuncture giving various viewpoints. I don’t think we have a theory as yet as to why it works assuming it does work.

    Here’s some suggestions.

    1. Replace the metal needles with non-conducting needles, say ceramic and see it that makes a difference. In other words, do something to rule out an electric effect of some kind. Perhaps there’s a chemical explanation, so use an absolutely inert substance as a needle.

    2. Put the needles in the “wrong” places. The patient won’t know the difference and see how that affects pain control. The Chinese have a whole system for where to put the needles. Violate the system and see what happens.

    3. More experiments on animals and babies.

    4. Look for chemical changes in the body after treatment.

    I asked my acupuncturist about (1) and he said that no one has done that.

    My acupuncturist started out as a doctor. I think he got his MD degree at Stanford University Medical School. He told me he was dissatisfied with the limitations of conventional medicine on an intellectual level as a medical student.

    Let’s not kid ourselves. Mainstream medicine is far from fully scientific. A lot of it is individual judgment and often wrong. At best its a combination of science, art and experience. For years doctors laughed at the notion that ulcers are mostly an infection. Then the specific bacteria was found that’s responsible for most stomach ulcer conditions. Today you get scoped and the doctor takes a pinch out of the stomach lining and sends that off to pathology. If you have the infection, you get cured with an antibiotic. What was once a joke is now the regular practice. Back in the 1960s ulcers were a serious problem. People died from it. Today most everyone gets cured quickly.

    • “3. More experiments on animals and babies”

      I’d unironically love more experiments on babies. Not conscious yet, stock is relatively quickly replenishable -but you think people find sticking an electrode in a kitten repugnant … yikes. Here’s hoping you’re a sign that norms are shifting in a more rational direction.

    • 1 has in effect been done, not by using ceramic needles, but by using fake needles that do not penetrate.

      2 has also been done.

      To my knowledge, when compared to these, “real” treatment has no effect.

  17. I have have the some wonders about acupuncture for a long time. I must say I do use it myself, both as acupuncture (with needles) and as accupressure (just pressing the points with the fingers), for my chronic back pain. Do not tell me about pain killers they simply do not work on my. There is something funny when I get my accup treatments: the same day of the treatment the back pain increases a lot, i mean really a lot to the point of having problems to walk out of the clinic. The next day is worse. The day +2 everything is fine and even better than the first day. I agree with your friend. Having the needles does hurt, specially when the put electricity as well.

    I have not reviewed the medical lit but I think there some evidence of changes on signalling on the nerves. Which may eventually make sense because the needles are often bang on your nerves.

    I also know for many people does not work, but same happened with main-stream medical treatments.
    What I do know is that I have been unable to walk and after 4 treatments (every two day) I feel no pain at all, I can sleep through the night again and be happy :). At least for me the treatment effects is large enough as to go for the needles and the electricity again and again !!!


  18. Andrew,

    A couple of thoughts from the molecular and computational biology angle. Full disclosure: I have no opinion about the medical efficacy of acupuncture; I spend my days analyzing large biological datasets.

    There do seem to be measurable, quantifiable effects at the gene transcription level. I’ve seen several papers in PubMed that ran control and acupuncture treatment groups, hybridized mRNA to microarrays, and detected changes in gene expression.

    Of course, those observations don’t imply that the gene expression changes are clinically relevant, but they do support the hypothesis that something physiological is happening.

  19. Let’s talk about sports massage for a second. I don’t know if there is any clinical evidence about the efficacy of sports massage, but I am strongly inclined to believe that it enhances performance. My next-door neighbor is a world-class masseur! In fact, right at this moment he is in Europe to massage the U.S. National wrestling team. He also works on the women’s swim team, as well as a bunch of elite Track and Field athletes. They fly him around the world to massage them. Pro cyclists and tennis players (possibly among many other sports, I don’t know) have masseusses/masseurs in their entourage.

    Now, suppose some Chinese statistician were to ask me “why do you believe that massage helps? There’s no known performance-enhancing mechanism, and there’s no clinical evidence that it helps. These people should use acupuncture instead, since we know it works.” Well, I don’t know about those claims one way or the other — although I do know that the claim that massage enhances blood flow such that toxins are removed from the body more quickly is not true — but what can I say, I’m still inclined to think sports massage improves performance. Otherwise why would all of those athletes get painful massages so often?

    If someone shows me convincing evidence that sports massage does _not_ improve performance, I’ll allow myself to be convinced, but just having some schmoe “there’s no clinical evidence that it works” is not enough to override my fairly strong prior belief.

    I suspect the Chinese people you’re referring to have more or less the same state of belief about acupuncture that I do about sports massage: there’s plenty of anecdotal evidence that it works, so given the known difficulties in performing a clinical trial of it the lack of clinical evidence that it works is not very persuasive of the claim that it _doesn’t_ work.

    Also, in either case (massage or acupuncture), if it works then it’s easy to think of reasons why it would.

    Me, I think there’s an 85% chance that sports massage helps, and a 15% chance that acupuncture helps, but I could imagine someone with different experience or culture starting out with the percentages reversed.

  20. Being married to a Chinese lady, I’ve had occasion to speak with her about this (and other things) at length, and you are exactly right, it *is* real, for (a good deal of) them, and acupuncture isn’t the only thing. The partial resolution in my case is the growing awareness that scientific consensus isn’t all it’s cracked up to be; it’s more like “consensus”, which can vary both over time and location. And to those who’d like to laugh at the silly notions of those Chinese (the 1.3+ billion of them), ask a Western scientist/statistician/nutritionist/whatever whether or not eggs are good for you.

  21. I don’t often read the comments as I read the blog via google reader but this piqued my interest a bit… Must say I’m disappointed at the general anti-science sentiment expressed here. The evidence is plentiful and poor for acupuncture (1.3 billion Chinese think… what, that rhinoceros horn is good for an erection?!). Even the best studies that acupuncture apologists point to show no benefit of “real” acupuncture to sham acupuncture. Sure, maybe there’s magic, but then why bother with statistics?

  22. I can very much recommend ‘Quackcast’, a (funny) podcast specifically about alternative medicine. The opinions are based on the most recent literature, discussed in detail. There’s a list here:
    episode 7 & 8 deal with the basics of acupuncture, but there are regular updates as new papers come in (episodes 21, 27, 33 and more). There are also extensive references for each episode

  23. I have to say I’m shocked by the number of commenters on a statistics blog who are determined to favour anecdotal evidence over the extensive robust trials that have been conducted.

    @Jay Kerns: I don’t think it’s a case of “laughing at the silly notions of those Chinese”; and I think your phrasing there, hinting that those of us who deny acupunture are racist, is disingenuous. 1.3 billion Chinese can be wrong: this is just a version of the argument from authority fallacy. Minimally, the sum of any two of 1.6 million Muslims or 2.2 billion Christians or 1 billion Hindus (and for my money, the all three, plus the rest) can be wrong.

    @Zarkov: “Mainstream medicine is far from fully scientific … For years doctors laughed at the notion that ulcers are mostly an infection … Back in the 1960s ulcers were a serious problem. People died from it. Today most everyone gets cured quickly.”

    I’m sorry, but you have described precisely why mainstream medicine IS fully scientific. Because they studied the condition, saw evidence that their previous opinions were wrong, modified their opinions and the treatment according to that evidence, and as a result massively improved their treatment. That’s the glory of modern medicine, and it’s why you (and the other medicine deniers in the comments here) will live on average twice as long as your grandparents.

    • Nicely put: we are wrong about everything and best to adopt ways to quickly get less wrong.

      It the good opportunities to get less wrong, that science provides, that makes it better in the long run.

      At Friday’s dinner party, I made the mistake of commenting there was no evidence (yet) for homeopathy and immediately heard from those who believe it is as effective as a placebo and should be promoted…

      When I worked in clinical research the guess was that 1/3 of what had apparent evidence of working – did not work.

    • @Stuart Brown: whoa, beep, beep, beep. It’s not clear what your “it” means, and I’d just like to state re: my comment that denial of acupuncture doesn’t imply/hint/anything racist. (?!?!) I said “silly notions”, and I was speaking from years of my personal experience being one of those who used to be doing the laughing. BTW, there are several more I’ve encountered that I still feel are silly. You are free to make up your own mind, as is everybody else. Re: authority fallacy – no argument there, that’s correct, and I said neither that they were right or wrong.

      My point is about scientific consensus, and that it’s not really what it’s cracked up to be. An interesting read in this vein is Paul Feyerabend’s _Against Method_.

      • Ironically, the consensus view in formal philosophy of science is that Feyerabend’s work, particular in Against Method, is not really what it was cracked up to be.

      • @Jay: OK accepted you didn’t mean to hint that. I read the “those Chinese” as implying condescension. That part of the comment is now withdrawn! Re scientific consensus not being “what it’s cracked up to be,” I’m afraid I really struggle with these arguments. The achievements of science blow the mind (and, should the silly doomsayers be believed, will be blowing the universe once the LHC gets up to its full 14TeV). True _any given statement_ must be viewed as potentially false, but I don’t find Feyerabend at all convincing because, simply put, he cannot propose an alternative by which we have come to have such incredible, powerful and above all USABLE knowledge of the world.

        I concur with the idea of anarchism, or at least a willingness to go very far out on a limb, for new hypothesis development: for instance, the endosymbiotic origin of eukaryotic cells. At any given point the consensus may also be pointing in the wrong direction, and the social systems underlying how we do science may aggravate that: Lee Smolin makes powerful arguments for this being the case with string theory. Also, the current fad topic of the so-called “decline effect” (= publication bias in my book) shows that we need time and repeated trials to really appraoch closer to the truth.

        But OVERALL, science has advanced to knowledge and abilities way beyond anything that could be imagined in the early Enlightenment (picking an arbitrary date for the start of the modern scientific programme). This is, really, Putnam’s No Miracles argument. Science WORKS. So damn well it’s astonishing. All the people who doubt it have lived to doubt it because they have not died of smallpox or of polio or of measles or of cholera or of any of the hundreds of other diseases that carried people off in their childhood a hundred years ago. We must (again, OVERALL) be getting it right.

        I think your egg example is flawed simply because of the vagueness of “good for you.” Science does not ask questions such as “is it good for you.” Ask a specific question: do eggs raise blood pressure, do eggs contain high levels of protein and science provides answers which you may then use to make a generalized judgement. I have an inherited condition of ridiculously high blood pressure. Eggs (sadly) most certainly are not good for me. If you are a bodybuilder with decent blood pressure, you may find a different answer. This application of scientific knowledge to the complex that is an individual’s health is the work of the practitioner end of medicine and, yes, is open to error, but (a) we should not decry scientific knowledge simply because it is not always easy to apply: it is still infinitely better than superstition or guesswork, and (b) as epidemiology and massive lifetime cohort studies improve, our knowledge of how to apply will (through consensus!) equally improve.

        • @Eamon :-) I wish there were a Like button for comments. You are right, and you’ve hit squarely on the reason why I said the resolution (for me) is only partial. Notwithstanding PF’s mainstream academic infamy, I like the spirit of the man and some of his ideas resonate with me.

          @Stuart I haven’t read Smolin, but I’ll check it out; thanks. I do know about the decline effect, and it’s troubling. Yes, I agree, we need time, and we need money, too.

          Re: eggs, your are right, and in fact, I regretted submitting my original comment because I forgot to add the phrase, “…for a /really/ good laugh, ask…”. The situation entertains me because the jury’s still out on what would mistakenly seem to be really simple questions. Another one: why do worms wiggle to the surface when it rains? Unknown. What causes lightning? Some definite ideas, but still murky. Geez, forget about the origin of the universe, the nature of life, the pending global warming catastrophe! Can’t we get settled on something more than ‘sun hot’, ‘rain wet’? (Of course we can, and have; I am being facetious.)

          The truth, in my view, is that the Earth and the people living on it are incomprehensibly complicated, East and West alike. We’ve made enormous strides, particularly in the last few centuries, often entirely and importantly by accident. We’re making them still. At an increasing rate. It’s an exciting time to be alive, and my children may yet live to learn the reason why worms wiggle.

          Back to Andrew’s question, as a result of this blog post I revisited the acupuncture issue with my wife (also a PhD statistician). She reiterated her position: it’s real, and more to the point, it *works*. For her, her family (my mother-in-law has regular treatments), and the families of her friends. She isn’t particularly swayed by evidence (or not) published in research journals 180 meridians away, in a field of specialization other than hers. She echoed @David Yeung’s sentiments, and reminded me that when we took our daughter to the hospital there were two windows at the front door: TCM and WM. Pay your money, take your choice. (She chose WM. I was busy gawking.)

    • Life expectancy at birth has increased significantly. That’s mainly because childbirth is much safer for the baby and for the mother. Childhood is safer too thanks to vaccines. But E[A | A >20] (A is age at death) has certainly not increased by a factor of 2. If you have evidence to the contrary please furnish. Compare force of mortality curves from (say) 1920 and 2000.

      I had two neurosurgeons give me completely different recommendations for surgery. Clearly that was a difference judgment. Pieces of medicine are scientific, but overall there’s a large component which is solely the judgment of individual practitioners.

  24. Andrew, how old are your Chinese statistician friends? Things are changing in a fast pace in China. 20,30 years ago, there were usually two big departments in a hospital, the Traditional Chinese Medicine(TCM) and the Western Medicine(WM). There were even hospitals called TCM hospital(although they had western medicine departments too). The common belief at that time,as I remember, is that TCM works well for chronic diseases, bone fractures, chronic pain and etc and WM works for acute diseases,epidemics and illness that needs surgery. I truly believed this cause that was what I was told. But as I grew up and pondered on whether TCM is really working, I gradually had more faith in WM. This is the trend too, as more and more hospitals are abandoning the TCM departments(ironically, some TCM hospital stopped to enroll TCM doctors). You can see debates on internet about whether TCM really works and whether we should still study it and whether we should totally abandon it.
    Sorry that i am talking about TCM instead of acupuncture. But acupuncture is really just part of TCM and maybe the only part that really seems to work.

  25. Amazingly, Science-Based Medicine has a very long explanation of why acupuncture is fake and then gives a citation that shows it is effective. Science!

    I searched Pubmed and found 140 meta-analyses of very varying quality for a stunning diversity of clinical questions, as always. Data! The results vary and there are many, many study design problems. Is sham acupuncture a treatment or a control? For non-acupuncture pain treatment should the groups receive usual care or identical care? How do you address blinding? There are more.

    That being said, it does seem like the most common conclusion of the better meta-analyses (eg, Cochrane) is that acupuncture is mildly more useful than not using it. (See pubmed id: 19216662, 17619914, 15838072, 20091527)

    Of course, RCTs demonstrate causation, not physiologic mechanism (relaxation). Of course, given the good data in favor of massage for low back pain (Cherkin et al Annals 2011), relaxation is as good a guess as any!

  26. OK – I got here by creating a Google Alert for Acupuncture, and I am a licensed Acupuncturist.

    I read through post and all the comments, and have a simple comment to make. Acupuncture, (Chinese Medicine), helps many people, that is why it persists. 25% of the world uses a form of medicine based on the foundation of Chinese Medicine as primary care. If you could read Chinese you would find much more information in “Scientific Literature” than you will find in English on the subject of Chinese medicine.

    Acupuncture will be very difficult to “prove” scientifically if by scientifically we mean measurement by physiological studies. Acupuncture involves more than the physical body. There is a basic assumption here – that Science proves – BUT Science does not prove anything – it can only disprove assumptions. In this instance I would suggest this “scientific” experiment: find a good, experienced acupuncturist and receive a reasonable number of treatments – say 12 in as many weeks. As yourself if they were helpful. That would be a more legitimate way to approach this question (without an ideological bias which is based on the assumptions which persist due to the Mind Body split that entered into the Western Tradition).

    I can tell you for certain that they will not be harmful – and that you will learn something more of the subject than this thread will ever provide.

    Newtonian Physics is true but so are the Physics of Quantum Mechanics, they are just describing different dimensions of the world in which we exist. This commentary reminds me of a quote I once heard: Someone asked the Physics professor at Oxford what he thought of the ideas that a mister Einstein had about Physics? His response was, “I don’t know much about that man or his theories, but what I can tell you for certain is that if his ideas contradict Newton – he’s wrong.”

    • In this instance I would suggest this “scientific” experiment: find a good, experienced acupuncturist and receive a reasonable number of treatments – say 12 in as many weeks. As yourself if they were helpful.
      See, J Goodman, it’s statements like this that make me _less_ likely to trust accupuncture, because they show that you simply have no grasp of placebos and the nature of RCTs.
      By your standard, little white sugar pills (or even better, little white bitter pills), are scientifically proven to work for pretty much any ailment in the world. And there are thousands of studies to prove this.
      So let me describe to you how sophisticated acupuncture studies work: They recruit a group of patients with similar symptoms – in the German acupuncture study lower back pain. They divide these (about 1000) people into three groups by chance. One group gets standard Western treatment. One group gets “scam” acupuncture – needles are not placed in any particular locations and not deep enough. One group gets “verum” acupuncture – i.e needles get place according to TCM guidelines. They receive this treatment over an extended period of time. In most such studies, acupuncture is “effective” in the sense that both groups of people who get “needled” report more improvement than the people with standard Western treatment. But there has not been, to my knowledge, a single credible study that showed a meaningful difference between verum and scam acupuncture.
      And not one of the commenters defending acupuncture in this thread has been able to explain why real acupuncture doesn’t perform better than fake acupuncture in a trial. Any takers?

  27. Really interesting comments. I particularly like the comments by Phil about sports massage and j about the difficulties of really testing acupuncture. I am Taiwanese and I have no particular theory about whether acupuncture is “real” or not. My wife uses acupuncture regularly and it works for her. I am not inclined to argue with her about it because (1) it does work for her and (2) I don’t want to screw with the placebo effect if that’s what’s working.

    It seems to me the real questions to test (for each class of malady, e.g. back pain):

    1. How often does someone going to a random “legitimate” acupuncturist get pain relief ?
    2. Is there alternate mainstream therapy that would be more effective for them ?

    If it’s placebo, it’s an effective one, and that’s not to be sniffed at. Even if there was an alternate mainstream treatment, my wife probably wouldn’t take it out of discomfort with doctors, hospitals, tests and drugs. A mediocre treatment taken is better than a somewhat better one ignored.

  28. I think that Phil’s point may be the key to this discussion. What I’ve seen of the scientific literature on this question compares acupuncture to drugs and surgery, and finds acupuncture wanting. But surely the more apt comparison is with other forms of massage. Generally speaking, the scientific study of massage is in its infancy, and as a non-Western form of massage that is usually practiced in tandem with various activities that are supposed to have a supernatural effect one would expect acupuncture to be a particularly difficult subject for researchers conducting such study. So I don’t think the literature warrants any definite claims about acupuncture one way or the other.

  29. I’m not sure I see the paradox. There exist statistical studies that say that acupuncture works, and some statisticians think that acupuncture works (for certain values of “works”). They are as uninterested in the underlying mechanism as you are (when you mention that there might be some backdoor way that it works, but you don’t care). They are not looking at it at the level of biology (does your chi energy move properly?) – they are looking at it at the higher level of medicine (does the patient feel better?), or even higher, epidemiology (is there statistical evidence that it works?). Reductionism has been a phenomenally successful way of explaining the material world overall, but it hasn’t been right in every instance (e.g. light ether), and the more complex the system, the less successful it seems to be (although it can still be very accurate, just less so). So just like a pharmaceutical researcher at Merck doesn’t worry about the effects of quantum uncertainty (even though they clearly exist), a well-educated acupuncture patient is perfectly OK not worrying about biological functions (even though they clearly exist).

    Douglas Hofstadter has an excellent example of the problem with reductionism: Let’s say that you create a very complex arrangement of dominoes, ready to fall over, with a complex series of paths all ending along the same line. The setup is arranged such that the first part knocks over every second domino in that final line, the next part knocks over every third domino in the final line, the next knocks over every 4th domino in the final line, and so on. You start the dominoes falling, and a strange pattern of some dominoes falling down and some remaining standing arises. Someone notices that the 61st domino is still standing after a while and asks why. Is the correct answer the reductionist one, namely, “when I push this button, an electrical impulse is sent to a motor, which…and then the first part of the setup works this way…the next does that..the next does this…,” or is the correct level to operate on the higher one, namely “the 61st domino still stands because 61 is a prime number.”?

      • Perhaps “uninterested” is the incorrect term, but you say:

        “my question here is not whether acupuncture could work (possibly through some backdoor mechanism like the needles stimulating your body in some useful way, or whatever) but on the evidence of how much it does work”

        If you are asking for evidence of how much it works, I assume that you would be willing to accept evidence at a higher level than, say, quantum mechanics. So it’s not that you are “uninterested” in lower-level explanations as to why acupuncture works, it’s that you are “not exclusively interested” in lower-level explanations, and I should have used that phrase.

        When I build a model, I would certainly consider a good reductionist causal argument as a piece of “evidence” that a relationship holds, but the more complex the system, the more likely I will look at the problem at higher levels (e.g. a statistical analysis instead of a chemical explanation) and accept that as “evidence” too. Especially for questions of human behavior, until we have bridged the explanatory gap between the physiology of the mind and human behavior, that’s the best evidence that we have :) While the relationship between the human body and health is at least an order of magnitude less complex than the relationship between the physiology of the brain and decisionmaking, it’s probably an order of magnitude more complex than most things that we readily accept non-reductionist, statistical arguments for.

        • Nick:

          Yes, that was my question right there. In general I’m very interested in the mechanism of how these treatments work. (I have aches and pains too!) I just was setting that aside for the purpose of highlighting the contradiction that I explored in the blog post.

  30. Clinical trials of many different forms of Alt-medicine are starting to show up in the rigorous medical literature, and they show, over and over again, that [insert your method here] doesn’t work. What does seem to have an effect is getting more one on one time from a practitioner. It may also be the case that people who see alt practitioners are (apparently) more likely to follow diet and exercise advice, and this helps (my thoughts, not something from any actual study).

    Wrt acupuncture, and any other alt treatment, I have a big concern with deciding that it’s OK even if all it does is stimulate some placebo effect, namely, that it is not ‘doing no harm’. One, it is taking money from people under false pretenses (isn’t that fraud?) and two, it can actually be physically harmful:

    One can make the argument that such harms are due to ill-trained practitioners, but you’d think that the President of South Korea would only see the best…

  31. I’m a Chinese and a statistician who grew up and spent my my early 25+ years in China. I tend to think that the way most Chinese (including your Chinese friends) believe in the Chinese medicines is pretty much in a semi-religion fashion, not hard science based. Sure, like other ancient cultures, some trials and errors during the past few thousand years in China should have come up with some sort of “solutions” or “explanations” for everything. In the end, any alternative medicine (including acupuncture) should have the word “alternative” removed if it is really science. Maybe acupuncture is like “Cargo Cult” for some Chinese?

Comments are closed.