Mar 29 2011
Acupuncture Revisited
Believers in acupuncture claim it is supported by plenty of published scientific evidence. Critics disagree. Thousands of acupuncture studies have been done over the last several decades, with conflicting results. Even systematic reviews have disagreed with each other. The time had come to re-visit the entire body of acupuncture research and try to make sense out of it all. The indefatigable CAM researcher Edzard Ernst stepped up to the plate. He and his colleagues in Korea and Exeter did an exhaustive study that was published in the April 2011 issue of the medical journal Pain: “Acupuncture: Does it alleviate pain and are there serious risks? A review of reviews.” It is accompanied by an editorial commentary written by yours truly: “Acupuncture’s claims punctured: Not proven effective for pain, not harmless.” (The editorial is reproduced in full below.)
Ernst et al. systematically reviewed all the systematic reviews of acupuncture published in the last 10 years: 57 systematic reviews met the criteria they set for inclusion in their analysis. They found a mix of negative, positive, and inconclusive results. There were only four conditions for which more than one systematic review reached the same conclusions, and only one of the four was positive (neck pain). They explain how inconsistencies, biases, conflicting conclusions, and recent high quality studies throw doubt on even the most positive reviews.
They also demolished the “acupuncture is harmless” myth by reporting 95 published cases of serious adverse effects including infection, pneumothorax, and 5 deaths. Some but not all of these might have been avoided by better training in anatomy and infection control.
Their analysis does not prove that acupuncture doesn’t work (negatives are hard to prove) but it unquestionably sheds serious doubt on the claim that it does work. Overall the evidence is inconsistent, and the results tend to be negative among those studies judged to be of the highest quality. Where the results are positive, the reported benefits can be explained by the surrounding ritual, the beliefs and expectations of patient and practitioner, and other nonspecific effects of treatment. There is no evidence to support the vitalistic concept of qi or the prescientific mythology of acupuncture points and meridians; it doesn’t seem to matter where you put the needles or whether the skin is pierced. More modern science-based explanations like increased endorphin production are not convincing, since placebo pills can produce the same effects.
I was delighted when the editor of Pain asked me to write a commentary to accompany the article. It gave me a soapbox in a major medical journal to say all the things I thought needed to be said about acupuncture.
My commentary was edited, but it was a very different experience from the kind of editing I experienced with O,The Oprah Magazine. It was a pleasant collaborative process aimed only at improving the clarity of the writing and strengthening the impact of what I wanted to say.
The journal thought our articles were important enough to warrant a press release. Both Ernst’s article and my commentary immediately got some attention in the media: Science Daily, Medical News Today, e! Science News, and the American Council on Science and Health all reported on them.
Believers in acupuncture will not be pleased. I expect a hostile response and am wondering if Ernst and I should invest in needle-proof vests.
Here is the entire text of my commentary. Thank you to the publishers of Pain, the IASP and Elsevier, for their permission to reproduce it here.
Acupuncture’s claims punctured: Not proven
effective for pain, not harmless
Commentary from Hall H. Acupuncture’s claims punctured: Not proven effective for pain, not harmless. PAIN 2011 Apr; 152(4): 711-712
© 2011 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. This article has been reproduced on ScienceBasedMedicine.org with permission of the International Association for the Study of Pain® (IASP®). The commentary may not be reproduced for any other purpose without permission. Permission to alter the article is not permitted. Permission to translate the article is not granted.
In this issue of Pain Ernst et al. [1], systematically reviewed a decade’s worth of systematic reviews of acupuncture. They found a mix of negative, positive, and inconclusive results. There were only four conditions for which more than one systematic review reached the same conclusions, and only one of the four was positive (neck pain). They explain how inconsistencies, biases, conflicting conclusions, and recent high quality studies throw doubt on even the most positive reviews. Ernst et al.’s analysis cannot prove that acupuncture does not work (negatives are hard to prove) but their study unquestionably sheds serious doubt on the claim that it does work. Overall the evidence is inconsistent, and among those studies judged to be of the highest quality, the results tend to be negative.
Acupuncture is based on pre-scientific concepts of a vitalistic entity (qi) and of meridians and acupuncture points unknown to anatomists. More scientific explanations have been offered as to how it might work, including a counterirritant effect or the gate control theory of pain. There is evidence that acupuncture can stimulate endogenous endorphin production, but there is evidence that placebo pills can do that as well. Importantly, when a treatment is truly effective, studies tend to produce more convincing results as time passes and the weight of evidence accumulates. When a treatment is extensively studied for decades and the evidence continues to be inconsistent, it becomes more and more likely that the treatment is not truly effective. This appears to be the case for acupuncture. In fact, taken as a whole, the published (and scientifically rigorous) evidence leads to the conclusion that acupuncture is no more effective than placebo.
Acupuncture research is inherently riddled with pitfalls. What constitutes an adequate control? People can usually tell whether or not you are sticking needles in them. Various controls have been devised, such as comparing ‘‘true’’ acupuncture points to ‘‘false’’ ones. The best control so far is an ingenious retractable needle similar to a stage dagger, where the needle just touches the skin and retracts into a sheath. Unfortunately, there is no way to blind the practitioner, so double blind studies are impossible.
The practice of acupuncture is also not sufficiently standardized, which makes it difficult, if not impossible to pin down reliably for objective study: there are various schools of acupuncture with different acupoints, and studies of acupuncture have included ‘‘electroacupuncture’’ (with or without needles), ear acupuncture, cupping, moxibustion, and other loosely related procedures. In their book, The Biology of Acupuncture, Ulett and Han [3] showed that transcutaneous electrical stimulation at a single arbitrary point on the wrist was just as effective as piercing the skin at traditional acupuncture points.
In more than one recent study, researchers have chosen not to use a sham acupuncture control group. Their reasoning? Since sham acupuncture has been shown to work as well as real acupuncture, then sham acupuncture must be an effective treatment too! Imagine applying this reasoning to a drug trial: if the drug and placebo got the same results, would you decide that the drug worked and that the placebo was just as therapeutic as the drug?
It does not make any difference where you put the needles or whether you use needles at all. Touching the skin with toothpicks works just as well. The crucial factor seems to be whether patients believe they are getting true acupuncture. It is becoming increasingly clear that the surrounding ritual, the beliefs of patient and practitioner, and the nonspecific effects of treatment are likely responsible for any reported benefits.
Is there really any need for more studies? Ernst et al. point out that the positive studies conclude that acupuncture relieves pain in some conditions but not in other very similar conditions. What would you think if a new pain pill was shown to relieve musculoskeletal pain in the arms but not in the legs? The most parsimonious explanation is that the positive studies are false positives. In his seminal article on why most published research findings are false, Ioannidis points out that when a popular but ineffective treatment is studied, false positive results are common for multiple reasons, including bias and low prior probability [2]. More studies are not the answer. No matter how many studies showed negative results, they would not persuade true believers to give up their beliefs. There will always be ‘‘one more study’’ to try, but there should be a common-sense point at which researchers can agree to stop and divert research time and funds to areas more likely to produce useful results.
Of course, advocates of acupuncture have argued that it is worthwhile even if it only produces a placebo response; and that it is harmless, so it does not hurt to try it. Ernst et al. however, have shown that acupuncture is not harmless. While many of the reported adverse effects could be avoided by proper training in sterile precautions and anatomy, they correctly point out that even one avoidable adverse event is too many. With any treatment, we have to consider the risk/benefit ratio. If there is no benefit, any risk is too much. And there are other harms that they did not mention: time and money wasted, effective treatment delayed, unscientific thinking encouraged.
Placebos are unethical: our patients trust us not to prescribe them. With the current state of the evidence, I do not think we should be recommending acupuncture to our patients. On the other hand, if patients ask about it and want to try it, we should not try to stop them. We have a responsibility to educate them, but not to make decisions for them. We can tell them that although some patients believe it has helped them, the evidence does not show that it works any better than placebo, and there is a small risk of infection and other complications. With this information, they can then make their own informed decision.
In summary, Ernst et al. have shown that the evidence for efficacy of acupuncture for the treatment of pain is questionable, to say the least, and of particular concern is that its use can be dangerous. If the 57 systematic reviews they surveyed had been for a prescription drug and a similar list of serious adverse effects had been reported for that drug, we would hesitate to prescribe that drug. Is there any reason not to hold acupuncture to the same standards?
Conflict of interest statement
I have no conflicts of interest to report.
References
- Ernst E, Lee MS, Choi TY. Acupuncture: Does it alleviate pain and are there serious risks? A review of reviews. Pain 2011;152:755–64.
- Ioannidis JP. Why most published research findings are false: author’s reply to Goodman and Greenland. PLoS Med 2007;4:e215.
- Ulett GA, Han SP. The biology of acupuncture. St. Louis, USA: Warren H. Green Inc.; 2002. 160p.
238 Responses to “Acupuncture Revisited”
Very good! I already summarized your comment in a blog of my own :
http://www.skepsis.nl/blog/2011/03/acupunctuur-tegen-pijn-dubieus-en-riskant/
Even if you can’t read the Dutch, you may appreciate the pictures.
I would like to take the opportunity to point out that Ernst et al. don’t mention 57 reviews, but 58. I checked all references mentioned in the beginning of section 3.1. There they give only 55 references, because they forget 67, 160, 161.
For my blog I did some more counting:
First the quality of the underlying studies:
7 reviews: poor
9 reviews: mostly poor
32 reviews: variable
10 reviws: good.
For the reviews themselves:
poor or moderate ones (23): 13 got a + , there were 4 times a +/- and 6 were -
good or excellent (35): 12 +, 9 +/-, 14 –
There were 6 reviews that were good, with underlying studies that were also good. Here the scores were 1 + , 4 +/- , 1 –
In 1989 three Dutch researchers (Ter Riet, Kleijnen, Knipschild, http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1371348/ , http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1021032/ ) examined acupuncture studies, not only for pain, but also addiction and asthma and so on, and they concluded that the number of positive trials decreases with increasing quality.
Dear Harriet Hall.
If I were the sole proprietor of the biggest oil company I should do all possible actions so that Los Angeles (the most “muscular” town in the USA) neither will ever have a public-private subway, nor electric, nor low fuel consumption cars.
If I were the sole proprietary of the biggest fast-food company I should plan the transformation of all US children into very fat adults, few able or completely unable to think autonomously, but in compensation for it very specialized to eat junk-food and move their bodies no longer, except for the minimum enough to drive high fuel consumption cars.
If I were the sole proprietary of the biggest drug company and I read the following very bad news for my company…
— The Times Higher Education Supplement (27 July 2007) reports an 31.5% increase in applications for ‘university’ courses in complementary medicine. Compare this with 19 per cent fall in applications for places on anatomy, physiology and pathology courses, and a relatively low 6 per cent rise in applications for pharmacology, toxicology and pharmacy.—
… I should engage the most intelligent medicine Professors, who are the ablest thinkers to transform the most innocent symptom of a patient into the heaviest imminent perception to be at death’s door.
“And if I were a good man, I’d talk with you more often than I do. ” (Pink Floyd)
Probably all CAMs are placebo medicines but MDs and all Well Trained people treating patients by them love children, nature and freedom.
Best regards,
Stefano Marcelli MD
(not paid to generate fear and diseases)
…so Dr. Marcelli is arguing that Ernst was paid by drug companies to produce this negative result?
You know what we should do? We should get Big Needle to conduct another meta-analysis to get positive results! That way, we can tell both sides of the story!!! WHAT A GREAT IDEA!!!!!
Sweet Jebus, maybe acupuncture just doesn’t work and it’s not all a giant conspiracy.
@ Stefano Marcelli:
Are you in Italy? Are you advocating the purposeful use of placebos? If so, I wonder whether this is a more common point of view on the European continent. Here’s a French physician unabashedly arguing a similar point for homeopathy:
This strikes me as something that physicians in North America (I can’t speak for elsewhere) have, for better or worse but as part of a broader cultural change, repudiated in the past generation or so: old fashioned medical paternalism.
No?
“There is evidence that acupuncture can stimulate endogenous endorphin production, but there is evidence that placebo pills can do that as well. ”
The pharmaceutical armamentarium includes dozens of medications with proven abilities to fight pain quickly, effectively and safely. Even if acupuncture proves to be mildly effective, why go through the risk, time and ordeal when far better options are available?
Thanks, Harriet Hall.
If it’s not a medical prodedure or treatment, I don’t want my doctor performing or prescribing it.
If the risk/benefit analysis is that acupuncture is a wash, we can allow acupuncture to be practiced. But that doesn’t mean I want to get it from my doctor.
There are ways to reduce the risks of acupuncture. The one pmoran consistently proposes is to have it be offered under the supervision of a doctor. This is as unethical as doctors prescribing sugar pills — or even unindicated antibiotics — and cannot be considered an acceptable solution.
Another way is to require anyone offering acupuncture to carry insurance and to have the insurance conditional on the acupuncturist respecting certain conditions (receive training in anatomy and sterile procedure and to limit practice to certain conditions). There are risks here as well — if acupuncurists must be licensed, then people are likely to believe it to be a legitimate medical prodedure — but those can be perhaps limited by requiring them to have a license and insurance that also apply to (say) tattoo artists and piercing studios.
To repeat from an earlier thread: the difference between acupuncture and sham acupuncture is small, probably too small to be clinically important. The difference between sham acupuncture and no acupuncture is moderate, probably large enough to be clinically important. You appear to be better off getting a fake acupuncture treatment than a fake steroid injection. Perhaps there is a difference between plain placebo and Extra Strength Placebo.
But there is a larger set of questions not touched upon by Dr. Ernst in his review article (and beyond the scope of that article). It is very difficult to study pain and its treatment. Inferences from randomized clinical trials (RCTs) are not straightforward, even when the methods are well reported. This is because there are issues in the analysis of pain in which there is no general consensus as to the best method to follow.
For example, it is usual and customary to analyze effect sizes in terms of changes in pain scores from baseline to various time intervals after the initiation of treatment. Generally this means taking some measure of central tendency (the mean) and its variance, and comparing these measurements between groups. However, it is likely that these comparisons miss at least some of the truth.
In an article titled “Methodological issues in clinical trials of opioids for chronic pain” (Neurology 2005;65 Suppl 4:S32-S49) Nathaniel Katz discusses some problems related to difficulties in interpreting studies that show that opioids are no more effective than placebo for pain conditions. Heterogeneity among patients with respect to responsiveness, for example, can make the average pain response an imperfect measure of the effectiveness of the drug. Knowing which endpoint to measure (and how to measure it) is not an easy or settled matter, and it would be misleading to imply otherwise. If clinical trials fail to show that opioids are no more effective than placebo, does that mean that they should be abandoned (since they have dangerous side effects), or does it mean that the effects may be missed by some commonly used analytical methods?
Similarly, how to handle missing data (which happen in all longitudinal trials) is not an easy issue to resolve. How you handle this (baseline observation carried forward, last observation carried forward, multiple imputation methods) can make a difference in what you conclude and what effect size you measure.
Most pertinent to the present discussion, the greater the placebo response for any intervention, the harder it is to show a difference between placebo and the real thing. This happens for statistical reasons, and imposes difficulties of interpretation for all clinical trials.
Dr. Ernst has done us a service in collecting reports of adverse effects of acupuncture; the 95 cases of serious adverse effects (5 fatalities) supports his conclusion that all acupuncturists should be trained adequately. However, the 95 cases are what is called numerator data; to interpret them properly, a denominator is needed. The number of patients treated with acupuncture would be the population at risk, and its magnitude is not reported and probably not known.
So my point in all of this is that we must avoid the seduction of the straightforward interpretation. If anyone claims that the interpretation of clinical trials is clear-cut, we should regard that claim with the same horror we would if a nominee for Chief Justice of the United States sat in front of the Senate Judiciary Committee and said that applying the Constitution to specific cases is just like an umpire calling balls and strikes. Sometimes the posts and comments on this website give an impression that once the data have been presented, the conclusions are clear. This is probably an unduly optimistic view of how science works.
windriven on human psychological quirks: “Even if acupuncture proves to be mildly effective, why go through the risk, time and ordeal when far better options are available?”
Good question, but people do.
I think it has to do with micheleinmichigan’s point that people are more likely to adhere to a treatment that they understand. Many people feel as though they have souls (I know I do) so presumably many people find it easy to extend that feeling to conceptualizing qi (which I do not do). Medicine is hard, and people are naturally suspicious of things they don’t understand. For people who feel that they understand qi and are suspicious of medicine, acupuncture may be “better” in that it is psychologically accessible to them.
Limiting the harm of a second- or third-rate treatment may be a lesser evil.
People do lots of things we don’t objectively think are good in themselves (drug and alcohol use, for instance) and some of the most effective responses of a pluralistic society appear to be harm-reduction in effect. (Welfare payments, union wages, subsidized daycare and the full participation of women in the workforce reduce the harm of alcoholism to families, making abolition less attractive. Needle-exchange programs are morally repugnant to many but appear to be objectively better from a utilitarian perspective than a war on drugs.)
pmoran has a point in supporting harm-reduction measures with respect to unsupported parallel practices. I just don’t accept his proposal that acupuncture be incorporated into actual medical practice.
Why wouldn’t make that statement a bit more pointedly (sorry for the bad pun) by explaining them that the degree to which a patient truly believes that acupuncture will work seems to be the key factor in determining its efficacy?
^^^ ugh… sorry for the missing words… only one cuppa so far today :-/
“if the drug and placebo got the same results, would you decide that the drug worked and that the placebo was just as therapeutic as the drug?”
Isn’t this what happened with antidepressants? I read that a number of meh trial outcomes were suppressed by the pharma companies that conducted them.
More conspiracy crap – ?
>>WilliamLawrenceUtridgeon
>>Dr. Marcelli is arguing that Ernst was paid by drug companies to produce this negative result?
No! Dr. Marcelli cannot judge Ernst as a person, but as a “science’s citizen” he can judge his public position.
So I must note that side effects of acupuncture compared to those of chemical (also herbal) drugs and radiations and surgical interventions are NOTHING. Consequently it’s not necessary to be Sherlock Holmes to think it’s a bit strange that Ernst – a university professor of CAMs – seems to go against his own specialty. What should you think about me as a cardiologist if I said OFTEN that my specialty is not only unuseful but also dangerous?
I am both an acupuncturist and the first skeptic in regard to acupuncture. See here please: http://tinyurl.com/6bpvb9v. Beside the fact I am paid directly by patients consulting me privately (placebo or not), as an independent researcher I found enough observational reasons to continue to be confident in acupunture theory of meridians.
BW
Stefano
Ed Whitney
“The difference between sham acupuncture and no acupuncture is moderate, probably large enough to be clinically important”
The difference in such an unblinded comparison cannot be considered clinically significant, except possibly for evaluating the strength of the placebo response.
“Perhaps there is a difference between plain placebo and Extra Strength Placebo.”
It’s a fairly well know fact that the more radical a treatment appears, the stronger the placebo response. (saline injection > sugar pill, etc.)
It doesn’t change the fact that prescribing placebos is unethical and a violation of informed consent. Additionally, the use of inert treatments has the serious potential to crowd out/ replace active, effective treatments.
“However, the 95 cases are what is called numerator data; to interpret them properly, a denominator is needed.”
Not when the benefit is zero. ANY risk with zero benefit is unwarranted, no further math needed.
@Dr. Marcelli
You’re not comparing the side effects of acupuncture to that of real treatment, that’s not a fair comparsion because real treatment has both placebo effects and real effects. You must compare the side effects of acupuncture to a similarly inert or placebo treatment such as homeopathy or reiki which have essentially no risks beyond the common, corrosive effect of such nonsense on critical thinking. The whole point of all these trials is to see if acupuncture has any actual effects beyond placebo. The suggested result is that acupuncture is an unusually effective placebo but one that causes harm. Therefore the appropriate choice is between acupuncture and all other placebos, picking the one that causes the least harm. Whether the benefits are worth the risk of real treatments depends on the intervention and the disease. Acupuncture is only an option when real treatment isn’t available – essentially it’s a needle-based emotional coping technique used to mark time while the symptoms regress on their own.
Dr. Ernst did a great service by evaluating the scientific research behind acupuncture and arriving at an empirical conclusion. If you said you were a cardiologist who found their discipline unuseful and dangerous, I’d suspect you were incompetent. Ernst, who started his career in CAM and only became a skeptic when he realized how shallow the empirical basis is, is doing a truly great service to the world because unlike other CAM practitioners he is attempting to find out if the interventions are useful rather than assuming they are.
The fact that you directly profit from what appears to be a placebo intervention, both in terms of money and the emotional gratification you must feel at “healing” your patients means I am far, far more likely to be suspicious of you than I am of Ernst. It’s a toss-up on who would be more likely to be intellectually dishonest, an acupuncturist or Big Pharma, because though the profits of Big Pharma are much greater, so is the oversight. You, on the other hand, have a direct motivation to lie to yourself and your patient because acupuncture is how you feed yourself. And your oversight is essentially zero. So don’t play the Big Pharma card as if it were a winning one, you’re applying an intellectually dishonest double-standard that you are apparently too blind to see. I’m not sure if this scares or angers me but either way I’m not happy.
drmarcelli,
Outside of twitter, there’s really no reason to use a URL shortener. I like to know where I am going before I click on a link.
For anyone who wants to know before clicking, drmarcelli’s link is:
http://www.geneticacupuncture.com/ifwetried.htm
Dr. Marcelli, As has been mentioned before, the side effect is not “nothing”. Risks of infection, delay in pursuit of rational treatments and wasting precious monetary resources on something that has no demonstrable benefit are all side effects.
You make a few inappropriate comparisons between oil companies, fast food and Edzard Ernst.
An example would be that oil companies don’t want anything but gas powered cars. Definitely not true –gasoline is but one product made from oil. Think of all the plastics other petrochemicals that are made during petroleum distillation. I believe we use plastic in our hybrid and EV cars no?
Also, fast food companies surely did not start up with the intent to make people fat. Customers who die from obesity related illness do not buy any more hamburgers. If the fast food companies really wanted that outcome they would not make the effort to include more healthful choices on their menus.
You call yourself an independent researcher. Does that mean you ‘research’ your own patients without any peer review? Dr. Ernst has his work strengthened by the process of critical review by fellow researchers. It is now out in the literature for the rest of us to critique and build upon –where is the fruit of your research so that we may take a critical look at it?
Also, do you have an autoclave in your office?
drmarcelli,
“So I must note that side effects of acupuncture compared to those of chemical (also herbal) drugs and radiations and surgical interventions are NOTHING.”
That’s the point, inert treatments can be expected to have zero side effects. Active treatments can have unintended side effects. It’s a matter of risk to benefit ration. If the benefit outweighs the risk, it’s generally worthwhile. Zero benefit requires zero risk.
However acupuncture is not always truly inert, as the posts says. On rare occasions, infection, pneumothorax, and even death are side effects of acupuncture.
Setting aside the ethical and other concerns…
For anyone advocating acupuncture because of its nonspecific (placebo) effects, if you are going to recommend an inert treatment, go with the one with the least risks and complications. Sham acupuncture therefore trumps actual acupuncture.
@drmarcelli,
“side effects of acupuncture compared to those of chemical (also herbal) drugs and radiations and surgical interventions are NOTHING.”
I agree. But acupuncture advocates have claimed that it is completely harmless, and Ernst has shown that that is clearly not true.
“I found enough observational reasons to continue to be confident in acupuncture theory of meridians.”
As Mark Crislip has said, “In my experience” are the 3 most dangerous words in medicine. Medieval bloodletters were confident in their observational reasons as they continued to kill patients. That’s why we need science. You seem to reject the scientific method. What would you think of a doctor who practiced bloodletting today because he had found enough observational reasons to continue to be confident in the theory of the 4 humors?
@drmarcelli: “it’s a bit strange that Ernst – a university professor of CAMs – seems to go against his own specialty.”
You don’t seem to appreciate that there is a difference between studying a subject and uncritically promoting it.
Ernst’s professorship was originally funded by Sir Maurice Laing, a CAM supporter, with a brief to conduct proper research into CAM. When the original endowment ran out after ten years, Sir Maurice provided further funding.
If you want acupuncture promoted, you should hire an advertising agency.
@drmarcelli
“I found enough observational reasons to continue to be confident in acupuncture theory of meridians.”
Would you care to share those with us?
@Alison
“Limiting the harm of a second- or third-rate treatment may be a lesser evil. ”
Indeed. I would never argue that someone should be prohibited from doing something to their own body so long as they are sane and have complete information available.
Excellent!
Despite this overwhelming evidence, the proponents of acupuncture and Chinese medicine are using political pressure, lobbying and legislation to increase their scope of practice.
For instance, in California, State Senator Leland Yee is proposing a new legislation (SB 628) to recognize the legitimacy of Traditional Chinese Medicine (TCM) because it is 5000 years old. SB 628 will change the name of the Acupuncture Board to the TCM Board.
Furthermore, the proposed bill would allow acupuncturists to practice traumatology!
If this bill passes, California TCM providers will be treating fractures, dislocations, Injury of muscle and tendon, internal traumatic syndrome, and “qi injury” (whatever that means!)
Thins will be a blow to science and rationality in medicine!
Anyone who lives in California, and is concerned with the public’s safet,y should contact their elected representatives, and ask them to oppose SB 628.
@Windriven and others wondering “why the expense and time and effort”? When there are so many clinically proven ways to treat problems.
The first answer is that there aren’t always. Either for that particular individual with the problem who isn’t finding the conventional treatment helps, or sometimes for a chronic or rare disease that simply has no treatment. I guess the suggestion is that those of us who fall into one of these camps should just sit tight and hope for a science based solution to come along before the condition or our natural life span comes to an end. There is nothing irrational about thinking that way. It is also something that could only be felt by someone who hasn’t had the experience of daily chronic pain, fatigue or some other disabling symptom. This population are also very vulnerable to exploitation by CAM ideas and providers, so I am not sure what the answer is about regulation/ licensing. People should be able to exercise agency in trying whatever they want to see if it works, and when you are desperate enough it stops mattering how it works.
Before you jump all over how ridiculous that sounds, consider that I inject, every day, a medication meant to slow the progression of my multiple sclerosis, and no one knows how it works or even where it goes once injected. They don’t know how your body excretes it. What they don’t know is more than what they do. What kind of crazy person takes something into their body every day when there is no scientific data explaining how it works (as per the company that makes it)? I take it because there is at least some data showing efficacy. And because it feels better than doing nothing at all to try and help myself. And because the alternatives are less desireable.
I also go to acupuncture. I go not because I am under the delusion that it will cure my disease or even change the progression, I go because it helps the day to day symptoms- a lot. And I really don’t care why as long as I am able to hold a pencil, walk normally and experience fewer vertigo episodes. If I can accept a pharmaceutical intervention despite not knowing how it works or why, is it so crazy for me to accept acupuncture when I have no definitive answer as to why it works? With Copaxone, I have good studies showing long term results- it does nothing day to day to show me it’s working at all. With acupuncture, I have no studies proving anything, but day to day improvement in how I feel.
@CarolM
“Isn’t this what happened with antidepressants? I read that a number of meh trial outcomes were suppressed by the pharma companies that conducted them.”
I think you are referring to this study: http://www.nejm.org/doi/full/10.1056/NEJMsa065779
I recently met the lead author of this study and he told me he is firmly convinced that antidepressants are effective and life-saving in major depression. He was concerned that selective publication would lead to unrealistic estimates of effectiveness, not that the drugs are ineffective.
For Karl Withakay: “ANY risk with zero benefit is unwarranted, no further math needed.”
Maybe there are differences in orientation between sensation (emphasis on facts and their clear meaning) and intuition (emphasis on multiple potential meanings and possibilities for facts). This was all described by Carl Withasee Jung, but it is perhaps peripheral to this discussion. I am not so certain that the effect of acupuncture is proven to be zero, and therefore still need the denominator for the 95 cases in the numerator for serious adverse effects of acupuncture. Further math needed, as I see it. See below for the reason.
For Harriet and the gang:
Compare and contrast Ernst 2011 with Hopton A, MacPherson H
Acupuncture for Chronic Pain: Is Acupuncture More than an Effective Placebo? A Systematic Review of Pooled Data from Meta-analyses
Pain Practice, Volume 10, Issue 2, 2010 94–102
One advantage of Hopton over Ernst is that Hopton gives a better accounting of what went into the quality assessment. Ernst cited criteria based on a 1991 article by Oxman and Guyatt in 1991 which have generally accepted things a good systematic review must do. It asks, for example, “Was the search comprehensive?” Hopton, like Ernst, uses Oxman and Guyatt, and also uses a more recently validated tool for systematic review quality, namely, AMSTAR, which is validated with a free link at: (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2131785/?tool=pubmed and at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1810543/?tool=pubmed ) .
AMSTAR gives more specifics for how the criteria are to be decided upon; for the comprehensive literature search, AMSTAR, unlike Oxman and Guyatt, specifies that there must be at least two electronic sources (e.g., Central, EMBASE, MEDLINE) with key words and/or MESH terms stated, and the search strategy provided. One of my frustrations with Ernst was that he cited the Oxman and Guyatt criteria without details as to how they were applied. Advantage: Hopton.
Now, Hopton and MacPherson restricted their analysis to systematic reviews which pooled data into meta-analyses. For short term results, they found acupuncture superior to sham acupuncture for low back pain, chronic osteoarthritis of the knee, and headache. In the longer term (6-12 months), the results were more uncertain and inconsistent; the effect sizes were small to moderate for knee and back pain.
Hopton was probably published too late for Ernst to have had it for his review of reviews. I cannot proclaim that Hopton is right and Ernst is wrong; the only clear thing is that different methods led to different conclusions, and that the methods were more clearly described in Hopton.
To sum up: It is a mistake to think that Ernst has provided conclusive evidence that the effect of acupuncture is zero. The conclusions giveth, but the methods taketh away, as they say.
@ CLK
” I go because it helps the day to day symptoms”
If you feel better after acupuncture, you are certainly justified in continuing it. That is very different from claiming that it works.
# windrivenon 29 Mar 2011 at 12:50 pm
Would you care to share those with us?
My observations are there: http://www.geneticacupuncture.com
Thank you.
Stefano M.
@Ed Whitney,
“It is a mistake to think that Ernst has provided conclusive evidence that the effect of acupuncture is zero.”
I didn’t say that. I said “Ernst et al. have shown that the evidence for efficacy of acupuncture for the treatment of pain is questionable.”
The point is that after several decades of study, the evidence is still inconclusive. If acupuncture were truly effective we would expect the evidence to be clear by now. The published evidence is compatible with the hypothesis that acupuncture is nothing more than a compelling placebo system.
Here is one typical example of drmarcelli’s observations:
http://www.geneticacupuncture.com/03.GBchannel&biliarytree.EN.htm
He imagines similarities of shape between a diagram of “gallbladder”acupuncture meridians on the head and the anatomical relationships of liver and gallbladder. He thinks this validates acupuncture theory and shows that the liver controls the brain. Magical thinking at its best. Actually pretty funny.
It’s like someone had MS Paint, some pictures of human anatomy, and too much time on their hands.
@ Harriet Hall: You did not say that the effect of acupuncture was zero; Karl Withakay said so, and that therefore no further math was needed. I need more math for the reasons I specified, when it comes to the need for a denominator to go under the numerator of 95 serious complications.
I agree that the data are compatible with the hypothesis that acupuncture is a compelling placebo system. Perhaps the data are compatible with the hypothesis that acupuncture is “nothing but” a compelling placebo system.
The data are also compatible with the hypothesis that something about acupuncture requires that we may need to have a more subtle thought system to help us approach an even more subtle reality system. Comparing Hopton and Ernst could be a profitable exercise in this direction.
@CLK
There’s a difference between your medication (don’t know how it works; validated that it does work) and acupuncture (don’t know how it works, don’t know whether it works).
Medication also has going for it a set of high prior probabilities (we know the body is made of chemicals, we know chemicals interact, we know certain chemicals alter the function of receptors which in turn alters the function of the body).
Further, if anyone has any condition where there’s truly no treatment, the use of low-cost, low-risk interventions to mitigate symptoms is reasonable and rational. I would never recommend acupuncture be made illegal, but I would recommend that the “magic” parts of it – qi, yin, yang, meridians, acupuncture points – be eliminated and the materials used be as safe as possible. No deep penetration, even use non-penetrating needles or simple pressure, sterile instruments and needle areas where the risk of harm is minimal (i.e. avoid the torso and ideally superficially needle over bones). But don’t call it a cure, don’t continue to invoke magic, discard the aspects that waste time, cost money and put patients at risk. Far more than any other part of acupuncture or TCM in general, that is what I reject – magic, waste, unnecessary training, the teaching of nonsense and the retention of worthless tradition. Anything the ancients did, we can certainly do better, faster, safer and for less money.
Ed Whitney,
RE: “I am not so certain that the effect of acupuncture is proven to be zero, and therefore still need the denominator for the 95 cases in the numerator for serious adverse effects of acupuncture. Further math needed, as I see it.”
Were the below not your words?
“To repeat from an earlier thread: the difference between acupuncture and sham acupuncture is small, probably too small to be clinically important.”
If you dispute that sham acupuncture is equal to actual acupuncture, fine, present your case (as you subsequently did), but if acupuncture and sham acupuncture are equivalent, than there is no reason to recommend verum acupuncture over sham, as verum acupuncture has a non-zero risk.
The fact that we can do study after study, review after review, and meta-analysis after meta-analysis and still get weak, inconsistent, and conflicting results (even if you don’t accept that the better controlled and blinded studies tend to be negative), would seem to indicate that the effects are weak and inconsistent at best and more likely just noise.
>>>
“For short term results, they found acupuncture superior to sham acupuncture for low back pain, chronic osteoarthritis of the knee, and headache. In the longer term (6-12 months), the results were more uncertain and inconsistent;”
I would think that this is consistent with (though not necessarily indicative of) a placebo response. Likely either that or people can develop a tolerance of acupuncture.
Ed Whitney
“The data are also compatible with the hypothesis that something about acupuncture requires that we may need to have a more subtle thought system to help us approach an even more subtle reality system”
The data are also consistent with my invisible, phase shifted, magic purple dragon choosing to empower the practice of acupuncture when and only when it feels like it, I suppose.
Occam’s Razor baby, Occam’s Razor.
I’m hoping you’re not invoking post modernism and different ways of knowing.
>>>Harriet Hall
That’s why we need science. You seem to reject the scientific method. What would you think of a doctor who practiced bloodletting today because he had found enough observational reasons to continue to be confident in the theory of the 4 humors?>>>
I’m sorry, dear all friends, I answer only Harriet Hall because she is the article writer and I have not all the energy required to think and write more than one continual hour in English.
Ten years ago, were you skeptic as you are today? Pause…
Did you act no medical behavior being not in accord to Science Based Medicine? Pause…
And twenty years ago? Did you need science as you now need to recruit me to it? Pause…
And before beginning medical studies? Long long pause…
If science based behaviors are continuing to fail (i.e. see this last irrational war and Japanese atomic failure) is because many MDs like me are mentally children or apes or worms, evo-devo biologically speaking. There are degrees in nature. Maybe I’m arriving to where Prof. Ernst and you are now, maybe I will go also beyond.
I dont reject scientific method, I need only to be recruited to skepticism. I am here to discuss with other MDs, to see things by other points of view.
Sincerely,
Stefano
For an example of how the placebo response rate can affect the contrast between true and sham procedures, see Peng B, Pang X, et al. A randomized placebo-controlled trial of intradiscal methylene blue injection for the treatment of chronic discogenic low back pain. Pain 2010;149:124-129.
This study looked at 72 patients with discogenic low back pain at a military facility in Beijing. Patients had diseased discs (documented by discography), and were randomized to intradiscal injection of either methylene blue (MB, n=36) or saline (n=36). The outcomes were pain on the Numerical Rating Scale (NRS) and disability on the Oswestry Disability Index. The MB group had a large favorable response for both NRS and ODI, and the saline group had practically no response on either outcome scale. The randomization, allocation concealment, and blinding were reported to have been done, which controls the most important sources of bias. (The success of blinding was not so clearly reported, and perhaps that is a problem). But the quality of the study was pretty good, going by the criteria the Cochrane Collaboration uses to judge risk of bias.
It looks pretty good for MB, but I know no one who is taking this study at face value. The placebo response was suspiciously low; if an injection is supposed to have a large placebo effect, this study certainly does not prove it. Something is amiss, in my opinion, but until the study is replicated (as is in progress now), no one should believe a word of it.
The reason to bring this up is that it is a sure thing that if enough people start injecting methylene blue into enough discs, some knucklehead sooner or later is bound to hit the thecal sac. Intrathecal MB can be very bad news; the adverse response rate would be very bad if it happens.
So I agree with the conclusion of Ernst: acupuncture should not be done by knuckleheads. And even a well-reported randomized trial should be suspect if it just plain looks flaky.
Methylene blue, as we know, does prove the truth of the law of similars in homeopathy. It is the treatment of choice for methemoglobinemia. You inject this blue medicine into that blue patient, and the patient turns pink! So there you go.
For Karl: “Occam’s razor, baby, Occam’s razor.”
Do you know what William of Occam died of?
(It was multiple causes.)
Also: I have met a postmodernist or two, and they were very literal when it came to the interpretation of the parts of their faculty contracts that dealt with fringe benefits.
drmarcelli,
What would you think of a doctor who practiced bloodletting today because he had found enough observational reasons to continue to be confident in the theory of the 4 humors?
@drmarcelli
Ten years ago I was far, far more credulous than I am today. However, a fair amount of reading on the sources of bias and the scientific method illustrated just how conceptually simple science really is, and just how powerful a tool it is for arriving at meaningful knowledge that can be used to improve lives. Even if in the past there were doctors and readers of this blog who were not applying the best knowledge and the best, science-based practice to the treatment of their patients – that doesn’t justify perpetuating the error. The whole point of science is to learn more and the whole point of being a science-based practitioner is being willing to change your mind in the face of convincing evidence. It is not a crime or a problem to make mistakes or operate under mistaken assumptions. But resolutely failing to change one’s mind irrespective the evidence, special-pleading with evidence to make it go away, willfully ignoring contrary evidence because you don’t want it to be true is harmful to patients and society.
Wars are not scientific failures, the are political. Science can make wars more deadly (or less deadly) and one day may help make decisions about wars (See Sam Harris’ The Moral Landscape for a baby step in the right direction) but right now “war” isn’t a scientific decision (chances are if it were, the decision would be “don’t, that’s stupid”). Similarly, the events in Japan were due to a 9.0 earthquake, and dykes that were a meter too short. Science built a reactor that provided power to millions, and helped build the untimately inadequate defences against the tsunami – but I don’t think traditional chinese medicine has much to offer against a 9.0 earthquake either. Humans are indeed apes, which is why science is necessary to create research that improves lives.
Today I ranted on SaveYourself.ca that the acupuncture debate
So is my mind closed? Damn straight. That’s the point. It didn’t used to be about this — but it is now, and for many good reasons.
“Should we keep an open mind about astrology, perpetual motion, alchemy, alien abduction, and sightings of Elvis Presley? No, and we are happy to confess that our minds have closed down on
homeopathyacupuncture in the same way.” — Mike Baum, The dangers of complementary therapy, Breast Cancer Res. 2007; 9(Suppl 2): S10@CLK
Best wishes on managing your MS and hopes for early progress in arresting and reversing its effects.
Part of Dr. Hall’s point, I think, is that acupuncture hangs in there because there are some who perceive benefits, illusory as those benefits might be.
But we return to the problems that arise with the study of pain and its response to treatment. Special study designs, such as enriched enrollment with randomized withdrawal, have been developed for the study of many pain medications. This involves enrolling all participants in an open-label run-in period, seeing how many respond to the drug with a meaningful (at least 30%) pain reduction, and then randomizing the responders to either continue with active medication or to placebo. It sounds weird, but it is a very common design nowadays, and is partly motivated by the idea that for certain interventions, analysis of the average pain response will obscure the fact that there is a subset of all participants for whom the drug is effective. There may eventually be genetic markers for this responsiveness, but for now there are none, and raw empiricism has to determine which patients will obtain pain relief. Nerve stabilizers like gabapentin and pregabalin have lots of such studies; even opioids have some of these study designs.
There are some conditions for which a subset of patients do respond and others have no response; the average response can obscure the clinically important response of a subset. For breast cancer, Herceptin works only on women whose tumors have excess expression of HER2. This is about 25% of breast cancers; this minority has an important response, and that response could easily be missed if all cancers were analyzed together.
So it is possible that there is some subset of patients who do respond to acupuncture, for reasons that are not known, and may someday become clear. It may be that CLK is perceiving a genuine effect of acupuncture, not an illusory one. We have to be careful not to deliver a message which could have a nocebo effect! It may be really working in CLK’s case. Acupuncture is not analogous to astrology or Elvis.
The study of pain is difficult! Important basic questions about methodology do not yet have a general consensus among researchers. Way too much certainty on this blog thread!
“it is possible that there is some subset of patients who do respond to acupuncture”
Sure, it is possible. But is there any evidence-based reason to think that it is probable? If more than a tiny minority of subjects responded, it should show up in the statistical analysis. And if it is only a tiny minority, would we want to recommend it to everyone knowing that most would not benefit? And would we want to apply the same reasoning to drugs and continue offering them to everyone when they failed controlled trials? It is possible that some subset of patients respond to homeopathy. Are we wrong to reject it? It is possible that some patients responded to bloodletting; were we wrong to reject that?
>>>Harriet Hall
What would you think of a doctor who practiced bloodletting today because he had found enough observational reasons to continue to be confident in the theory of the 4 humors?<<<
I think that doctor lives and works in a theoretical dimension very different from that of MDs working in modern hospitals, though some of them are promoting the integration of EBM with CAMs. That doctor practices bloodletting or phlebotomy "off label" (intended not for hemochromatosis and polycythemia) for patients consulting him. In case there were a real danger the right authority can legally stop this practice, as recently happened in the USA. Anyway acupuncture is not "tout court" identifiable with bloodletting or phlebotomy and researches in this field continue, seen that Elsevier and Wiley journals continue to publish them.
drmarcelli,
So, do you think any doctor should use any treatment that he believes is not dangerous? Would you condone a colleague who used bloodletting to balance the humors? If journals continued to publish bloodletting studies would that alone convince you that it was effective, even if the results were as inconsistent and questionable as the acupuncture studies?
If a doctor “lives and works in a theoretical dimension very different from that of MDs working in modern hospitals” don’t you think truth in advertising should require that he identify himself as a practitioner of magic rather than letting patients assume he is practicing the kind of medicine MDs practice?
“Of course, advocates of acupuncture have argued that it is worthwhile even if it only produces a placebo response; and that it is harmless, so it does not hurt to try it. Ernst et al. however, have shown that acupuncture is not harmless. While many of the reported adverse effects could be avoided by proper training in sterile precautions and anatomy, they correctly point out that even one avoidable adverse event is too many. With any treatment, we have to consider the risk/benefit ratio. If there is no benefit, any risk is too much. And there are other harms that they did not mention: time and money wasted, effective treatment delayed, unscientific thinking encouraged.
This paragraph assumes that we can know what benefits a placebo response can deliver in any given case. We have to know that before we can make the cost/risk/benefit calculation.
We don’t. We cannot. None of the studies referenced are designed to fully elicit placebo responses, yet even so, as Ed, points out, there are indications of very acceptable overall effect sizes from sham versions of acupuncture as compared to “not acupuncture” groups.
I agree that we don’t need further studies that essentially only tell us whether certain features of otherwise unsupportable Chinese medical theories are necessary to any “activity” of what can be quite complex ongoing “treatment” programs.
Yet further questions need to be answered for the sake of a fuller understanding of medicine in all its glory.
We need reassurance that the Newtonian “working better than placebo” yardstick is all we need to practice medicine with full effectiveness and humaneness, especially while our treatments are lacking in so many respects. We need also. I believe, need a better understanding of the role that CAM may be playing within medicine as it exists in the year 2011 before we decide upon policy and attitudes towards it.
>>>WilliamLawrenceUtridgeon 29 Mar 2011 at 2:59 pm
>>>Ten years ago I was far, far more credulous than I am today. It is not a crime or a problem to make mistakes or operate under mistaken assumptions.
I thank and applaud you.
>>>Wars are not scientific failures, they are political.
About this point I disagree. Some geniuses, “a posteriori” penitent, among them Einstein and also the Italian physicist Fermi “presented” the atomic research’s results to the US Army for a Science Based World War. Also today part of science and most wars are based upon fear and circumstance rather than evidence and fraternity. IMHO, of course.
>>> Harriet Hall
If a doctor “lives and works in a theoretical dimension very different from that of MDs working in modern hospitals” don’t you think truth in advertising should require that he identify himself as a practitioner of magic rather than letting patients assume he is practicing the kind of medicine MDs practice?<<<
Dear Harriet,
you are right, I agree that doctor should identify himself as a practitioner of magic, also simply to be correct and honest toward his patients!!!
Medicine is practiced at different positions with different degrees in freedom of therapy. For example:
- University scholar and researchers don't inject IV vitamin C in cancer. They could not do it also if they believe it works.
- Hospital MDs could inject IV vitamin C in cancer but generally they don't. When one "magically" believes it works can give vitamin C a shot.
- GP can inject IV vitamin C – and other non evidence based (but anyway legally sold) drugs in cancer. Only few of them do it, the so called orthomolecular MDs for exemple, so researches on Vitamin C in cancer continue.
But in this case the problem is not with Vitamin C or glutathione or possibly acupuncture, the problem is with cancer's complexity.
The same instinctive-irrational medical behavior happened in the past for diabetes type I, before insulin were discovered. People suffering from diabetes (MDs did not know the word "hyperglycemia") needed however to be helped because of the may problems diabetes comported. Obviously "magic" therapies were the rule. But patients who received the raw dandelion felt a little bit better because (today we know why) dandelion contains pharmacologically active compounds that lower blood sugar and combat infections (all group B vitamins and C).
I apologize for being a strongheaded MD. Now I go to sleep
29 Mar 2011 at 01:07 am in Italy
This thread is starting to get tapped out, but before it does, I continue to be curious about the implications for clinical practice in the management of chronic pain. I agree that a large percentage of patients with a meaningful difference between real and sham acupuncture should show up in the statistics, but the interesting thing still seems to me to be the difference between acupuncture and no acupuncture that does show up when a “no acupuncture” arm is included in the trials.
I have not heard Harriet acknowledge the problems inherent in pain studies; the lack of a general consensus on how to approach some critically important methodological issues (e.g., how to handle missing data, which endpoints to use and when to measure them, what to do when recruiting participants to a trial becomes problematic) mean to me that dogmatism about methods is unwarranted. True, RCTs give you the best possible insurance against biased comparisons between randomized groups, but they are still very difficult to interpret and translate into practice.
I think that many contributors to this thread think that the results of the acupuncture trials mean that physicians should tell their patients, “No, don’t even consider acupuncture; just keep taking your NSAIDs and going to PT. Just keep doing what you have been doing, because there is no difference between that and acupuncture. If you feel it has helped you, you are probably just responding to an illusion known as the placebo effect.” I do not think that the data warrant that interpretation.
EBM should not mean that the old patriarchy of “Doctor knows best” should be replaced with a new patriarchy of “Cochrane knows best.” (Come to think of it, most of us have seen some pretty flaky Cochrane Reviews.) The patient’s perception of reality should have strong standing in the decision-making process. Sometimes those perceptions seem to fit poorly with the data in the literature. That means that there needs to be flexibility and subtlety on the part of the practitioner.
Hell, if the interpretation of clinical trial data were unambiguous, there would be “an app for that,” and we could just do it all on our iPads.
Ed, I think what most contributors to this thread and those who understand SBM would not abandon is science and rational thought. RTCs are not difficult to interpret. The RTCs of acupuncture clearly show that real needles are not better than sham needles.
Substitution of “Doctor knows best” with “Quack with pre-scientific magical thinking knows best” is not an improvement.
If a patient with chronic pain is unsatisfied with treatments that have been tried so far, that patient should discuss treatment options with their doctor. Good and ethical doctors don’t prescribe placebos. Good and ethical doctors don’t lie to their patients. Good doctors don’t lie to themselves and pretend they know things for which there is no evidence.
@Ed Whitney,
“I have not heard Harriet acknowledge the problems inherent in pain studies”
I do acknowledge the problems inherent in pain studies, but I don’t think that is an excuse to ignore studies and go by beliefs rather than by evidence. I am certainly not a worshiper of Cochrane: on the contrary, I tried to explain that the existing body of acupuncture RCTs is unconvincing because of many factors including the problems inherent in pain studies. But RCTs, properly designed and properly interpreted, are still the best tool we have to avoid error.
“The patient’s perception of reality should have strong standing in the decision-making process.” Have you considered the full implications of that idea? If a patient is out of contact with reality, should the doctor participate in or encourage his delusions?
If you merely mean the patient’s perception of what seems to help, I can go along with that. I have no objection to people using acupuncture or anything else that they perceive as helping their symptoms. It would be cruel to confront them and say it doesn’t work. I only object to doctors who misrepresent the evidence and recommend placebos.
To be explained: Real acupuncture and sham acupuncture are the same, but sham acupuncture is better than none.
One interpretation: the trials with two arms, real and sham acupuncture, have been comparing two active interventions; the differences have been too small to bother with. Trials with three arms, one of which is “usual care,” show differences in which usual care is less effective.
Imagine that you were conducting a series of trials of SSRIs for depression, and in each of the trials, you compared two different SSRIs; perhaps you used Celexa as the drug you were investigating and you gave Prozac to the control group. After numerous trials, which were pooled with the best meta-analysis methods available, you conclude that Celexa does not work, since it is no better than sham Celexa. Unwittingly, you would be comparing two active interventions with small differences between them.
The descriptions of sham acupuncture require special scrutiny. Often, the sham acupuncture is done by acupuncturists who simulate many aspects of the real acupuncture (see Cherkin et al Arch Intern Med. 2009;169(9):858-866 for an example). The problem of imperfect placebos has been recognized as one of those damn persistent difficulties in pain literature. The sham acupuncture may have enough characteristics of the real thing to constitute an active intervention in itself.
No, the interpretation of these trials is not easy. The placebo which is not a placebo may be an active treatment. Sounds Taoist, doesn’t it: the placebo-not-placebo!
I meant, Harriet, that the patient who says that an intervention helped is to be believed; the patient is the final judge of what alleviated pain. The doctor should trust but verify. For example, the patient should be requesting fewer refills on meds, and should be up and about and doing things rather than lying in front of the TV. Forget about it being “cruel” to tell them that the intervention had not worked; it would be inaccurate.
RCTs are, as I said, the best way to avoid biased comparisons between groups; this is the one error against which they protecct us. They leave us open to all too many other errors of inference.
EdI meant, Harriet, that the patient who says that an intervention helped is to be believed; the patient is the final judge of what alleviated pain.
I can’t go along with this. We should nearly always accept the patient’s observation that the pain is better now. We are under no obligation to accept the patient’s understanding of why it is better.
Although, as Harriet says, we don’t necessarily have to treat a patient’s statement as we might a serious scientific contention.
@ Ed Whitney,
“the patient is the final judge of what alleviated pain”
No, the patient is only the final judge that the pain has been alleviated. The patient is not the final judge of what alleviated the the pain because he can’t rule out post hoc ergo propter hoc errors.
The Prozac/Celexa example is a false analogy, but if it were valid, then the sham/verum acupuncture trials would still show that acupuncture doesn’t work, since the sham procedures don’t involve the things that acupuncture theory rests on (meridians, acupoints, skin penetration, de qi sensations, etc.). The most parsimonious explanation is that a placebo response is being elicited by both verum and sham procedures. They both “work” in the sense that placebos “work.” If the sham procedures really “worked” beyond placebo, that would only prove that none of the rigamarole of acupuncture is necessary.
We have posted a response to Ernst et al on The Journal of Chinese Medicine’s website (http://www.jcm.co.uk/drum-tower-archive/article/a-response-to-acupuncture-does-it-alleviate-pain-and-are-there-serious-risks-a-review-of-reviews-by-e-ernst-myeong-soo-lee-and-tae-young-choi-pain-r-volume-152-issue-4-april-2011-1606/).
It is written by Mark Bovey, Research Co-Oridinator of the Acupuncture Research Resource Centre. He reports major problems with Ernst’s paper (“inadequate methodology that at times crosses the border into misrepresentation”) that clearly call its conclusions into question.
I was careless: the patient is the sole judge that the pain is better, not in attributing the relief to a specific cause. Big problem with vertebroplasty, where the effects are often as dramatic as effects get: the patient with a very painful fracture has the procedure, feels better almost immediately, and attributes the improvement to the procedure. This is a different topic, having to do with such things as patient selection, but it illustrates a major current dilemma for spine practice.
I almost immediately thought better of the Celexa/Prozac analogy. Maybe testing an SSRI with a MAOI as the control would be closer to the case. The MAOI does not selectively inhibit norepinephrine reuptake, and therefore does not involve the mechanism upon which the SSRI theory rests. MAOI would still be an imperfect placebo which could mask the truth, unless there were a third treatment arm with no treatment at all, which would probably show a difference with the third group coming off the worst. The acupuncture studies with two arms and the studies with three treatment arms seem to follow this pattern.
We need not speculate back and forth about the imperfect placebo problem. It seems that it is testable. Perhaps it has already been tested; I am not all that acquainted with the literature. But here we go: group A is randomized to acupuncture by a trained acupuncturist, who is told to place the needles according to usual practice standards. Group B is randomized to placebo acupuncture by the file clerk who is given an equal number of needles and told to poke the patient in several places with no further instruction. This would be a real placebo. You predict that group A and group B will have the same degree of pain relief. I predict that group A will do better than group B. If acupuncture is pure pseudoscience, the file clerk or maintenance man should do just as well as the acupuncturist.
Trouble will be getting the IRB to sign off on the design. No sense of adventure with those guys.
Ed, as I understand that trial has been done, but even more rigorously. They used trained acupuncturists for the sham leg so they would not put needles in the acupuncture points. There was no apparent difference in outcome.
Real needles in the right places, real needles in the wrong places, fake needles; they all give the same results. If needles are not necessary and the right places are not necessary, what is “necessary” about acupuncture?
On the same Journal of Chinese Medicine website I can conveniently purchase ear candles and gua sha products (skin-scraping, toxin- removing ‘therapy’) . Here are some interesting links on these two products:
http://www.hc-sc.gc.ca/hl-vs/iyh-vsv/med/ear-oreille-eng.php
http://youtubeskeptic.wordpress.com/2008/05/15/gua-sha/
http://www.cmjournal.org/content/5/1/5
The first link states …”both Canada and the United States have issued directives that ban the importing of ear candles.” (Health Canada)
Gua sha is, at best, a crude exfoliator and, at worst, a recipe for severe bruising.
Peter Deadman – credibility on a science blog is hard to attain, even harder if you are peddling quackery.
“They used trained acupuncturists for the sham leg…”
I can’t avoid picturing a highly ingenious placebo intervention here. Real needles and real acupuncturists, but a sham leg.
If you could prove that the place of the needles is irrelevant a lot would be gained. A test which would specifically test this (double blind) would go as follows.
The licenced acupuncturist (LA) indicates for the complaint several places, only one of which is correct, and the others are not even on known acupuncture points. S/he marks and numbers ten places with a circle. There is no talking and the circles are drawn in a random order. The LA leaves.
The test supervisor selects randomly a number from 1…10. An assistent who is properly instructed how to stick acupuncture needles in people, but with no knowledge about acupuncture points, enters the room, and puts the needle in the center of the randomly chosen circle and leaves only after the needle is taken out and all circles are erased.
Of course many variants of this scheme are possible. Another possibility is that one hundred true and one hundred false points are preselected, the LA selects from the one hundred true points a few treatment points, and unbeknownst to the LA a similar selection is made from the false points. Then the assistant gets after randomisation either the instruction to put needles in the correct points or in the falsle point. In this manner the LA doesn have to see the patient at al to mark the puncture sites.
When it has been established that the location of the points doesn’t matter, the training of acupuncturists can be simplified to a short course (four half days, would that be enough for someone with a knowledge of anatomy?) where only needs to be explained where you can’t stick the needles. All ‘Colleges for acupuncture’ and diplomas can be abolished. We don’t have colleges for bicycle riding either. Or do we (here in the Netherlands we don’t)?
Essentially what is off the table is all the “theory” accompanying acupuncture. What is still on the table is whether sticking people with needles has specific effects. The toothpick trial suggests that puncturing the skin is unnecessary. The dramatic ritual trial suggests any effect that might exist is enhanced by elaborate ritual. Therefore the best option seems to be nonpenetrating but pointy stimuli on safe locations with some sort of elaborate setup. Perhaps substitute the qi nonsense with some sort of pre-poking relaxation, or an accompanying gentle massage. That seems to be what the best evidence says – train people not to injure their patients (which would be much shorter and cheaper than the current program with it’s unnecessary vitalistic nonsense), make it a time-consuming and relaxing ritual (but one that doesn’t invoke magic and it’s accompanying corrosive effect on critical thinking and nonsense criticisms of real medicine) and make sure the equipment is safe to use (pointy enough to feel like something, blunt enough that it does not penetrate the skin).
That’s what the scientific understanding of acupuncture appears to distill down to currently.
It seems that we need to resort to some philosophy to advance this discussion, since the placebo concept is so much at the heart of the matter. Clearly, “placebo” is a noun, but it does not necessarily follow that placebos are things. Alfred North Whitehead discussed “the fallacy of misplaced concreteness” in Science and the Modern World. (It has been about 40 years since I read this book, so bear with me, but Whitehead intended it as a precaution against mistaking the abstract for the concrete.)
The discussion of placebo acupuncture has been proceeding as if it were analogous to a placebo in pharmacologic research. In those studies, placebos are things; they have mass and extension in space and they displace air when sitting in a room. Furthermore, you can have confidence that they are truly inert: a gelatin capsule has no remotely plausible mechanism of action in the body. They are perfect placebos. No problem with misplaced concreteness with them.
With “placebo” acupuncture, the analysis needs to be done on a different logical level. One must make do with imperfect placebos; the placebo in this case is not a thing but a complex interaction between the practitioner, the instruments, and the patient. You cannot count on anything in the equation being absolutely inert. Calling it a placebo does not make it analogous to a gelatin capsule.
The problem is this: these RCTs may be comparing two active interventions with competing (or similar) mechanisms of action. Acupuncture may yet have a mechanism of action, with placebo acupuncture having a distinct but physiologically real mechanism of action. The gate control theory comes to mind; this is a pretty elastic theory and comes up constantly in discussions like this. Whatever the phenomenon is, if you can’t explain it with the “gate control theory,” shoot, you just aren’t trying.
If sham acupuncture has a biological mechanism of action, using it for a “control” is not problematic in any way; using it as a “placebo” is full of problems. Just as an MAOI used as a control for an SSRI could obscure the effect of the latter through having a competing mechanism of action, sham acupuncture could obscure the effect of real acupuncture through a competing mechanism of action.
Richard Owen wrote a commentary in JAMA back in 1982 (vol. 247(18): 2533-4) on “Reader Bias.” He talks about things like “personal habit bias,” overrating or underrating a study that supports the reader’s habits. If we read that chocolate and red wine may be good for us, are we going to get all huffy and say, “Well, correlation is not causation” ? Hell no! We will say, “Very interesting study!”
There is “do something bias,” which favors a study which demonstrates effectiveness (common among clinicians, according to Owen), and its converse, “do nothing bias,” common among academics. I am personally inclined to favor the final one, “I am an epidemiologist bias,” which means repudiating any study containing any flaw in its design, analysis, or interpretation. Not having any dogs in this particular fight, I do not feel any temptation to “bankbook bias” or “territory bias.”
We bring our biases to the interpretation of studies. If you have long believed that acupuncture is bogus, the Ernst study nails it, so to speak. If you believe that there is something fascinating about acupuncture which has yet to be discovered, the study raises more questions than answers. If you are a true believer in acupuncture, you may see the study as missing the point, so to speak.
So, Ed essentially admits that there’s not specific effect of acupuncture and it’s all a sham to produce a placebo effect. Yet somehow avoids the conclusion that it’s therefore unjustifiable to practice…
Scott,
Whether it’s “justifiable” to practice is irrelevant: people do practice it.
There are other relevant questions though.
Whether it’s justifiable to offer acupuncture as a medical treatment in a medical context or by a medical doctor.
Whether it’s justifable to make acupuncture illegal.
Whether it’s justifiable to regulate acupuncture in some way, and if so, how.
Ed, what is your definition of “placebo”? The definition that I like is “a treatment that has positive health effects where those health effects are not mediated through pharmacology, surgery or other physical means”.
By my definition, psychotherapy is a placebo because it is a treatment that has positive effects where the effects are not mediated through pharmacology or physical means. The positive effects are mediated through communication with the psychotherapist.
Psychotherapy is an ethical placebo because the therapist is not lying to themselves or to the patient and patients do get better and psychotherapy is recognized as being effective under a standard of care.
Acupuncture is an unethical placebo because acupuncturists are lying to themselves and to the patients and it entails risks of infection not compensated for by actual therapeutic effects and acupuncture is not recognized as being effective under a standard of care.
Alison Cummins “Whether it’s justifiable to regulate acupuncture in some way, and if so, how.”
I’d vote for regulations requiring licenses and certification that requires safety and infection control education. This should include Government inspections that are paid for by acupuncture license fees. Insurance to cover customer injuries should also be required.
I believe this would be similar to tattoo, piercing salons, hair salons, cosmetology, etc.
I guess I was thinking that was the current situation in the U.S. But, I’ve never checked into it.
Doesn’t make sense to make it illegal. Consenting adults and all that.
Proper regulation of acupuncture must also include barring false claims. In particular, claiming that it has any specific effect, or that qi/meridians exist, or that the needles are relevant.
Placebo is relative to the hypothesis.
* Hypothesis = acupuncture points/meridians exist; placebo = points on the body that are not “true” acupuncture points for a condition
* Hypothesis = skin penetration is necessary; placebo = nonpenetrating needles; toothpicks
* Hypothesis = needling is inert; placebo = an inert or comparable treatment that has an equivalent dramatic intensity, attention from the practitioner, ritual and “exoticness” (all of which are known to enhance the placebo effect)
Each requires a different test, a different placebo, a different analysis. When people say “acupuncture doesn’t work”, they could be saying acupuncture points don’t exist, needling doesn’t manipulate qi, acupuncture doesn’t reduce acute pain, acupuncture doesn’t reduce chronic pain or acupuncture’s effectiveness is based solely on nonspecific effects. Part of the problem is the lack of distinction between acupuncture’s theory (qi, points and meridians), it’s practice (the very act of needling) and the nonspecific effects of practitioners (their bedside manner, the time and attention paid to patients, their nonverbal behaviour).
I believe the British Medical Acupuncture Association has adopted acupuncture (in the sense of using solid needles for the treatment of symptoms) and dropped the “theory” (http://www.medical-acupuncture.co.uk/Portals/0/Public%20Docs/BMASCodeofPracticev9Dec2009.pdf). Their emphasis is on scientific knowledge rather than tradition, and their training courses are only four days long.
Rather entertainingly, they appear to be headquartered at one of the British homeopathic hospitals.
For daedalus2u:
Ed, what is your definition of “placebo”? The definition that I like is “a treatment that has positive health effects where those health effects are not mediated through pharmacology, surgery or other physical means.”
The physical means, while interesting, are of secondary importance in my understanding of what EBM tries to do. I am mostly interested in interpreting comparisons between groups with different exposures or different levels of the same exposure. Many purported biological mechanisms, even if confirmed, are less important than clinically meaningful endpoints.
The most plausible alternative to the “acupuncture sucks” interpretation of most sham-controlled RCTs is the one I have tried to put forward: the studies are comparing active interventions with different mechanisms of action. Now, I may read about acupuncture enhancing generation of nitric oxide in a randomized double-blind crossover study (Tsuchiya et al, Anesth Analg 2007;104:301-7) and read that NO generation increased in the true but not the sham acupuncture intervention. What do I do with this? Do I interpret it as saying, “See! Acupuncture works after all! Told you so!” Not perzackly.
Rather, I look at their data and say, “Hmmm. Interesting. This may be a mechanism of beneficial effect, or it may be an artifact, or it may be a physiological response which has no clinical meaning.” Having a measureable physiological mechanism gives us one of those old Bradford Hill considerations for causality, namely biological plausibility. I am more concerned with strength of association (effect size) and with the factors that can obscure the association. Having a comparison intervention with an established mechanism of action (gate control theory), making it highly likely that it is an active intervention, is one of those factors that can cloud an association I am interested in.
For Scott: “Ed essentially admits that there’s not specific effect of acupuncture and it’s all a sham to produce a placebo effect. Yet somehow avoids the conclusion that it’s therefore unjustifiable to practice.”
What I essentially admit is that the specific effect of acupuncture is unproven. Plausible mechanisms exist, but I am not convinced of their relevance to any therapeutic effect of acupuncture. Does acupuncture increase NO synthase activity? Jolly good! Does that constitute proof of its clinical mechanism of action? See above for discussion.
When “the specific effect of acupuncture is unproven,” and as much effort has been put into looking for it as there has been, the correct conclusion (albeit provisional as in all of science) is that there is no meaningful specific effect.
“But… but… there MIGHT be!!!” is unavailing. And talking about “plausible mechanisms” for an effect where the evidence against said effect existing is so strong is the worst sort of Tooth Fairy science.
These are the statements of someone ideologically invested in the proposition that it works, and unwilling to consider facts and evidence that demonstrate that it does not.
I have just joined the Skeptics, and have been intrigued by your writing. I was drawn to your refutation of “Why we get fat” but found it to be flawed, and actually agree with a couple of commentators that you may not have read the book.
On accupuncture I think you are being fair in reviewing the study of the studies (what we used to call “meta-analytical”). I am a regular user of acupunture, and have a positive view of its efficacy. I really like your crack about needle proof vests. I think there are huge hurdles to overcome before acupuncture can be studied, and only a few of them are the limits of the western scientific method as it has developed. The incredible variety among practitioners, training and treatment methods etc. contributes more that the lack of understanding of the priciples of acupuncture that I have found in the very few studies I have read.
There is one glaring error in yout work that I feel I must comment on. I also hope that you will indulge me in reading further and letting me share an anectdote about acupuncture with you as well. So, to the error:
“In summary, Ernst et al. have shown that the evidence for efficacy of acupuncture for the treatment of pain is questionable, to say the least, and of particular concern is that its use can be dangerous. If the 57 systematic reviews they surveyed had been for a prescription drug and a similar list of serious adverse effects had been reported for that drug, we would hesitate to prescribe that drug. Is there any reason not to hold acupuncture to the same standards?”
I think this is a bit niave. I remember reading an article last year about the FDA approving a new and dangerous drug. The FDA said that although the drug fared slightly worse than the placebo effect, it would be wrong to deny the drug to the small portion it would help. When you mentioned in the paragraph above that one that placebo’s were un-ethical it reminded me of a study that showed the placebo effect still occured even when patients were told they were getting a placebo! So yes, there is a lot that science can’t (yet) exlplain about medicine. Or about space for that matter.
If you have bourne with me this long I hope you will tolerate an anecdote. I fitst went to acupuncture for foot pain, and my acupuncture wanted to know why I wasn’t as concerned about my blood pressure and my asthma ( a life threatening condition, especially in the wilderness!). In discussing this with my sister she mentioned that she too had her running curtailed by pain. When my acupuncturist examined her he told her that her knee pain was being caused by a severley inflamed galbladder, and that she should see her physician right away, there was nothing he could do.
My sister called the doctor that Friday, and her doctor laughed. He told her that he would love to see her just to prove the acupuncturist wrong. He scheduled her for an ultrasound two weeks out. That Sunday she spent the night in the ER due to a severly inflamed galbladder. It was so bad that the only treatment option they offered her was surgery, done about 48 hours later. That is moving pretty quickly for our HMO!
I appreciate your writing, and would love to write as well as you do.
Scott “Proper regulation of acupuncture must also include barring false claims. In particular, claiming that it has any specific effect, or that qi/meridians exist, or that the needles are relevant.”
Clearly, I’m not a lawyer, but I would guess that you could handle it somewhat like the FDA does with a statement similar to “this is not proven to have any benefit” required.
I don’t see how you could make claiming that qi exists off limits, when it’s easy enough to offer the defense “I was speaking metaphorically, about a mental process or visualization process, etc”
but what do I know.
Ed, you brought up nitric oxide, not me.
http://daedalus2u.blogspot.com/2007/04/placebo-and-nocebo-effects.html
The physiology of the placebo effect is mediated through neurogenic production of nitric oxide.
It is the neurogenic triggering of the end of the “fight or flight” state. It tells physiology to re-allocate resources to healing and tissue repair and away from keeping them ready for “fight or flight”.
If you don’t have a clear understanding of what a placebo is, just say so. If you do have a clear understanding then you should be able to define it, and your definition needs to be consistent.
If something is equivalent to a placebo, then it is a placebo. Real therapies can have placebo-type effects. Part of the therapeutic response to a treatment is due to the placebo effect. That is why “real” therapies are tested against placebos or against known effective treatments. If a treatment only has therapeutic effects through the placebo effect, then it will have therapeutic effects no better than a good placebo.
What I want to do is invoke the “placebo effect” pharmacologically by raising the NO/NOx status of people. Invoking the placebo effect via pharmacology will work better than any ritual.
Michele,
I should have been more precise, and instead of “claiming that qi/meridians exist”, referred to the more exact “claiming that inserting needles into specific points along meridians improves the flow of qi, thereby producing beneficial effects.”
For daedalus2u:
I am not much good at interpreting studies with surrogate endpoints like NO synthase activation, and usually ignore them, following the GRADE Initiative’s approach to evidence, in which the “indirectness” of an outcome results in its being downgraded in support of an intervention. Therefore I cannot interpret why that Tsuchiya study found NO increased in the acupuncture plasma but not the sham acupuncture plasma. When is a placebo not a placebo?
Personally, I never had acupuncture (don’t much believe in it). When I fell recently and had excruciating low back pain, I did not bother with placebo pills or needles. I used placebo Feldenkrais exercises, with lasting beneficial effect. This was due to a bias that Owen in his JAMA paper did not mention, namely “cheapskate bias.” Also, I never take large, oblong, chalky placebos; they do absolutely nothing. I take only small, round, colored placebos (red works best).
One of the puzzling things about the Ernst article was his puzzlement over acupuncture working for some conditions but not for others. This is only a problem under an assumed “panacea” model of acupuncture. I am unaware of any intervention of proved effectiveness that is good for what ails you. The point of studying an intervention is to figure out what, specifically, it is good for.
It so happens that the conditions I am most interested in, namely low back pain, neck pain, and osteoarthritis, showed overall superiority of acupuncture over placebo acupuncture; other conditions, including rheumatoid arthritis, cancer pain, fibromyalgia, and shoulder pain, were not shown to respond to acupuncture over placebo. The stuff I need to worry about was on balance either inconclusive or positive for the comparison between the real and the sham interventions. Table 1, ignoring the poor quality systematic reviews, had conclusions like “effectively relives ( I am sure hoping this is a typo for ‘relieves’) back pain,” “limited evidence to support,” and “sham-controlled trials show statistically significant benefit.” The latter was in one of the ‘excellent” systematic reviews of 16 trials for peripheral joint osteoarthritis, one of the conditions I have to focus on.
For Scott:
The time has come: I gotta ask you whether you have actually read the Ernst article. I have been wrestling with it for a couple of days (it was just published), and I do not interpret it the way most of this thread seems to be interpreting it. For the conditions mentioned above, there seems to be no reason to deprecate acupuncture.
Ernst does say that most of the serious adverse effects came from Asia, where (1) acupuncture is widely practiced, and (2) Asian therapists may be poorly trained in Western stuff like sterile technique. I still look very seriously at adverse events, and, as previously stipulated, think that it should never be practiced by dummkopfen.
.”The toothpick trial suggests that puncturing the skin is unnecessary.”
Toothpicks might not work as well in practice, and not only through the introduction of a banal element into acupuncture proceedings.
If the endorphin release is a major part of the action of acupuncture, it is quite likely that the thought that needles are going to be inserted into the skin will be an adequate trigger. If so, the toothpick studies may actually be comparing like with like.
The soothing effects of the endorphins and possibly other neurotransmitters may trigger a cascade of other responses leading to a reevaluation of pain levels or reducing pain perception though other mechanisms.
Now, I don’t know how valid this will prove, but neither does anyone else. We have to factor the possibility in to some degree or other.
Ernst seems to embrace “”classical medical skepticism” (of which I was a founding member), for whom a toothpick study is seen as somehow completely discrediting the whole field of acupuncture including any claimed benefits. It is thought that the most rigorous levels of scientific inquiry are being applied, but biases are leaking through, nevertheless.
The toothpick result only confirms that acupuncture has equivalent therapeutic effects to toothpicks.
To continue to accept acupuncture, you have to change your physiology paradigm such that whatever the unknown physiological mechanism of acupuncture is, toothpicks activate it and needles in random places activate it and that activation occurs in ways that are indistinguishable from the activation by acupuncture.
This result does not provide evidence for qi flowing in channels underneath the skin because the stimulation works where the channels are not supposed to be, and works without breaking the skin. The toothpicks and random places demonstrates that affecting the flow of qi in channels cannot be the mechanism by which acupuncture works because toothpicks can’t affect qi (unless qi is even more magical than has been supposed).
That toothpicks and random spots are just now found to be equivalent to standard acupuncture, implies that all prior acupuncture “experts” didn’t know that they were equivalent, and all the “theory” that they developed to justify and predict differences is simply wrong. Acupuncture isn’t based on any “wisdom of the ancients”, it is based on the “foolishness of the ancients” and the “Emperor’s new clothes effect” on every “expert” who passed down that “wisdom” without testing it.
The alternative is to acknowledge that since acupuncture works the same as a placebo, acupuncture is a placebo. If you have nothing better to offer someone, say you are stranded on Gilligan’s Island and need to treat someone for a broken leg, maybe a placebo would be ok. I would opt for kind words and gentle hand-holding myself.
But the studies with three arms (not only two), with usual care as the comparison–the sham acupuncture should not be better than usual care. Let’s say that usual care means NSAIDS, physical therapy, maybe a nerve stabilizer or a tricyclic, not an uncommon mix of things for many kinds of chronic pain. Shouldn’t all those things have a placebo effect, too? A study with three treatment groups, all of them with placebo effects, ought to be equal in outcome, assuming that a placebo is a placebo is a placebo. Why should the placebo effect of acupuncture or sham acupuncture out-perform the placebo effect of usual care? If acupuncture is a placebo, why isn’t usual care just as good a placebo?
For the conditions of back pain, neck pain, and osteoarthritis, the available data in Ernst 2011 support acupuncture having more than a placebo effect. Hopton 2010 (cited here yesterday), having reached similar but not identical conclusions as Ernst, has a most pertinent question: “is it now time to shift research priorities away from asking placebo-related questions and shift toward asking more practical questions about whether the overall benefit is clinically meaningful and cost-effective?” Then questions such as, “Do acupuncture patients start to reduce their opioid consumption?” are worth exploring, and are quite practical–the kinds of question Hopton probably has in mind.
Harriet mentioned that RCTs are our best protection against error; I repeat that they are our best protection against one kind of error: biased comparison of group outcomes. They do not protect us from the error of asking the wrong questions, or from neglecting to ask the right questions.
“Comparing like with like” insofar as that the endorphin cascade you speak of would be the same between the two modalities. However, you seem to be asserting that “sticking a needle in someone” is the same as “not sticking a needle in someone” – and that is patently false.
The fact that you don’t know “how valid this will prove” is exactly the problem. I know how valid it will prove – invalid! This does in fact show us that acupuncture does not work.
To try and assert we are comparing “like with like” here is just asinine. Just as you may tense up and have a catecholamine rush at someone throwing a very convincing punch at your face, but stopping just short as well as if they actually connected and decked you, I think you might agree that there is a distinct difference between getting punched in the face or not. But by your logic my “punch you in the face” versus “almost punch you in the face” treatment would be comparing “like with like” since the cascade of catecholamines would be the same and indeed it would be the anticipation and expectation causing said cascade. But if you would like to assert that this is still a “comparing like with like” scenario perhaps you’d like to sign up for a randomized trial of the two “equivalent” face-punching modalities.
Yes, but it is your biases, not Ernst nor ours that are leaking through.
A good question… answered by:
Based upon the mountain of equivocal data the question should be “Why is there a lack of additional placebo effect in the “usual care” treatment arm.”
The fact of the matter is there are myriad answers: different placebos affect people differently, perhaps the people seeking acupuncture also used some usual care, perhaps the “usual care” isn’t actually very good, and of course, since we are asserting that the placebo effect of acupuncture is mediated through the ritualistic interaction, one could expect that to have a higher placebo effect than the “usual” prescription of some pills. Also, you must remember expectation people with chronic pain have been using “usual care” with limited success… then they hear about “something new” and give it a go. The expectation is that it may work, especially with such equivocal results being given credulity. If you notice, most of the studies show the effects wane after time. If you look at studies that look at objective measures of function there is no change, despite reported decreases in pain. And there is that well known fact that patients will often tell their practitioner what they think the practitioner wants to hear.
But the best part? None of that matters! The point is when you have a treatment, whose very premise is that needles inserted into meridians modify chi flow and thus generates an effect, and you have no evidence (or even plausibility!!) that meridians or chi exist in the first place (and indeed historical evidence to offer an explanation for their location and existence without invoking the magical thinking of TCM) and a mountain of data that shows equivocal results, AND studies showing that the very core premise of needle insertion AND specific meridian points is nonsensical (the sham acupuncture) you are done! Wham, bam, thank you ma’am – it is clear that there is no (remotely significant) effect beyond placebo to acupuncture.
Think about it – I give you aspirin to swallow and tell you that it will relieve your pain and thin your blood by inhibiting COX which will then inhibit synthesis of pain mediating prostaglandins and thromboxane, which is necessary for platelet adhesion. So you take it and have more effect than, say, bloodletting which has been the “usual care.” Then, along comes a molecular biologist and finds out that COX doesn’t exist, that prostaglandins don’t mediate pain, and that thromboxane isn’t necessary for platelet aggregation. Then, we do studies and show that whether you ingest the aspirin or if I just wave a fistful of pills over you, the effect is the same – but both have more effect than “usual care.” And further study shows us that all the results with aspirin have been equivocal with only a couple showing some odd mildly positive results. And on top of that, someone like Ben Kavoussi shows us the history of it all and demonstrates that the original purveyors of aspirin actually called it an offering to the COX god and that the “prostaglandins” and “thromboxane” were just permutations of old understanding of simple anatomy superimposed on it (meridian lines and veins).
But wait! There is more effect than “usual care!” We should clearly keep trying to find out how aspirin works!
Can anyone say “Tooth Fairy Science?”
– — assuming that a placebo is a placebo is a placebo
Not tenable. Plenty of evidence to the contrary — that the strangth of placebo responses are influenced by both characteristics of the placebo treatments and other aspects of the therapeutic interaction.
is that an argument? I even clearly stated that not all placebos are created equal. Ben Goldacre has a lot to say on that as well. What is your point pmoran? You seem to be lacking one (no pun intended).
Are you trying to advocate for the intentional use of placebo since one placebo may be better than another placebo? Ever heard of “medical ethics?”
Pay attention, Nybgrus! That was a response to the quotation from Ed, as you should have recognised, having responded to it yourself in a similar way, in an overlapping post.
apologies then, pmoran.
Isn’t “acupuncture” a distraction here?
If the point is to induce an objectively greater placebo response, then shouldn’t any further medical research be invested in how to induce that heightened response safely and ethically? (Massage, for instance?)
I haven’t read any of the papers, so please correct me where I’m wrong. If the heightened placebo response induced by [sham] acupuncture is restricted to temporarily reducing reported pain and does not actually improve functional outcomes, then a parsimonious explanation is that hope that the pain will end makes pain more bearable. When it turns out that a treatment isn’t working, hope is dashed and pain tolerance drops again.
For a doctor to string patients along by offering them false hope would be unethical.
For chronic pain patients to flit from one practitioner to another in hopes that maybe this will help… No? Then maybe that… That’s understandable, and perhaps a doctor should stand back and not interfere. But to lie to a patient that “this will help” because hope feels good (until it is proven unfounded) is not ethical and I would be very distressed to think that medical professional organizations endorsed this practice.
Don’t have access, but any single article is irrelevant. The overall evidence base was overwhelming years ago.
Remember, any completely ineffective treatment, if tried enough times, will produce some positive results. Check out Ionnidis’ work; it applies perfectly to the acupuncture literature.
Eh it’s the S&M play that makes a lot of people feel better for a time thanks to some medical interaction. The more intense the adversive, the more trust and surrender for the bottom. At some point the bottom dissociates, which is like getting a hit of morphine. The pain is still there but it seems distant and less intense.
I can’t quote anything from PubMed, but I’ve talked to people involved with alternative lifestyles.
I’ve also watched little kids play doctor and hypnotize-U (“you are now under my power”), and WWII movies. White coats and black uniforms can become discriminative stimuli or fetishes that trigger the dissociative state.
Most humans are fascists in their deepest darkest hearts. We, the old farts, really hate fascism because of what our grandparents lived through. But the next generation is clueless. They’ve no idea of the addictive power of social rank or grades. They’ve no idea what happens when mysticism –a shared representation of an idealized human mind– is combined with political power.
The traumatized dissociate and therefore long for guidance. The worst thing you can do, IMHO, is to take on the role of shaman for these people. That’s what they want. But it’s not good for them. It simply reinforces their habit of tripping out all the time.
My reason for asking if people had read the Ernst article was that it is being taken as saying that acupuncture is bogus when it does not say that acupuncture is bogus.
Ernst reported that acupuncture was superior to placebo for low back pain and osteoarthritis, but not for other conditions, including sciatica, cancer pain, and shoulder pain.
These “contradictions and doubts” are puzzling to Ernst because “there is no plausible reason why acupuncture should reduce pain in some conditions while failing to work in many others.”
They are puzzling to me as well, but perhaps not as deeply. It is a bit puzzling that NSAIDS should work well for most kinds of pain but not very well for neuropathic pain. It is not very realistic to expect any intervention to work for everything. Sometimes you can construct a plausible reason for different effects in different pain situations: why epidural steroids could be effective for sciatica but not so much for back pain, as an example. Sometimes you can’t. Often the explanations are constructed after the fact, which does not make them bogus; it only means that they have an element of speculation and may not be completely correct. Nature does the darndest things.
About the ethics of placebos: some national practice guidelines formalize distinctions that are made intuitively by clinicians in daily practice: there are recommendations, suggestions, and options. You would recommend smoking cessation to everyone, suggest water aerobics for others, and be willing to discuss acupuncture as an option with others. The circumstances of the patient would have a small impact on recommendations, a moderate impact on suggestions, and a large impact on options. The level of certainty about the effectiveness of the intervention also plays a role; the more certain the effect, the more it moves from option to suggestion to recommendation. It may be ethically questionable to recommend something with a large placebo effect, defensible to suggest it, and justifiable to discuss it as an option.
There are potent modifiers of the placebo effect, assuming that this is related to an alignment of the patient’s expectations and conflict-free goals. Many RCTs do not enroll patients if they have pending litigation for personal injury or disability compensation. Why? Well, part of the reason is that placebos do not work for them; getting all well can create a conflict of interest when there is a lawyer standing by working on a contingency basis for a large settlement. It would be foolish to discuss acupuncture as an option in such circumstances. If, on the other hand, the patient’s daughter is getting married in a month and he wants to walk her down the aisle, and has been suffering from back pain that has been responding poorly to usual care, and if he heard that his wife’s hairdresser’s astrologer tried acupuncture with success, and wonders if it may work for him, that is a different situation and a different response could be appropriate.
@Ed Whitney,
” It is a bit puzzling that NSAIDS should work well for most kinds of pain but not very well for neuropathic pain.”
It is far more puzzling that acupuncture would work for musculoskeletal pain in one part of the body but not in another. If the studies on NSAIDS were as inconsistent as the acupuncture studies and showed that they relieved musculoskeletal neck pain but not musculoskeletal shoulder pain, wouldn’t you start to wonder whether they really worked at all?
Keep in mind the problem of multiple comparisons! When you try something for enough indications, it will seem to work for some of them. Going from there to try and figure out why it works, instead of considering that those are statistical anomalies, is wrong-headed.
seorsa,
” I was drawn to your refutation of “Why we get fat” but found it to be flawed, and actually agree with a couple of commentators that you may not have read the book.”
For crying out loud! That’s an outright insult. I read the book! I couldn’t have commented on its details the way I did if I hadn’t read it. And if you found my analysis flawed, you should have commented on that thread and explained what you thought the flaws were.
“The FDA said that although the drug fared slightly worse than the placebo effect, it would be wrong to deny the drug to the small portion it would help.”
I don’t know what drug you’re talking about, but the FDA could not have said that unless it had solid evidence that the drug was effective for a subset of patients. Evidence much more credible than the acupuncture evidence. Do you have any evidence that some subset of people can be identified who benefit from acupuncture?
Yes, sometimes people get a placebo response even after being told they are getting a placebo, but that doesn’t change the fact that placebos are unethical. Your comment about “science doesn’t know everything” is a truism that doesn’t contribute anything to the discussion.
@ Harriet:
Continuing the battle of the analogies, I would indeed expect that NSAIDS should work on any musculoskeletal pain sharing similar pathophysiology. NSAIDS, being systemically absorbed, ought to go to every part of the body.
Acupuncture may have a general relaxing effect in the hands of a good practitioner, but the action (by whatever mechanism) may be more local than systemic. I, like most contributors to this thread, do not buy into models of “meridians” and “qi” that classical acupuncture uses as its explanatory principles. However, investigations into relationships between “meridians” and connective tissue planes, neurovascular bundles, and the like, are beyond my ability to assess critically, but at least they are based on general principles in which I have some familiarity, even if the details are long forgotten.
If the mechanism of action is more local than systemic, then it would be surprising, but not quite shocking, to find that different musculoskeletal conditions responded differently to acupuncture interventions. It would, in any case, be less shocking than if NSAIDS acted only in some parts of the body and not in others for similar kinds of pain.
I actually doubt that you and I are all that far apart on acupuncture; I did think that your editorial was more negative on acupuncture than the Ernst meta-meta- analysis concluded. Acupuncture seems to have a limited, secondary role (not a first-line choice) for some common musculoskeletal pain conditions. There are contrasting points of view between people who emphasize efficacy (with specificity of mechanisms and strict control of experimental conditions) and effectiveness (with emphasis on what works for patients in the unstructured settings of community practice). I say contrasting, not contradictory; the differences are more of emphasis than substance. Most contributors to this thread are oriented to efficacy, as a perusal of the thread will confirm.
Adverse effects, naturally, are of concern, but the risks of acupuncture are not all that great in a country where aseptic technique is nearly universal and training requirements meet some good standards. (There is a case report in press in Spine in which a cervical hematoma from dry needling resulted in quadriparesis; that was reported in Korea but could happen here.) Your editorial emphasized the numerator of 95 adverse events; I wanted the denominator as well as the potential confounding factors (setting and training of practitioners).
Bottom line: I do not think that acupuncture has been shown to be just a placebo, and I am certain that it has been shown not to be a panacea. It can be good, but not good for what ails you.
Ed Whitney said “I do not think that acupuncture has been shown to be just a placebo.”
I don’t either. I just think it has not been shown to be more than a placebo. And the longer we go without such evidence, the less likely that it will be found. As Scott commented above: “But… but… there MIGHT be!!!” is unavailing.”
@Ed Whitney,
“I actually doubt that you and I are all that far apart on acupuncture;”
It is said that the devil is in the details.
Methanol and ethanol are not all that far apart in chemical structure, but they have radically different outcomes if you drink them.
Dr. Hall beat me to it.
Swing and a miss. The point of Dr. Hall’s analogy was that pain arising from your calf muscle, neck muscle, bicep, etc should all respond to the same analgesia since there is no difference between the mechanisms for pain perception based on location. Neuropathic pain is different because it does not involve the same mechanism for generation and perception of pain as muscular pain. This is why when an NSAID works for MSK pain but not neuropathic pain, we aren’t surprised. You miss the point entirely when you compare the “systemic” effect of NSAIDs vs the “local” effect of acupuncture. If you injected an NSAID locally into a muscle it would work in any MSK pain to which it was localized. The fact that acupuncture works for MSK pain only in certain locations of the bodies is extremely surprising surprising enough that it makes it very, very likely that those positive results are false.
You also miss the basic point of medical ethics vis-a-vis risk/benefit analysis. If you have a treatment with a definable, measurable, and mechanistically explained benefit then you would look at the number of adverse events divided by the number treated (and the number helped). If that ratio becomes too high, then even though the treatment is actually beneficial it is no longer given. However, when you have a treatment that shows nothing better than palcebo, has essentially no definable or mechanistically explainable benefit then the denominator becomes pointless. Any adverse event, even just one, tips the scale. You may think “95 events if the denominator is 100,000 people being treated with acupuncture means .095% of people have an adverse event. That’s vastly better than aspirin! This is a great treatment” But the real calculation is “that is 95 events with a placebo – that is unacceptable.” Try and think about this – if I had two pills, one sugar and one drug X. And I did a study where drug X had the exact same effect as the sugar pill (real vs. sham acupuncture) but drug X had a .095% adverse effect rate (including a death and maybe some quadriplegia) but the sugar pill had exactly 0 adverse events, would you still give drug X? If you are a patient and knew this, would you want me as a physician to give you drug X? If you answer yes to either of these questions, you are truly lost on the topic.
@ Karlwithakay
Methanol and ethanol differ by one letter…reminds me of the sixth grade when us kids went around the playground at recess singing, “We never stagger, we never fall, we sober up on wood alcohol…” We didn’t know what it was, but it had been handed down by the big kids and that was good enough for us.
@ Harriet et al.
Hopton 2010 concluded that the accumulating evidence from recent reviews suggests that acupuncture is “more than a placebo” for commonly occurring chronic pain conditions. And Ernst 2011 says a most interesting thing in his discussion, namely, “The majority of the early reviews arrived at negative conclusions, while the majority of the 57 recent reviews were positive.”
That last sentence sounds almost innocuous, but this is a rather remarkable state of affairs. Almost at the same time that John Ioannidis published his widely quoted 2005 article on why most published research findings are false, he published a paper on contradicted and initially stronger effects in widely cited clinical research. This paper discussed a now widely appreciated phenomenon: studies that show large treatment effects tend to get published early, while later studies show smaller treatment effects. Carotid endarterectomy, immediate angioplasty, coronary artery stents, showed large reductions in morbidity and mortality when first published; later studies eroded the initial effects and led to considerable erosion of the early estimates of treatment effect.
So it is highly interesting to see the trend in acupuncture going in the other direction. Early studies tending to be negative, more recent studies trending positive…what in the world could be going on?
Ernst 2011 is not the final nail in the coffin that will bury acupuncture once and for all. Many are tempted to conclude the discussion, close the books on acupuncture, and move on. But just as premature closure of the epiphyses stunts growth, premature closure of the acupuncture question stunts progress in pain research.
@Ed Whitney,
After the “early studies negative/later studies positive” comment, Ernst et al. go on to say
“These findings should be seen in the light of recent results from
high-quality randomized controlled trials.” with a fuller explanation later in that paragraph about exactly “what in the world could be going on.”
The article will not spell the end of acupuncture because of people like you who keep saying “But… but… there MIGHT be!!!” We are not dealing with a straightforward question of pain research, but with a belief system amounting almost to a religion. If acupuncture were a new drug that people had no emotional investment in, it would have been given up by now.
@Ed Whitney
Ethanol and methanol differ by a little bit more than just one letter.
CH3–CH2–OH vs CH3-OH There’s an entire added methylene (CH2) group in ethanol, so three letters or two letters and a number.
“we sober up on wood alcohol”
Permanent blindness would be somewhat sobering.
The jury came in on acupuncture some time ago. It’s not pining for the fjords; it’s dead. If people wouldn’t keep nailing it to the perch, it’d be pushing up the daises.
Ernst 2011 is not the final nail on the coffin; it’s not even the icing on the cake. It’s the aperitif your drink after you finish the cake and the flowers you put on the grave of the buried coffin.
However no amount of negative research will convince some people that it is an un-parrot.
Jonathan Swift said, “It is useless to attempt to reason a man out of a thing he was never reasoned into”
Is this actually from ‘Letter to a young clergyman’,
http://www.readprint.com/chapter-49128/Writings-on-Religion-and-the-Church-Vol-I-Jonathan-Swift/14 ?
Harriet : The article will not spell the end of acupuncture because of people like you who keep saying “But… but… there MIGHT be!!!”
That’s not quite true. While I also disagree with Ed’s interpretation of the evidence, acupuncture will persist because a lot of people think that it helps them and because there is no clear practical medical significance to the phrase “works no better than sham/placebo”, with more complex interventions like acupuncture.
A simple example: why would not stimulation at that mystical P5 acupuncture point distract some patients from that growing feeling of nausea after surgery, sufficient to have a useful effect on post-op nausea? Why should that ever be regarded as a bad thing, and not something not worth trying in preference to adding on more drugs?
Why would a cotton wool plug in the left ear (an example I provided previously) not distract some people from their sea sickness, and quicker than any pill?
If a method like acupuncture has distractant, and counter-irritant potential, along with enabling periods of enforced relaxation, and opportunities for ongoing socio-medical interactions, as well as perhaps being a good provoker of endorphin release, why would it be dismissed as “only ” a placebo? WTF does that mean for practical medicine?
I don’t want to get into the tricky business of defining just what “placebo” means in all possible settings. I would like to draw attention to a problem in communication, with there not being a sufficiently well differentiated vocabulary for these matters.
“Placebo” has a negative taint that is difficult to shake off. Once something acquires that label it is automatically ignored by the “scientific” skeptic, even when it has plausible medical potential and the sham versions i.e. those that have been merely divorced from TCM theory have quite significant effect sizes within clinical studies.
pmoran,
I agree that placebos have plausible medical potential. I’d love to find a way to take advantage of that potential without lying to patients. Can you suggest a way to do that?
Harriet, how’s this:
“Acupuncture might help –we don’t know why but it could simply be placebo. Are you ok with that?”
@ tanha,
That’s better than claiming it works, but I still wouldn’t feel right about recommending acupuncture just as I wouldn’t feel right about recommending homeopathy or prayer or therapeutic touch. I wish there was a way to elicit the placebo response while offering the best available evidence-based medicine or while acknowledging that evidence-based medicine had nothing to offer but still finding a way to offer comfort.
Because, pmoran, once it is identified as placebo, regardless of whether you want to quibble over whether that means it “works” or not, it is unethical to recommend or use it! If we had no clue that acupuncture was almost certainly only a placebo effect then your arguments would hold water. However, in using the best and most current data, we can say that most likely acupuncture is only a placebo effect ergo WTF that means for practical medicine is we cannot in good consience use it. Pretty simple.
Because you would know that the purpose is a distraction device and tell the patient that. And, also importantly, you would also know the level of risk with that is very low. If you told the patient that the cotton wool would channel chakras into the ear and heighten the vestibulocochlear acuity to compensate better for motion sickness and that this was based on “ancient wisdom” that would be employing a placebo effect under false pretenses and be unethical. If you further went on to claim that cotton wool rubbed on your knee would active the CW-6 point and change the chakras there to decrease knee pain, well… then you’d have cotton wool acupuncture! And that is the point that people seem to be willfully missing here.
Harriet, you raise someone’s NO/NOx level and you will trigger the placebo effect pharmacologically. You have to raise the NO/NOx level via physological means, and I know of only two ways to do that, meditation (which not everyone can do and so it can’t work on everyone) and my bacteria applied topically.
I appreciate I don’t have any “data” to show it yet, just “theory”, but it is a lot of “theory”.
Sorry squire, I’ve had a look ’round the back of the shop, and uh, we’re right out of parrots.
I was careless again, and Karl caught me dead to rights; I should have said that “ethanol” and “methanol” differ by one letter, and that ethanol and methanol differ by a carbon and two hydrogens. “The use-mention distinction” isn’t “all that important most of the time,” but sometimes, as in the present example, it does matter.
When I see phrases such as “people like you,” followed by allusions to “a belief system amounting to a religion,” with quotes from Jonathan Swift about the futility of reasoning a man out of a thing he was never reasoned into, I have to be very careful not to read too much into the thing, and not to infer that anyone thinks that I have a belief in acupuncture amounting to a religion. We all depend greatly in these websites, where we do not know one another, on the ability of supposedly educated people to read. We may not have the chance to have a beer together in the Rose Garden at the White House. So we have to read one another carefully.
The only reference I have knowingly made as to my personal convictions in the matter was when I confessed I had never tried acupuncture, not much believing in it. Thanks to pmoran for disagreeing with my interpretation of the literature, which is what this is all about. Much appreciated!
The studies comparing acupuncture, sham acupuncture, and no acupuncture have consistently shown small differences between the first two and appreciable, clinically relevant differences between the last two. If this had happened only once or twice, it would not be of interest, but when it happens repeatedly, I get fascinated and see it as an opportunity to learn something new.
Having the “no acupuncture” arm in the RCT helps one to separate out many factors that usually get lumped together under the term “placebo effect.” These include the natural history of the condition, regression to the mean, the effect of being under observation in a clinical trial, and, if the follow-up assessments are done correctly, the potential for reporting and other biases. Having that third arm in the trial clarifies things that may otherwise be mixed in together.
Some of the literature on “placebos” distinguishes the effects of an intervention into two components: the specific effects, attributable to the theoretical mechanism of action of the supposed active agent, and the nonspecific effects, which are triggered by expectations, interactions between provider and patient, the context in which the intervention takes place (pending personal injury litigation is different from anticipation of an upcoming vacation), and similar vaguely characterized factors. In this setting, they can be approximately calculated by taking the difference between the sham acupuncture and usual care. The nonspecific effects repeatedly turn out to be a majority of the total effects (calculated by the difference between real acupuncture and usual care). If you do not care to tell patients that acupuncture is a “placebo” (which is a nonspecific term, when you get right down to it), you can tell them that its effects are nonspecific. It is just as accurate and less pejorative to some ears.
The size of these nonspecific effects in relation to specific effects is greatly annoying to some observers, and greatly fascinating to other observers. In general, patients, their families, and their clinicians want to observe the total effects, and are only mildly inconvenienced if the bulk of those are nonspecific. The persistence of acupuncture in the health care system will depend on the total effects, not on the specific effects, which means it is likely here to stay. Whether to be fascinated or vexed by the phenomenon is a matter of individual choice.
There seems to be an assumption that you cannot tell a patient that they are receiving a placebo, if you wish to observe a placebo effect. While ritual and deception may increase the placebo effect, telling a patient that they are receiving a placebo will not eliminate observed placebo effects.
Harriet pmoran,
I agree that placebos have plausible medical potential. I’d love to find a way to take advantage of that potential without lying to patients. Can you suggest a way to do that?
I would be lying to a patient if I said that Laetrile would cure their cancer. Characterizing the use of placebo medicines in other contexts as always involving ” lying to the patient” is a “loaded language” way of bolstering a case that could often go either way and probably depends very much on the individual case. What doctor has not had some simple soul ask, at the end of a long explanation as to why no treatment is needed: “so you are not going to give me anything for it?”.
But independently of this, the issue with acupuncture is whether it is quite accurate to regard it as placebo. For a long time now I have been referring to it as “mainly placebo”, in deference to the range of likely physiological responses it may provoke or enable.
One could go further than that and ask who makes the decision that something with those potentials is “only” a placebo (other than nybgrus, of course), and on what basis? The evidence allows us to say that it does not posses any unique therapeutic activity, also that it does not work as TCM theorists claim. But it is a complex intervention with some plausible potential above and beyond the expectancy responses that are generally regarded as underlying the placebo “effects” of a sugar pill.
So one answer to your question is to be more discriminatory as to what kinds of treatment we classify as purely placebo. Otherwise we may be trying to practice medicine with one hand tied behind our back.
“Placebo” is just another word for “ignorance.” Prove me wrong.
Even if we were to provide placebos, surely we’re obliged to use the placebos with the least risk of adverse events. That will certainly not be acupuncture. Reiki and other “energy healing” would probably be the safest placebo of all.
So even the “ethical placebo” argument can’t justify acupuncture.
Placebo research means sifting apart the variables in the treatment setting associated with symptom improvement.
So, for example, we can give one group red sugar pills and another group blue sugar pills, to see if pill color has an effect.
Once we know the effect of pill color, pill color is no longer part of what we mean by “placebo.”
So what remains? All the variables with uncertain effect upon symptoms. In other words: ignorance.
Ignorance surrounds us. There’s nothing really wrong with ignorance. But it is a mistake to *value* ignorance. It’s a mistake to build social structure that depend upon ignorance as their raison d’etre. Example: the intelligent design movement.
CAM is analogous to ID in its commercial exploitation of gaps in understanding.
As the person making the positive claim, the onus is on you to name the variable associated with acupuncture that is having an effect beyond patient expectation.
“But it [acupuncture] is a complex intervention with some plausible potential above and beyond the expectancy responses that are generally regarded as underlying the placebo “effects” of a sugar pill.”
Plausible in that it is not doing absolutely nothing? Is acupuncture really much more biologically plausible than sugar pill? Sugar pill increase blood sugar levels and caloric intake shortly after consumption, so it has some biological action, it’s not completely inert either. What is the plausible biologic model for the supposed action of acupuncture? To be biologically plausible, it needs to have a plausible method of action relative to the claimed effect, not just some demonstrable biological effect.
Really, if we can accept that the wise, old, ancient practitioners of acupuncture were wrong about the meridian/chi concept underlying acupuncture, and that they were probably wrong about the requirement of needling, why do many people assume they were right about the effectiveness of acupuncture itself?
Did the ancients use a different way of knowings for determining the various different aspects of the practice of acupuncture? Did they not use rigorous methods to determine these mythical chi points? Presumably those points were the points they determined to be most effective for needling, through trial and error, right? I mean the way they determined chi points couldn’t be more akin to astrology, could it? And if the way they determined those chi points was less scientifically rigorous and more akin to astrology, then surely they used more rigorous, scientific methods to at least determine the effectiveness of acupuncture, right? It wasn’t all just mysticism and mythology, right?
Lactose probably is one of the better placebos because it mimics one of the archetypal placebos, being nursed.
When an infant is nursed, it knows that all is right with the world. It is being held by its mom, and she is calm enough to lactate, meaning that she feels safe. If you are an infant and your mom feels safe and is holding you, you are as safe as you are ever going to be.
Karl: Plausible in that it is not doing absolutely nothing? Is acupuncture really much more biologically plausible than sugar pill?
Of course it is. I have listed several ways in which it may have added effects over a sugar pill. They are all ways that are accepted as having limited medical value in other contexts, yet EBM princesses can still only feel the “P” under all the mattresses.
In any rigorous scientific approach, the toothpick and random puncture point studies don’t actually show that acupuncture-like activities (to be precise) “don’t work better than placebo” (they actually beg the question “what placebo?) — they merely show that certain elements of TCM theory are not necessary to any medical activity that such treatment programs may possess.
(For GlaDOs and nybgrus — the “medical activity” is suggested by patient and practitioner testimonials and thousands of comparative studies that skeptics have often chosen to dismiss as “just showing placebo effects”. We can explain away a lot of those observations as mere illusion, but it is going well beyond the available science to claim that there is nothing else. )
It would not even be surprising if inserting needles into more sensitive parts of the body, such as the tip of the nose, elicited stronger therapeutic responses, and such confounding may explain some of the scatter in the results of RCTs that compare real and sham versions both of which are capable of delifvering much the same generic medical activity.
We have referred to EBM’s overemphasis on RCTs as an evidence base as “methodolatry”.
There may be another version of EBM that is over-preoccupied with mechanisms, having an unreasoning prejudice against methods that don’t have simple, unique physical or chemical effects on human physiology or pathology.
PM, would you consider applying a culture of ammonia oxidizing bacteria? There is some plausibility associated with that as a treatment for non-specific symptoms. I think more plausibility than is associated with acupuncture, and certainly more than is associated with reiki or homeopathy.
AS we have discussed before, Daedalus2, I don’t accept your (very cute) hypothesis as a yet tenable explanation for placebo-related phenomena. The latter clearly require the mediation of higher cerebral centres, and they don’t involve “healing” in the sense in which you seem to want to use the term.
You’ve lost me.
Ok so if acupuncture hurts more it might provoke a greater sense of benefit in the patient, for a period of time.
Seems a reasonable idea. But not an awesome idea. I’d rather spend my research dollars on something else.
PM, I appreciate that you don’t accept my cute NO hypothesis. But I presume you also don’t accept the disproven “hypotheses” of homeopathy, reiki, or acupuncture, yet you seem willing to accept their use because of a placebo effect. A placebo effect you are unwilling to consider might be triggered by my bacteria, either via my cute NO hypothesis or via the physiology of the “real” placebo effect which you do not have an explanation for.
Could you explain why you have a double standard that favors traditional CAM over scientific CAM? (scientific CAM is a term I have just coined to to label science based treatments that have not been shown to be Medicine in clinical trials (Medicine being all treatments that have been shown to work better than placebo, CAM being all treatments that lack clinical trial evidence that they work better than placebo)).
Y’know Freud was all over this placebo thing like white on rice. He played around with hypnosis, suggestion, and ritual. Initially he was keen to find a few tricks to speed up his talking cure. But later he abandoned the effort.
D’you know why? Because he came to understand something he termed the transference.
We cannot see the mind of another. We have to hallucinate it. Mother nature gives us a basic template for agency –friend or foe. Upon that template we add enough details to give us a feeling of what the other person is thinking.
Turns out this placebo thing can work both ways. It can provoke trust and symptom relief in patients. But it can also provoke mistrust, anxiety, abandonment fears, and resistance to getting better.
Freud witnessed many patients initially benefitting from placebo interventions only to become terribly stuck later on. He wasn’t interested in filling his practice with neurotics coming regularly for their mental subluxation adjustments. He wanted people to get better. He wanted patients to become partners with him in the work of understanding their own feelings. That work requires being awake and thinking.
tl;dr: If you make heavy use of the power of suggestion with your patients, you risk fostering a dependency that you cannot cure.
GLaDOS: I lost him a long while ago.
nybgrus, I’d much rather argue with pmoran that the little group of chiropractors that visits this site.
Yes but we have the onus rule for a reason.
Claims can be proven, unproven, or disproven. Sadly the set of unproven claims is infinitely large. But thanks to the labor saving power of the onus rule, we get to lump “unproven” with “disproven,” as “stuff we can ignore for now,” and so carry on with our lives.
GLaDOS, that onus rule is only a heuristic. It is not always lead to correct results.
Ed Whitney:
Yes but patients also expect MDs to base their opinions on sound scientific evidence.
Thinking rationally is difficult. Individuals can’t think clearly over long periods of time apart from an entire culture around them that is vigilant of the rules of organized thought. The introduction of dissonant values and language games into medicine by wealthy and politically motivated theocrats is presently disrupting that culture.
No, acupuncture is merely a symptom of a much larger problem within the house of medicine. Either rules of evidence matter or they do not.
I don’t know what you mean. The onus rule dictates who bears the burden of proof. It doesn’t lead to any “result.”
In the same way as practitioners of acupuncture want very much acupuncture to be proven right, some of you behave like a bunch of denialists, because you want acupuncture to be proven wrong.
An example of that is the link proposed by Peter Deadman was quiet interested, but dismissed because of were it was coming from
If you had taken the time to read the article and argument presented by Marc Bowey, you could have seen that using Ernst as the only valued reference in scientific evaluation of CAM may not be such a good idea.
heretix
“…you could have seen that using Ernst as the only valued reference in scientific evaluation of CAM may not be such a good idea.”
If you think any of the blogers (or any of the serious commenters) on this site are doing anything remotely close to relying on Ernst as a sole source of scientific evaluation of CAM, you cannot have read very many other posts besides this one.
Ernst and everything he has done could be wiped from history by a causality violation, and it wouldn’t significantly change the scientific evaluation of CAM or acupuncture specifically by the bloggers here. Ernst just adds the the plethora of evidence and information we already have.
I read Bovey’s response to Ernst’s article. I was not impressed. Among other things he complains that the method is not defined in the paper – although it is defined in a footnoted reference. He criticizes it for not including an Australian meta-analysis – but meta-analyses are not the same as systematic reviews and do not belong in a review of systematic reviews.
He misattributes an argument in Ernst’s article (any risk is too great) as coming from me.
It struck me as little more than a biased attempt at damage control from a true believer.
I don’t rely solely on Ernst, but I value his conclusions because he has revised them over the years in response to the evidence. He has a strong track record of rigorously following the evidence wherever it leads and being willing to change his mind, rather than consistently trying to support a preconceived conclusion like the acupuncture advocates do.
I was a bit more impressed with Bovey than was Harriet, but not enormously more impressed. I must defend him from one implied criticism, though.
“He criticizes it [Ernst] for not including an Australian meta-analysis – but meta-analyses are not the same as systematic reviews and do not belong in a review of systematic reviews.”
Actually, Ernst did include some meta-analyses (Madsen 2009, Reference #89), but he did not attempt to pool effect sizes himself. All meta-analyses are SRs but not the converse. Bovey correctly identified Hopton as having clearer methods, and I concur. Hopton (way up in this thread) used the same framework for evaluating SRs as Ernst (Oxman and Guyatt 1991), but supplements that reference with the AMSTAR tool, which spells out in more detail what an SR has to do to satisfy the Oxman and Guyatt criteria.
Bovey, though, is focused narrowly on defending acupuncture rather than looking at the broader question of how to create comparison groups for complex interventions, of which acupuncture is only one example. Free access to one article by Paterson and Dieppe at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC558023/?tool=pubmed .
These authors suggest that the randomized placebo controlled trial, invaluable for the testing of the efficacy of new drugs, has been used to evaluate complex non-pharmacological interventions, using physiotherapy, psychotherapy, and acupuncture as illustrative examples. As was suggested above, effects can be divided into “specific” and “nonspecific” effects, which they call “characteristic” and “incidental” effects. They further suggest that elements which are incidental to drug trials may be integral to non-pharmacological trials. These include the process of interaction with the practitioner, the type of talking and listening which is permissible, and the way that a consultation indicates to patients that everything about them is relevant to the diagnosis and treatment plan. Using acupuncture as an example, they assert that needling, talking, listening, and the attention to physical, emotional, and social concerns, are all part of the characteristic factors, rather than incidental.
If Paterson and Dieppe are right, then sham controlled acupuncture trials, which include most of the same complex interpersonal interactions as true acupuncture, are controlling for only one of the characteristic factors of acupuncture. The sham acupuncture session, under this way of looking at it, is like a partial agonist of a drug, producing some or even most of the effects of the test drug. Both compounds are more effective than a placebo pill or no treatment, but they may differ from one another less than both differ from the dummy pill. They could look very much like RCTs with true and sham acupuncture and with a usual care group.
Some comments on this thread have spoken of sham needling as “random”, as if ” needles in random places activate it and that activation occurs in ways that are indistinguishable from the activation by acupuncture.” However, none of the sham studies have done this. The placement of the needles has in each case been systematic in some fashion. “Random” would mean, by definition, that every square centimeter of the body had an equal likelihood of being needled. I have seen no sham controlled trial that has done this.
Much of the debate over the “implausibility” of acupuncture has arisen from its classically described model of the human body, with “meridians” and “qi” flowing along them. If, however, these were surrogates for factors that have physiological existence (connective tissue planes, cutaneous nociceptors, and other sensory receptors), then the plausibility of acupuncture changes.
There is good evidence that yoga is effective for many kinds of back pain and other musculoskeletal conditions. There is no comparable scientific evidence for chakras and prana, at least not as described in classical yoga. However, the way that yoga strives to align and integrate these entities could have biomechanical effects of which chakra and prana are surrogate descriptors. If yoga is to chakra what acupuncture is to meridian, there may be two active interventions whose classical explanatory principles are pre-scientific, but partly corresponding to normal physiologic variables. Quaint and obsolete language to describe them both should not directly affect our estimates of biological plausibility.
There is one pervasive kind of bias that seems to permeate almost all acupuncture RCTs and warrants mention here. Randomization, while protecting threats to internal validity by preventing biased comparisons of treatment groups, leaves itself open to other kinds of bias. Notably, participants in acupuncture trials are generally interested in acupuncture in some way, are often recruited through public announcements, and are enticed to participate in the RCT by assurances that they may cross over to the acupuncture treatment at the end of the trial if they are assigned to one of the control groups. They are, in short, more likely to be placebo responders, at least to sham acupuncture, and to expect good results from it. An RCT composed entirely of subscribers to the Skeptical Inquirer would have an unbiased comparison of treatment groups, but the effects of sham acupuncture would be much less, and much closer to usual care. Then leave it to Cochrane to combine them all in a meta-analysis, calculate that the heterogeneity was high, and choose a random effects model to estimate the overall effect, giving its readers one more damn mess to try to make sense of.
@ Ed Whitney,
“they assert that needling, talking, listening, and the attention to physical, emotional, and social concerns, are all part of the characteristic factors” only emphasizes that it is these factors that induce a placebo response that are characteristic of acupuncture.
The plausibility of how specific effects might be caused is not even an issue while the existence of specific physiologic effects has not been convincingly demonstrated.
The hypothesis that acupuncture is a complex suite of placebo and other nonspecific effects fits all the published data.
“If you think any of the blogers (or any of the serious commenters) on this site are doing anything remotely close to relying on Ernst as a sole source of scientific evaluation of CAM, you cannot have read very many other posts besides this one. ”
Although Ernst is not the only reference, he is the one that is the most frequently referred to.
It seems to me that the skeptics are out to “prove acupuncture wrong” rather than give it a chance and a proper evaluation;they dismiss it as a placebo rather than considering the psychosomatic impact of such a treatment; it is also because of this psychosomatic impact that it is difficult to evaluate this kind of treatment with RCTs set up in the same way as they would be for testing drugs.
The skeptic movement seems to be a predominantly anglosaxon thing with no real equivalent in non-english speaking part of the world; thus there is a tendancy to let research done in the english language predominate and ignore good research from other non-english speaking countries.
heretix, what experiment would you propose to distinguish between a psychosomatic effect of a treatment like acupuncture and the psychosomatic effect of a placebo?
If there is no experiment that can distinguish them, how is it useful to think about them as different things?
More that it has already been given a chance and a proper evaluation, proven wrong, and we don’t close our eyes to the evidence.
That IS what makes it a placebo.
False. The exact form of the control is different (sham acupuncture instead of a sugar pill) but RCTs are absolutely applicable.
“It seems to me that the skeptics are out to “prove acupuncture wrong” rather than give it a chance and a proper evaluation”
Actually, if you’d been reading this blog for a while, you’d know that acupuncture was one form of CAM the Dr. Gorski once considered plausible, not because he believed in the chi model, but because it was an active, invasive biological intervention. Only after rigorous research and review of scientific studies did he come to the conclusion he now holds.
Acupuncture has been given a fair chance already and has had numerous proper evaluations. The strongest positive spin that anyone can possibly reasonably make at this time is that the evidence is, at best, still inconclusive and contradictory, which at this point in the game means it’s really time to move on.
In spite of that, it is my understanding that the bloggers here are more than willing to give any new evidence objective consideration. Nothing so far has withstood the crucible of scientific scrutiny.
heretix and Ed Whitney seem to think that acupuncture is the green jelly bean.
Karl: “In spite of that, it is my understanding that the bloggers here are more than willing to give any new evidence objective consideration. Nothing so far has withstood the crucible of scientific scrutiny.”
I also cannot see obsolete TCM hypotheses ever being seriously resuscitated.
Nevertheless, a fresh problem looms. I predict that we will now see a lot of “pragmatic”, real-world studies comparing acupuncture and other CAM methods with drugs for various common conditions wherever the drug treatment is of “proven”, but limited benefit.
I further predict that this will be a close-run thing. There are already a smattering of such studies about, most notably studies showing a similar prophylactic effect for acupuncture as drugs on migraine*, using the fairly objective measure of a migraine diary, with considerably more side effects and drop-outs in the subjects allotted to the drugs.
Such studies have the potential to seriously challenge the mainstream’s scientific and ethical positions on treatments having mainly placebo/non-specific ‘effects’ and I don’t know why we are not seeing any discussion of them.
If the results proved consistent in quality studies, we might have to either admit that both methods are actually placebos, or allow them similar status within practical medicine, deciding between them on cost/risk/benefit grounds.
* http://www.ncbi.nlm.nih.gov/pubmed/19160193
“The hypothesis that acupuncture is a complex suite of placebo and other nonspecific effects fits all the published data”
If this means that all nonspecific/incidental effects of acupuncture (interaction with a practitioner who takes the time to listen to the patient, who touches the patient in a therapeutic manner prior to placing the first needle, who pays attention to how the patient is responding during the session and adjusts the treatment approach accordingly, and creates an atmosphere of participation in which the burden of illness can be shared and partly relieved) are counted as “placebo,” that is consistent with the data. Paterson and Dieppe suggest that they be counted as an inextricable part of the entire encounter with the acupuncturist, and thereby count as specific/characteristic effects, but their suggestion need not be taken. It is just one way of looking at how to analyze a complex intervention.
The key word is “complex.” We appear to be in agreement that it is not likely that you can get the same outcome by giving the patient a sugar pill and saying, “Here, swallow this.” And the methods for investigating the effectiveness of complex interventions ought to be of interest to us all. Physiotherapy, psychotherapy, cognitive-behavioral therapy (CBT), multidisciplinary rehabilitation, all are complex and lend themselves poorly to the kind of placebo control which works so well for new drugs. If CBT is as effective as instrumented fusion for patients with chronic disc degeneration, that is a big deal, but the simple inert placebo control cannot be used in its evaluation.
I realize that for the most part, this web site is not mainly about general problems in research methodology. It is focused on one particular kind of problem facing the science-based medical community, to wit, “woo.”
Trouble is figuring out exactly what criteria an intervention has to meet before qualifying for classification as “woo.” I am working on that. It seems that the intervention has to purport to be a form of medical practice, no? Yoga does not count, in spite of the chakras, tai chi does not count, and chai tea does not count. Fair enough. Astrology for deciding when to propose to your beloved, no; astrology to guide decisions about surgery for a herniated disc, yes. Foot reflexology to delay the onset of labor so that the newborn child will be a Sagittarius instead of a Scorpio, most probably.
A biologically ludicrous proposed mechanism of action? Sufficient but not necessary—homeopathy qualifies on this criterion alone, but not acupuncture. Potential adverse effects—helpful, but neither necessary nor sufficient. Unanimous, categorical rejection by all responsible medical professional organizations—works for homeopathy but not for acupuncture.
Will keep working on it. Honest.
“heretix and Ed Whitney seem to think that acupuncture is the green jelly bean.”
If I were to come to the defense of the green jelly bean or any other kind of pseudoscience, I would follow the postmodern pseudoscience advocacy guidelines of the American College of Charlatans and Snake Oil Salesmen (ACCSOS). This would take the form of a series of posts in roughly this form and sequence:
1. Rambling diatribe about paradigm shifts, Cartesian dualism, the holographic brain, quarks and/or whale songs, supported by mangled, out-of-context quotes from Thomas Kuhn, Paul Feyerabend, and Michel Foucault, plus reminders that Isaac Newton dabbled extensively in alchemy, followed by links to websites dedicated to the postmodern philosophy of Deleuze and Guattari
2. Something about the Heisenberg Uncertainty Principle, chaos theory, the butterfly effect, Bell’s inequality, nonlocality, and Schroedinger’s cat, accompanied by quotes from Rupert Sheldrake, David Bohm, Ilya Prigogine, and the Flower Garland Sutra, with self-congratulatory comment about refusing to be grounded or compromised by referentiality, plus additional diatribe against Western Science since the time of Francis Bacon for inciting man’s violent gang rape of nature
3. Non-sequitur fart joke in response to criticism of post #2
4. Follow-up post four days after #1, to the effect that since no one responded to the cryptic but devastating quote from Foucault, no one was capable of responding, and therefore the contributors to this thread must be mentally myopic, morally dishonest, or both. Then make a petulant resolution to start a new blog to promote own point of view; only to die while jaywalking across Sepulveda Boulevard in order to purchase homeopathic Cadmium iodatum in lieu of potassium iodide to ward off effects of radiation coming from Japanese earthquake.
I fail to see how anyone can acknowledge this, yet not then conclude that the acupuncture part of acupuncture (you know, the needles bit) provides no demonstrated benefit, carries risks, and therefore cannot be justifiably practiced.
Nobody disputes that a soothing interaction with a helpful person in a relaxing environment can help people feel better. But it’s a gargantuan leap from there to the claim that sticking needles into people, and charging them money to do so, is a good idea.
Take out the acupuncture bit, substitute efficacious medicine, and you have what I would call a good patient-physician encounter. The fact that the current levels of physician engagement with their patient in this regard are sub-par at best does not advocate the substitution of acupuncture for efficacious medicine in order to capture and deliver those missing aspects the patient-physician encounter. No, I see it as a call to improve ourselves as physicians. When you have a scenario in which all of those aspects are present in both an acupuncture setting and an SBM setting then it is clear to me who will win out. Trying to divine some sort of place for acupuncture and rationale for its usage in light of the data and this knowledge is akin to saying that because a particular town has a lot of bad interventional cardiologists we should do away with PTCA and stenting for MI and have someone take the time to listen carefully to the person having a heart attack, touch them in a therapeutic manner, pay attention to how much their chest hurts and adjust the treatment approach, and create an atmosphere of sharing the pain of the MI with the patient, all whilst waving crystals or giving extremely dilute and succussed epinephrine.
No, I think I would simply rather train those cardiologists to do their job better and not give in to the woo in the meantime.
Not so difficult really. You are missing a key point here – the distinction is one made between individual free choice and something being a medically accepted fact.
Because yoga is great exercise and people can exercise their chakras whenever and howver they like on their own time. If you were to start claiming that yoga cured cancer, then we would have woo. Ergo, when you start claiming that acupuncture can cure this, alleviate that, etc… you have woo. If all you claimed was “Sticking needles in you is a fun way to relax and make you feel better” then I hardly think anyone here would care to say much about it. Massage is relaxing (and I am way overdo for one) but my girlfriend can’t stand them. It is literally torture for her. Some may find the arbitrary sticking of needles in them quite relaxing and enjoyable, others not so much. Inasmuch as it stays in that realm, there is no problem. But if massage suddenly is claimed to cure disease, or if needles in your skin is claimed to reduce pain and increase function then it becomes a matter of scrutiny. When the claims do not match the results, and the claim continues to persist, then we have woo. Simple enough.
“…yoga is great exercise and people can exercise their chakras whenever and however they like on their own time. If you were to start claiming that yoga cured cancer, then we would have woo. Ergo, when you start claiming that acupuncture can cure this, alleviate that, etc… you have woo. “
So here is the thing: Yoga has been claimed to alleviate chronic low back pain.
See Williams K, Abildso C, et al. Evaluation of the Effectiveness of Iyengar Yoga Therapy on Chronic Low Back Pain. Spine 2009;34(19):2066-2-76.
90 patients (69 women, 21 men, mean age 48) treated for chronic nonspecific low back pain in the orthopedics department of the University of West Virginia. Eligibility criteria included age 18-70, BMI<37, low back pain at least 3 months, Oswestry Disability Index (ODI) score between 10 and 60, pain VAS between 3 and 8 cm (30 to 80 mm), and ability to get up and down from the floor without assistance. There were lots of exclusion criteria: spinal stenosis, tumor, infection, osteoporotic fracture, ankylosing spondylitis, spondylolisthesis with radiculopathy, kyphosis, scoliosis, radicular pain with weakness, failed back syndrome, pregnancy, fibromyalgia, major depression, substance abuse, current litigation, having practiced yoga in the previous year, or an open workers compensation for low back pain.
They were randomized to yoga (n=43) or control (n=47). The yoga group got 24 weeks of twice-weekly classes of 90 minutes duration, a yoga DVD, and was asked to practice yoga 30 minutes per day. The control group continued with self-directed standard medical care. The yoga group did poses like Supta urdhva hastasana and Parsva pavanamuktasana. There were about 30 such postures listed in the article. (Aside: what do you get when you cross the Godfather with a yoga instructor? Someone who makes you an offer you can’t pronounce).
ITT analysis for ODI showed a mean decrease of 7.3 points (29%) in the yoga group at 24 weeks, with a decrease of only 2.3 points (10%) in the control group; for pain VAS, the corresponding decreases were 17.6 mm (42%) and 4.4 mm (10.7%). At 6 months the yoga group continued to have lower ODI and VAS scores than the usual care group.
About 63% of the participants had previous experience with “CAM” but the nature of the woo was not further described. The research assistant who collected the outcome data was blinded to group assignment. The method of randomization was not as clear as you would like, since the participants were given envelopes with group assignment but allocation concealment was not clearly described.
Overall, the article was adequate for evidence that yoga can improve nonspecific chronic low back pain. And there you have it: a study claiming that yoga can treat a common back disorder.
I asked specifically if yoga was “woo” precisely because the claim made in this study is no different from that made in most of the acupuncture literature I read. Nonspecific neck pain, nonspecific back pain, and peripheral joint osteoarthritis are alleviated by both true and sham acupuncture. I have seen no claims that acupuncture cures cancer; if I did, I would approach them with the assumption that the studies were not credible, and that the methods were unsound. Most acupuncture studies exclude specific pathologies from inclusion; far from claiming to cure cancer, they exclude tumors and the other diseases that were excluded from the yoga study.
Acupuncture makes the same low-key claims that yoga makes. Acupuncture probably works by stimulating nociceptors in a particular pattern (hence toothpicks can reproduce most of the effects of needles). This is a perfectly reasonable mechanism of action, and piercing the skin may not be required. After all, no one chastises medical students on their neurological examinations if they failed to puncture the skin during the sensory exam.
No chakras, no prana, no meridians, no qi, just modest claims to alleviate symptoms of some common musculoskeletal conditions. No grandiose claims to rid the world of all known diseases, just moderately effective adjuncts to reasonable treatment plans for some forms of chronic pain.
So I am still stuck: acupuncture is “woo” but yoga is not. I do not get it.
Ed Whitney,
Discussion of tai chi and fibromyalgia here:
http://www.sciencebasedmedicine.org/?p=6617
My take on the acupuncture ruckus is:
1) Actually, it’s not clearly that good for anything in particular. It might help some people but it doesn’t seem to help much; otherwise, clinical results would be less equivocal.
2) I thought the nocioception hypothesis hadn’t been sustained, perhaps I’m wrong. If it hasn’t been sustained then it really just is the attention and deliberate relaxation and the poking itself is not part of any beneficial effect. If poking is not required — even with toothpicks — then it’s unethical to pretend that it is and any risks associated with poking are unjustifiable.
3) Even if acupuncture does offer significant pain relief, does that make it medicine? Massages and sex can often provide pain relief as well but are not considered medical procedures or therapies and are not routinely offered in medical clinics. People can obtain these things outside the clinic in a non-medical context and get at least as much benefit from them there.
But I think it’s mostly 1). IF it doesn’t actually do much, THEN there’s no point in discussing how it works.
*** *** ***
I do find the area of nonspecific effects troublesome. In ngybrus’ ideal of medical practice, nonspecific effects are leveraged by practitioners of scientific medicine. One of the major contributors to nonspecific benefits is human physical touch, such as happens during a routine physical exam or when blood pressure is measured.
From the comments in this post on the usefulness of the routine pelvic exam:
http://www.sciencebasedmedicine.org/?p=11218
OP:
“Hands-on interactions and the perception of “doing something” can be reassuring and can enhance the doctor/patient relationship. But can’t a caring clinician attain those same benefits within the realm of science-based medicine?”
Alison Cummins:
“What hands-on interactions do you recommend for routine visits? Blood pressure?”
Harriet Hall:
“It is more cost-effective for BP to be taken by others with lower salaries. The MD’s time is better used in activities requiring more education. It would be presumptuous of me to recommend any hands-on interactions that have not been studied and are not evidence-based; but things like shaking hands and touching/examining areas of concern to the patient would seem to be appropriate. Human interactions will vary with culture and custom and with individual styles.”
If blood pressure is taken by a nurse who uses a special chair and doesn’t need to actually touch the patient, then physical contact between doctor and patient may be limited to a handshake.
Most doctors do not offer a handshake.
“I realize that for the most part, this web site is not mainly about general problems in research methodology. It is focused on one particular kind of problem facing the science-based medical community, to wit, “woo.”"
This site deals with science based medicine and non-science based medicine. Although the major focus is what is often called woo, it also deals with more mainstream, conventional medicine, like the recommended schedule for mammography, etc.
“So here is the thing: Yoga has been claimed to alleviate chronic low back pain.”
It’s a form of light exercise, so that’s not implausible. It’s reasonable to study that. The practice of yoga is not woo; it’s light exercise. However the traditional underlying philosophy and understanding of yoga is indeed woo.
Likewise the traditional underlying philosophy and understanding of acupuncture is clearly also woo. The actual practice of acupuncture has been studied and been found to be no more effective than the controls of needling in “wrong” locations and no needling at all. It has so far failed to be scientifically validated.
It is another question entirely if any certain aspects of, or similar to, the practice of acupuncture may have some scientific validity, such as needling. The studies show that poking works just as well as needling, so we don’t even have an aspect of acupuncture that may be valid. At best, what we have is some possible practice derived from/inspired by acupuncture, such as poking in random locations, that could potentially be valid, but the plausibility doesn’t seem to great for that either.
“Acupuncture probably works by stimulating nociceptors in a particular pattern (hence toothpicks can reproduce most of the effects of needles).”
That’s an interesting speculation, but I wouldn’t even call it a hypothesis at this point. For one thing, you’re working backwards from the assumption that acupuncture works and speculating on remotely biologically plausible mechanisms whereby it could act. I can try to determine where the Loch Ness Monster and Bigfoot came from, what kind of animals they are, and how they live, but I really need to establish that they exist first.
If you accept that needling is not required, I don’t see why you’re even discussing acupuncture anymore rather than advocating randopoking. Needling is unnecessarily invasive compared to poking. Even if you think that poking only reproduces MOST of the effects of needling, I would think that we need to determine when poking is sufficient and when the added risk of needling is warranted. Such needling, if scientifically valid, would be medical needling, not acupuncture as key aspects of the practice of acupuncture, such as chi points, would be absent.
Alison Cummins – “Massages and sex can often provide pain relief as well but are not considered medical procedures or therapies and are not routinely offered in medical clinics. ”
As part of my physical therapy for two conditions (Plantar fasciitis and hip/leg pain related to SI joint issues) was deep tissue massage. Both of the therapists also showed me how to do some form of at home massage (a tennis ball is your friend), for when I noticed the condition flaring up. Paid for by insurance, so I will assume that some targeted massage for tendon inflammation can be considered a medical therapy, at least here in the states.
Yes, and presumably that was a specific effect of that form of touch on that part of the body for that problem.
I’m thinking more of the kind of massage that you get at a spa which feels really good and has you walking on air afterwards. That kind of massage can help all kinds of aches and pains (though not for long).
Regarding handshakes. Recently (say in the last five or eight years) I’ve noticed that doctors are offering handshakes more. Particularly upon first meeting, but often at the end of an appointment too.
I’m not a big handshaker myself. But in this case, I do think it’s a nice touch. It sort of puts the appointment in a more respectful, professional light.
Actually, massage is a topic I would love to see covered here, from a science based medicine perspective, of course. It’s a topic in need of some separating of wheat from chaff. I think it is a practice (or numerous practices) that stretches from reasonably scientifically base (or at least not un/pseudo-scientific base) all the way to uber-woo including Reiki and Craniosacral therapy.
Any skeptical/ Science Based massage experts out there? (Do I hear crickets chirping, or is that the minor narrow frequency hearing loss ringing in my right ear?)
On massage, specific vs general. Agreed.
Also, I would say that a 45 minute spa type deep tissue massage offers some relief and increased mobility for back muscle spasms. It was often enough take the hard edge of the worst first day of a three or four day muscle spasm cycle. For me, the relief offered was about the same as that of gentle exercise that raises the heart rate (say walking on a treadmill). Although the massage is more pleasant, the walking is cheaper and possibly more accessible (driving with muscle spasms is not fun).
Never tried acupuncture, though. Can’t compare the non-specific results.
Karl:”If you accept that needling is not required, I don’t see why you’re even discussing acupuncture anymore rather than advocating randopoking.
It’s tricky. Remember poking with toothpicks (or retracting needles) only works as well as “real” acupuncture if the patients think they are being needled.
I agree the word acupuncture has lost any clear meaning once we divorce it from TCM.
@karl: exactly what I was to say. Yoga is not woo for back pain because yoga actually is exercise and actually specifically stretches the back, ergo very plausible and likely mechanism. If you claimed that it was the chakras re-aligning that fixed your back pain, now we are back to woo. Also I would like to add that yoga can be a very intense workout – I do the P90X version and it kicks my butt!
pmoran: that is exactly the point, isn’t it? When you divorce acupuncture from its TCM roots it suddenly just becomes a complex human interaction that has non-specific effects. Arguing for the efficacy and utility of such a method is akin to saying we should have shamanistic voodoo dances in our waiting rooms since those will have non-specific effects as well.
Intriguingly enough, Nybgrus, when challenged, nearly everyone agrees that it would be OK to recruit the services of the local shaman should a member of an isolated tribe want him to participate in their hospital treatment.
We merely deny a similar courtesy to our own culture, even when there are clearer placebo/non-specific/??specific benefits to be had from treatment programs like acupuncture in some settings (see Ed’s effect-sizes of 0.45).
I also, by training and bent, approve of medicine trying to confine itself to EBM -endorsed methods. But let’s be absolutely clear as to why and how we draw certain lines, and consider how we should deal with certain inevitable consequences (which happen to include wider use of CAM).
I respectfully disagree. I think I know what you are referring to, but I also think that was more of a tolerance for practices deeply imbued in a person under your care, not an endorsement of their use. The idea being that denying what would otherwise be a non-harmful (but non-beneficial) “treatment” from the family would cause undue mental and emotional stress. That is a tolerance on a individual basis and things like that are indeed denied to the patient if there is evidence of likely harm or other problems (like burning sage in an ICU room or something). The lines you speak of seem reasonably clear to me, but muddied to you for some reason.
If a patient came into my care who was in hospital and adamantly demanded acupuncture, I would state that to my knowledge I think it would do nothing, that the patient should procure and pay for the acupuncturist, and that I would examine the patient afterward to ensure no deleterious effects, but I would not deny the acupuncture. That is wholesale different from actually endorsing acupuncture as a viable treatment modality and offering it to the patient.
nybgrus, it turns out that the external skin has the highest nitrite level of any tissue compartment. At least in rats. I think that is true for all mammals and relates to reduction of nitrite to NO in the region of skin that is most hypoxic.
The skin is the tissue compartment that is among the most hypoxic in the body. The external skin gets O2 from the external air, and there is a minimum O2 concentration where O2 diffusion from the internal blood supplies the O2.
I think massage is a way to increase lymph flow and increase the transport of nitrite in the skin into the blood stream. I think that cupping is a way to put that lymph and blood containing nitrite in an even more hypoxic state so that the nitrite is reduced to NO. Reduction of nitrite to NO is inhibited by O2. That is probably the source of hyperemia when skin is pressed on and blood flow is prevented. Blocking blood flow causes local hypoxia, nitrite is reduced to NO, the NO causes vasodilatation until the hyperemia restores the local O2 level.
“you’re working backwards from the assumption that acupuncture works and speculating on remotely biologically plausible mechanisms whereby it could act.”
Well, I said “nociceptors” when I should say “cutaneous afferents.”
I was basing my assumption that it worked based upon a number of studies, the best of which in my opinion was Haake (Arch Intern Med. 2007;167(17):1892-1898). This was one of the largest studies, and gives the best support for the equal effectiveness of true and sham acupuncture. They randomized 1162 patients to true acupuncture (n=387), sham acupuncture (n=387) or standard therapy (n=388). Patients in the conventional therapy group “received a multimodal treatment program according to German guidelines.” The guidelines provide “the treating physician with recommendations about the treatment algorithm and assess the various therapy forms according to the degree of evidence based on a literature search and recommendations of the specialist associations.” The no acupuncture group is getting good evidence-based treatment in this study, and is therefore a suitable comparison group for the two acupuncture groups.
The acupuncture was done by physicians of various specializations who had at least 140 hours of acupuncture training: 55% had undergone basic training (mean, 213 hours) and 45% had advanced training (mean, 376 hours). The study physicians had practiced acupuncture for 2 to 36 years (median, 8.0 years). “Sham acupuncture on either side of the lateral part of the back and on the lower limbs was also standardized, avoiding all known verum points or meridians. As with verum acupuncture, 14 to 20 needles were inserted, but superficially (1-3 mm) and without stimulation.”
The primary outcome was treatment response 6 months after randomization, defined as 33% improvement or better on 3 pain-related items on the Von Korff Chronic Pain Grade Scale or12% improvement or better on back-specific functional status measured by the Hanover Functional Ability Questionnaire. Patients who had recourse to additional treatments other than rescue medication were classified as nonresponders, as were unblinded patients. The interviewers were blinded to the treatment group of the study participants.
The primary outcome was attained by 47.6% of the true acupuncture group, by 44.2% of the sham acupuncture group, and by 27.4% of the conventional therapy group at 6 months after randomization. This analysis included all patient on an intention-to-treat basis. The differences between true and sham acupuncture were not significant; the differences between both acupuncture and conventional treatment were. In terms of the success rate, sham acupuncture was 1.6 times as successful as the evidence-based guideline multimodal care. True or sham, that damn acupuncture is doing something that can be measured by a blinded observer using a conservative method of analysis.
“I don’t see why you’re even discussing acupuncture anymore rather than advocating randopoking. Needling is unnecessarily invasive compared to poking. Even if you think that poking only reproduces MOST of the effects of needling, I would think that we need to determine when poking is sufficient and when the added risk of needling is warranted.”
Assuming that randopoking means “random poking,” we have no data about it and cannot speculate on it. This assertion of random needle placement has been made before. Random has a formal definition; it means that every element in a population is equally likely to be selected in a sample. The eyeball is just as likely to be selected for needling as the T11-T12 intercostal space on the right or the tip of the index finger.
The “sham” acupuncture in the Haake study was done by the same trained physician acupuncturists that did the true acupuncture. They may have chosen points that are not identified as acupuncture points, but the points were chosen deliberately, and they did not operate haphazardly or randomly. For one thing, they are physicians, bound by the Hippocratic Oath, sworn to do no harm and to seek the well being of all who come to them. Even when conducting a study, they are not just poking around. In order for the RCT to be valid, this has to be the case. If you had the control group get poked by Larry the cable guy when he came to install the cable TV setup, and gave him 14 needles and told him to go poke that lady over there, you would have an invalid study; there would be two variables to account for instead of one: the provider and the technique.
If this can be done non-invasively, and produce a 60% increase in the success rate of treatment of chronic low back pain, that would be a very good thing. Hell, I might even sign up to have it done when my back bothers me.
A while back Harriet quoted Carl Sagan’s great skeptic’s motto: “Extraordinary claims require extraordinary evidence.” True enough; it is also true that mundane claims require mundane evidence. We are not talking about cancer or HIV or Alzheimer’s disease here: these studies are done on low back pain after serious pathology has been ruled out. A non-specific modality alleviates symptoms of a non-disease more effectively than conventional standard care. Not by a miraculous amount, either, but by enough to matter. A mundane claim has been made, and mundane evidence has been supplied in support of it. No wu, no mu (well, maybe endogenous ligands of mu receptors are involved), no mysticism. A pretty boring condition responds to a pretty boring treatment. The treatment is not represented as good for spinal cord compression from multiple myeloma, but for garden variety mechanical back pain. You can use yoga or you can use acupuncture; if you choose the latter, you can choose true or sham acupuncture, whichever is cheaper. Just make sure you are treated by a trained acupuncturist!
How about yourself? Do you see an acupuncturist for pain?
nybgrus The lines you speak of seem reasonably clear to me, but muddied to you for some reason.
They once seemed quite clear to me, too. Yet if complex interventions are proving to have similar or better outcomes to normal EBM-endorsed medical care, possibly even with fewer ill effects than commonly employed drugs, I am required by the same standards of evidence to look at the “lines” more closely.
Although, I wish that in studies like the one Ed has just described they would place more emphasis on more objective outcomes, such as analgesic consumption and how long off work? Those results will be less contaminated by reporting biases– patients supplying what they think are the expected answers.
GLados: How about yourself? Do you see an acupuncturist for pain?</i?
No. There may first need to be a special series of studies using hardened skeptics as subjects.
I have allowed that acupuncture is "mainly placebo".
No, it does not. Random means that there is a distribution. It most certainly does NOT mean that the distribution is uniform. The sum of two dice rolls is random, but non-uniform. A Gaussian is random but non-uniform. And so on. Not a key part of your point, but sufficiently wrong that it needs correcting.
Man, you just can’t get away with carelessness on this blog! I should have specified that I meant a simple random sample of the body surface. If the average 70 kg man has a body surface area of 2800 square inches, then each square inch in a simple random sample is equally likely to be selected for a random needling. Scott is correct to point out that the essential thing about a random variable is that it is described by a probability distribution.
Main point is that the so-called sham acupunctures in RCTs are done in a way that makes them comparable to drugs with similar mechanisms of action. A carefully planned sham acupuncture done by a trained practitioner is like a partial agonist of a drug receptor. That is what makes the selection of suitable controls for acupuncture so controversial.
@Ed Whitney
“Random has a formal definition; it means that every element in a population is equally likely to be selected in a sample. The eyeball is just as likely to be selected for needling as the T11-T12 intercostal space on the right or the tip of the index finger. ”
Did you really think I meant truly random poking at any location on the body, including the eyeballs and not understand that I used the term randopoking simply to differentiate from the chi point based needling of acupuncture?
If so, let me just clarify for you that the term “rando” was indeed derived from the word random, but was not intended to imply chaotic randomness in point of application, but was only meant to imply the use of non traditional application points (in other words, the non-use of traditional points of acupuncture).
Perhaps it was my fault that I left some straw, clothing and matches laying around. /endsnark
“They may have chosen points that are not identified as acupuncture points, but the points were chosen deliberately, and they did not operate haphazardly or randomly. ”
I’m curious then, what was the basis for the selection of the points?
I would think that for the control, you would define acceptable application regions, but that the specific points used within those regions should indeed be assigned randomly, possibly even different for each test subject to minimize the possibility that you might discover a new, unknown chi point.
@ Karl
If you want me to go on arguing, you have to pay for another five minutes.
Seriously, folks, the sham acupuncture is being done systematically by trained practitioners in a systematic, planned fashion, and that makes the true/sham comparison more subtle than is the case in placebo-controlled drug trials. It is a bit like telling a trained gymnast to go out on the mat and do the routines in a clumsy manner; they will depart from their usual routine, but do it gracefully. An attentive observer can immediately tell the difference between a trained gymnast acting like a klutz and an actual klutz.
Main soapbox is that the interpretation of studies is not standardized or completely straightforward. Imagine the horror you would feel if a nominee for Chief Justice of the United States sat in front of the Senate Judiciary Committee and said that judging Constitutional law was just like an umpire calling balls and strikes. What a nightmare that would be! Just hypothetical, of course.
Oh, wait…
Ed Whitney,
“If you want me to go on arguing, you have to pay for another five minutes.”
Awesome response!
Thanks to this web site there’s a chance anyone Googling “acupuncture” will learn that the procedure works about as well as a kindly doctor offering reassurance, the tincture of time, chicken soup, back rubs, etc. –i.e., stuff a lot cheaper than a $1000 series of acupuncture visits.
So when you say “skeptic,” do you mean someone who has been exposed to this kind of information?
“Although, I wish that in studies like the one Ed has just described they would place more emphasis on more objective outcomes, such as analgesic consumption and how long off work? Those results will be less contaminated by reporting biases– patients supplying what they think are the expected answers”
That is true of a lot of studies out there; something like analgesic refills and return to work are reported less often than the self-rated stuff. The Hanover Functional Ability Questionnaire was used as a functional outcome, and looks similar to Oswestry and Roland-Morris scales that turn up a lot in the US literature—can you reach up and get a book off a high shelf, pick up a light object from the floor, wash your hair over a washbasin, etc. I can only find it in German on the internet. In Googling it I ran across something just made for the SBM website:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3004618/pdf/jpr-2-075.pdf is free access. Anthroposophic therapy for low back pain, from the Institute for Applied Epistemology and Medical Methodology in Freiburg. Gotta love those Germans—an Institute for Applied Epistemology!
This is honestly very interesting, intelligent, and detailed analysis about these sorts of things. Ed is very much applying thoughtful detailed analysis at a level much higher than your average Bear. (And no, that is not at all a snark).
However, the main problem here is that it is tooth fairy science. Missing the forest for the trees. Assuming something and working very intelligent discourse around it.
And pmoran:
Indeed. But medical ethics keep those lines quite nicely in place.
GlaDos:Thanks to this web site there’s a chance anyone Googling “acupuncture” will learn that the procedure works about as well as a kindly doctor offering reassurance, the tincture of time, chicken soup, back rubs, etc. –i.e., stuff a lot cheaper than a $1000 series of acupuncture visits.
That may be said, but is it an entirely accurate reflection of the available evidence? What do relevant studies actually show?
So when you say “skeptic,” do you mean someone who has been exposed to this kind of information?
Not really.
Harriet Hall summarized the acupuncture studies like this:
You said acupuncture probably won’t work on skeptics. So to clarify I asked if by “skeptic” you mean people who accept Dr. Hall as a qualified expert and who happen to read her article above, for example.
So help me out here. What do you mean by skeptic in this case, with respect to acupuncture?
that is exactly my point glados… considering that I view the evidence as not supporting the inherent efficacy of acupuncture, that I think the mysticism surrounding it is BS, and that all the is there is the ritual… would the ritual stil have an effect on me? Do you think that if I, personally, went to an acupuncturist I would experience any sort of positive effect from it? If not, then how can you be arguing for its use on patients? When the entire “effect” of a treatment evaporates the moment you no longer believe in the ritual what do you have left?
Nothing. And nothing you can recommend to patients. And nothing you can feel good knowing people are paying for. That is the definition of placebo I like. And “works.” If something stops “working” because you have become educated as to exactly how it works, then it never “worked” in the first place. And I would be willing to bet that it wouldn’t “work” on me if I were to get it tomorrow.
Of course, you can then claim that I don’t have the proper condition, or mindset, or whatever you can come up with and thus it would work in a certain subset of the population, of which I am not a member. And thus you could liken it to, say an antidepressant, of which any given agent only “works” in about 60% of the population. But that would be disingenious – the key difference here is that once the antidepressant does work you cannot suddenly make the effect go away by teaching the patient about the mecahnism of SSRI and norepi in the brain. But if you do have a person on whom acupuncture “works” and you educate them to it, I am willing to bet it will suddenly stop working. That is the key difference.
I may honestly go out and get some acupuncture soon just to try and prove the point.
“You said acupuncture probably won’t work on skeptics. So to clarify I asked if by ‘skeptic’ you mean people who accept Dr. Hall as a qualified expert and who happen to read her article above, for example.
So help me out here. What do you mean by skeptic in this case, with respect to acupuncture?”
Well, I was going on the basis of the fact that expectations of benefit are an important component of the placebo response, and that the placebo response would be less in skeptics insofar as that component was less. Because a high placebo response is one factor that can obscure a true treatment effect, a specific effect of acupuncture, if there is one at all, would be easier to detect in a population with a low placebo response rate. This would be a good study to do. You could not use the usual incentives for enrollment, that whatever group assignment the participant receives, he or she will be eligible to receive real acupuncture at the end of the trial. Maybe offer tickets to Penn and Teller; I don’t know.
Ralph Horwitz at Yale once estimated that one third to one half of RCTs in medicine are negative, even when the intervention is known to be clinically effective. I don’t know how he came up with that figure (it does not seem to apply to published RCTs), but sometimes it certainly does happen. R. Buynak et al (Expert Opin Pharmacother 2010;11(11)1787-1804) randomized 965 patients with chronic low back pain to tapentadol, oxycodone controlled release, or placebo. The attrition rate was high, but tapentadol and oxycodone were both superior to placebo in the average reduction on the Numerical Rating Scale from baseline to 12 weeks (2.9 for tapentadol, 2.9 for oxycodone, and 2.1 for placebo). When responder rates were compared (defined as either 30% or 50% reduction in pain), the rates were higher in tapentadol than in placebo groups, but oxycodone did not differ significantly from placebo: for 30% response the rates were: placebo, 27.1%, tapentadol, 39.7%, oxycodone, 30.4%; for 50% response, the rates were: placebo, 18.9%, tapentadol, 27.0%, oxycodone, 23.3%. Because many researchers consider it better to report response rates rather than mean pain reductions (due to the distribution of pain scores), this large RCT could be interpreted to mean that OxyContin is no different from placebo.
When first reading this, my heart leapt and I was excited: I could start my own Broward County, Florida type “pain clinic.” You know, the ones on the OxyContin express, visited by people from several neighboring states, located in strip malls between a pawn shop and a tattoo parlor, where you can go and stock up on hundreds of pills in an unregulated and unmonitored fashion. I could go and rent a place in a strip mall right next to the Skin Cabaret and dispense gelatin capsules for $25 a pop! Then I remembered what Ralph Horwitz said and my heart sank again. That damn thing that happens when placebo response rates are high: the efficacy of known valid treatments is obscured. Of course, OxyContin is much more effective an analgesic than gelatin capsules, and it was the conditions of the RCT and the high attrition and the high placebo response rate that made the difference between them “non-significant.”
Funny word, “sham” acupuncture…
All the talk of “sham” acupuncture for RCTs in which the sham group had superficial or alternative needle placement, or toothpicks instead of needles, has rested on one critical, and from my observation, unexamined assumption. That “acupuncture” refers to an exact set of procedures done in accordance with a manual of some kind, such that superficial needle placement, rather than being called a variant procedure, is called a “sham” procedure. This is rather odd. It is as if you were listening to spine surgeons talking among themselves, and finding that some of them routinely used pedicle screws in certain situations and others did not. You might not think that they were discussing variations on a single kind of operation, and might think that some were doing real and some were doing sham fusions. After all, if they do not do them all exactly the same way, then what else are you to think? The guys who use pedicle screws are doing “true” fusion and the others are doing “sham” fusion, right? Wrong! These are variations of fusion, not qualitatively different procedures!
The Ernst paper mentioned Cherkin (who used toothpicks for the sham acupuncture procedure); Cherkin had three acupuncture groups, one individualized, one standardized, and one sham; the fourth group had usual care. The three acupuncture groups were all superior to usual care, and the percentages of patients with 3 point improvements on the Roland-Morris disability scale at the end of 6 months for the four groups were 62%, 58%, 58%, and 44%; usual care was significantly different from the three acupuncture groups, but those three groups did not differ from one another. With Cherkin, the usual care group “received no study-related care—just the care, if any, they and their physicians chose (mostly medications, primary care, and physical therapy visits).” Cherkin mentions Haake 2007 and Brinkhaus 2006, notes that they are all coming up with similar results, and ponders on the meaning of it all.
Sham acupuncture better than usual care, and about as good as true acupuncture–strong evidence, as Cherkin sees it, that this pattern points to a real phenomenon. All three authors discuss the small clinical significance of the details of needle placement. Most interesting of all, Cherkin says that the use of blunt needles was discussed in a classical Chinese book on acupuncture needling. Because penetrating the skin has been construed as something essential to acupuncture (otherwise you couldn’t call it “puncture”), the equal effectiveness of the toothpick sham procedure has been interpreted as devastating to the hypothesis that acupuncture is an effective treatment. If Cherkin’s history is correct, this finding need not prove anything dismaying to the practice of acupuncture
Main point: perhaps all the debate over “true” and “sham” acupuncture will turn out to be a debate over different variations of the same intervention. Just as the guy who operates without pedicle screws is doing true fusion, not sham fusion, the guy who uses toothpicks (AKA blunt needling) could turn out to be doing a variation of what the guy does who uses needles.
Now for the big irony:
For all three studies, I see a graph with three bars: the one on the left is true acupuncture, the one in the middle is sham acupuncture, and the one on the right is standard care. The two bars on the left are about the same height, and both are noticeably higher than the one on the right. If these were studies of a drug, I would conclude that the ones on the left worked better than the one on the right, and represented a similar class of drug. They are higher, and are about the same size; they appear to be similarly effective, and better than the third one with which they are being compared.
Others look at the same three graphs, notice the bar on the right, and see that, even though it is shorter than its neighbors on the left, the latter two are about the same height; they say, “Aha! I told you the ones on the left did not really work. Look! They are about the same size! I know this is true because I understand the one on the right but not the ones on the left!”
These studies are an examination of some long-held assumptions about acupuncture technique, and this examination is long overdue.
These three studies, all of them well done, lead the authors to do what one would expect of scientific physicians: re-examine some of the assumptions of their practice, particularly in the emphasis traditionally placed on the exact location of the needles, and on the need for depth of penetration. More superficial penetration may be just as effective as the deeper penetration in the “true” acupuncture tradition. It is likely to be safer.
Perhaps these studies will lead to changes in how acupuncture is practiced. It is just a matter of getting acupuncturists to abandon long-held practice patterns and entrenched habits. No problem.
What is so hard about that? After all, when needle biopsy was shown to be just as effective as open biopsy for tissue diagnosis in suspicious mammograms, all the surgeons changed their traditional practices just like that, didn’t they!
Uh-oh. Wait a minute…
The person making the claim bears the burden of defining his terms. The people who believe acupuncture has some benefit defined “acupuncture” as a procedure involving needles penetrating the skin along “meridians.” You might call this an “assumption,” but the TCM guys would not accept that. They have many books with complex line drawings and instructions regarding needle placement. They charge people a lot of money to learn this material.
Using points apart from the qi lines is a way to test the hypothesis about needle position in pain relief.
Because the sham procedure worked about as well as the true procedure, we can say that needle position is not important. The TCM guys got this bit wrong.
Next up: needling verses poking the skin. TCM guys got this wrong also.
Given these failures, a reasonable person should conclude that the TCM guys don’t actually know what they’re talking about. If they say anything useful it will be sheer luck rather than understanding.
You know what else sometimes works by sheer dumb luck? Dictionary fishing. That’s when you close your eyes, ask a question, then pop open a dictionary at random. You drop your finger on the page and open your eyes. The closest noun to your finger is your answer.
You want to do a bunch of studies on some treatment that rests upon a method of investigation about as reliable as dictionary fishing?
That has been pretty well addressed by Ben Kavousi’s posts on the topic. I suggest you read those before you continue.
Once again, if you look at Ben’s posts that will help you understand why it is reasonable to do so. It is very important to define what you are looking to measure and what your measure is. If you have a large enough sample of data and try and work backwords to find “something” in the data, you always will. A more accurate anology would be if you have 3 bars of data, one with a Drug X in a gelatin capsule, one with just an empty gelatin capsule identical to that used to contain Drug X, and one with a placebo sugar pill. You look at the 1st two and see they are identical and the last bar is much smaller. Then you would say that indee, Drug X doesn’t actually work. You would be lead to think that it is the capsule. Ergo, acupuncture doesn’t work – it is the wrapper that it is in.
Because a spinal fusion is not defined by putting in a pedicle screw. It is defined by the outcome in terms of measurable and observable differences in the vertebra and using a pedicle screw is one method of achieving that. Read the actual definitions of what acupuncture is, how it is actually taught, and how it is actually practiced. That is “true” acupuncture. “Sham” would therefore take away the integral part of “true” and thus offer a suitable comparison group. Thus, all you are left with is the placebo and non-specific effect of the ritual associated with it – nothing of “acupuncture” is left behind.
Dear Harriet,
as a founder of Science Based Medicine you should absolutely write to Gen. George W. Casey Jr. US Army’s Chief of Staff to stop this medieval bloodletting medical activity against US ill soldiers.
At the moment US Army is the most powerful army in the world. Maybe because of a sense of inferiority from the Yellow Face is US Army seriously thinking to acupuncture as a pain-reliever method for its soldiers?
http://www.army.mil/-news/2010/11/01/47448-acupuncture-gains-momentum-in-army-health-care/
The father of (western) Battle Acupuncture is Richard C. Niemtzow, Colonel, MD, Consultant of Alternative and Complementary Medicine to the Air Force Surgeon General. His works are these:
http://www.n5ev.com/battlefield_acupuncture.htm
http://www.n5ev.com/PDF%20BATTLEFIELD%20UPDATE.pdf
As a direct descendant from an ancient Roman General http://en.wikipedia.org/wiki/Claudius_Marcellus, whom should I believe and trust? A Science Based Medicine or an Intelligence Based Army, to which scientists of all times gave their best discoveries (from Archimedes’s Mirrors, to Leonardo’s cannons and Einstein’s bombs?)
As I am here to learn to be a honest skeptic with no trespassing into the cospiracy theory that the introduction of acupuncture in the US Army is an idea of BigPharma, please answer me: could I call Colonel Niemtzow a medieval bloodletter?
As so friendly as military
,
Stefano Marcelli MD
and ex captain of (NATO) Italian Army
My observations on Gross Anatomy and Acupuncture are here:
http://www.geneticacupuncture.com
Acupuncturists believed acupuncture worked.
They wanted to validate their beliefs with science.
They devised the best studies they could think of.
They devised a sham acupuncture procedure that they had every reason to think was a good placebo control. Indeed, patients couldn’t tell which they were getting.
The studies showed no difference between their sham and true acupuncture.
They were surprised and shocked.
Since they believed that acupuncture worked, the only way they could salvage their belief was to postulate that sham acupuncture worked too.
Then, after the fact, they came up with hypotheses about the depth of stimulation, etc.
Reminds me of the patient who thought he was dead. Doctor asked him if dead men bleed; he said no. Doctor pointed out that patient was bleeding. Patient said “Wow, dead men DO bleed!”
There’s a name for this logical fallacy.
“You might call this an “assumption,” but the TCM guys would not accept that. They have many books with complex line drawings and instructions regarding needle placement. They charge people a lot of money to learn this material.
Using points apart from the qi lines is a way to test the hypothesis about needle position in pain relief.
Because the sham procedure worked about as well as the true procedure, we can say that needle position is not important. The TCM guys got this bit wrong.
Next up: needling verses poking the skin. TCM guys got this wrong also.
Given these failures, a reasonable person should conclude that the TCM guys don’t actually know what they’re talking about. If they say anything useful it will be sheer luck rather than understanding.”
“Read the actual definitions of what acupuncture is, how it is actually taught, and how it is actually practiced. That is “true” acupuncture. “Sham” would therefore take away the integral part of “true” and thus offer a suitable comparison group. Thus, all you are left with is the placebo and non-specific effect of the ritual associated with it – nothing of “acupuncture” is left behind.”
Aha! I think I see part of the problem. Our perspectives may not be all that different after all. When I have been reading these well-done studies and concluding that they support the use of “acupuncture,” you have been thinking that I am concluding that they support the use of TCM. We may have been talking at cross purposes.
I rated the Brinkhous, Haake, and Cherkin studies as good quality because they do what I ask of RCTs of other kinds of intervention in controlling the risks of bias that threaten internal validity. I rated Haake the highest quality, because he did something that is often lacking in RCTs of pharmacology. In Table 3, he checked the adequacy of blinding, and found that most patients did not correctly identify or did not know whether they had received true or sham acupuncture. (Yes, I know, CONSORT 2010 has deleted this from their guidelines for reporting RCTs, but Haake was working in 2007 and his Table 3 shows that he was making every effort to control bias).
In addition, he did something at the very beginning of the trial similar to what they did in SPORT when they were studying surgical and nonoperative treatment of herniated discs: he had a separate cohort study of acupuncture which prospective participants could choose to enter. This was done so that patients with positive expectations of acupuncture and negative expectations of conventional therapy could be excluded from the outset.
Each of these authors, in their discussion sections, speaks of the need to re-examine traditional assumptions about needle placement. Earlier, I praised them for doing so, and for seeking less invasive methods to do acupuncture. This should have dispelled any impression that I was sticking up for TCM, but my communications are not always as clear to others as they are to me.
I used to see patients who had gone to TCM practitioners for their health complaints, and the impressions they gave me were not favorable. They would be given bags of herbs with no explanation of what was in them, how they worked, or how to recognize if they were not working. When they returned to their TCM “doctors,” they would be given another bag of herbs, again with no explanation of what they were or what they did. TCM looked like a rigid, closed, authoritarian system of practice. And the replies of GLaDOS and nybgrus show that they and I are of one mind with respect to that system. Those TCM docs were pursuing the worst habits of the old paternalistic “Doctor knows best; don’t ask questions” of some practitioners of scientific medicine, with none of its virtues.
Cherkin, Brinkhaus, and Haake, in their eagerness to open up a closed system, are major threats to the rigid system of TCM that uses ancient charts of meridians and qi to dictate practice, charging lots of money for teaching this system. I was about to add this to my last post, but it was already running a bit long. If these authors go ahead with the research they want to do, it is safe to say that their strongest opposition will come from the ranks of TCM. These guys are threatening their territory, and they will not like that one bit. Like any other authoritarian system, TCM says that there is only one right way to do things, and they are the custodians of that right way. GLaDOS, nybgrus, and I are on the same page with respect to that. I think.
More than one right answer, of course, does not mean that “anything goes.” It does not mean that any random or arbitrary answer is as good as any other answer. Quadratic equations have two right answers and cubic equations have three right answers, but that does not mean that anything goes; it only means that the space in which they are represented in analytical geometry has more than one dimension.
Saint Augustine gave an example of an eternal truth: seven plus three is ten, has always been ten, and will always be ten. In setting up the discussion of truth in this manner, he laid the ground for a rigid dogmatism which had an inflexible model for considering what it means to look for truth. Most of the time, we are not confronted with problems of the form “What is seven plus three?” We are confronted with problems of the form, “Which two numbers, added together, equal ten?” Here, if the answers are constrained to be positive integers, there are only a few right answers, but more than one. If constrained to be rational numbers, there are an infinite number of right answers, but not all answers are right.
TCM authoritarians have a real problem with that. They will freak out if Haake and his colleagues proceed in the direction they are going. You can absolutely, positively bet your virginity on that, as my old first sergeant used to say.
“They devised a sham acupuncture procedure that they had every reason to think was a good placebo control. Indeed, patients couldn’t tell which they were getting.
The studies showed no difference between their sham and true acupuncture.
They were surprised and shocked.
Since they believed that acupuncture worked, the only way they could salvage their belief was to postulate that sham acupuncture worked too.
Then, after the fact, they came up with hypotheses about the depth of stimulation, etc.
Reminds me of the patient who thought he was dead. Doctor asked him if dead men bleed; he said no. Doctor pointed out that patient was bleeding. Patient said “Wow, dead men DO bleed!”
There’s a name for this logical fallacy.”
I don’t know how shocked they were, and cannot guess what they wanted to salvage, not having ESP and lacking the ability to channel thought waves. I propose another possibility: their data showed that sham acupuncture was better than standard care, and they concluded that there is more than one right answer to the question about how needles should be used and placed. A shock to TCM, perhaps, but not to anyone else.
What exactly did Cherkin, Brinkhaus, and Haake do wrong?
That bar graph representing the response to conventional therapy (the short one), to the right of the two bars representing true and sham acupunture (the taller ones), remains to be examined and explained. For some reason I am reminded again of my army days, when the drill sergeant used to lead marching cadence with, “Look to your right and what do you seeeeeeeeee”
@ Ed Whitney, “their data showed that sham acupuncture was better than standard care, and they concluded that there is more than one right answer to the question about how needles should be used and placed.”
I just don’t get this. Why should they conclude that? Why not conclude that the reason both interventions were better than standard care was that both interventions evoked a stronger placebo response than standard care? That is the most parsimonious and logical explanation of all the observed data.
Nope. Acupuncture is the procedure under discussion, not some other TCM procedure.
BTW, if you weren’t sure whether I was talking about TCM or acupuncture, you might have asked and so saved yourself a teal deer (tl;dr aka “too long; didn’t read).
Allowed potential conclusions must be stated PRIOR to any study to avoid the sharpshooter’s fallacy. Once you go looking for patterns in the data AFTER the study is over, you’re starting from scratch as if there had been no study.
You can use patterns in study results to generate new hypotheses. But you can’t use those patterns to draw conclusions.
“Allowed potential conclusions must be stated PRIOR to any study to avoid the sharpshooter’s fallacy. Once you go looking for patterns in the data AFTER the study is over, you’re starting from scratch as if there had been no study.
You can use patterns in study results to generate new hypotheses. But you can’t use those patterns to draw conclusions.”
Fair enough; I will be as brief as possible. Haake and Brinkhaus were done in Germany and had no registration of their trials on clinicaltrials.gov. Cherkin was done in the US and the trial registration is at the end of the abstract. I make it a rule to look at the trial protocol to rule out the Texas sharpshooter fallacy, when those protocols are available. http://clinicaltrials.gov/archive/NCT00065585/2005_06_23 has ruled this out.
From your having said, ” You might call this an “assumption,” but the TCM guys would not accept that. They have many books with complex line drawings and instructions regarding needle placement. They charge people a lot of money to learn this material,” I thought you were somehow talking about TCM. But I did not ask if this were the case.
Busted again.
>90% of mechanical back pain, even recurrent pain, will eventually go away. The chronicity of back pain in these circumstances and it is a well understood phenomenon that it has to do with expectation and incentive. Patients either expect that they will not get better, are afraid of their back pain and too much movement and thus perpetuate their sick role or they have incentive to continue with the sick role from worker’s comp, insurance payments, or even how their family treats them because of their back pain. I actually spent a week in med school learning about these things so that we can adequately handle these issues and avoid contributin to chronicity.
So your study specifically includes only a population defined by a self limiting disease process, which has been shown would be ameliorated by expectation of getting better (i.e. acupuncture is new maybe it will work!) and incentive to get better (they’ve enrolled in the study, maybe they want to get better but have been afraid of their back pain). The point being, you are looking at a study on a process in a population that is expected to have a variable course but with a measured outcome we expect to get better regardless of the treatment modality.
I have this awesome tea you can drink. It works for almost everyone and it will cure your cold/flu in 7-10 days. Maybe we should study it?
“So your study specifically includes only a population defined by a self limiting disease process, which has been shown would be ameliorated by expectation of getting better (i.e. acupuncture is new maybe it will work!) and incentive to get better (they’ve enrolled in the study, maybe they want to get better but have been afraid of their back pain).”
This was precisely the point I tried to make earlier, that the principle of “extraordinary claims require extraordinary evidence” does not apply. A self-limiting condition resolves faster with slender needles than with NSAIDS and PT. This claim, mundane though it is, should be supported with mundane evidence, and that has been done. Rule out chance, confounding, and bias, and you are entitled to attribute the faster resolution to the study intervention.
“I have this awesome tea you can drink. It works for almost everyone and it will cure your cold/flu in 7-10 days. Maybe we should study it?”
Cold/flu remedies are judged effective if they can cure you in 4-6 days. If you can resolve a self-limiting condition more efficiently with one intervention than with a comparison intervention, you have done something useful. Won’t get you a Nobel or Lasker, but worth someone looking at.
” Why not conclude that the reason both interventions were better than standard care was that both interventions evoked a stronger placebo response than standard care? That is the most parsimonious and logical explanation of all the observed data.”
“Placebo” is an explanatory principle, belonging to the world of concepts. It lacks a concrete referent in the world of extra-verbal reality. To speak of “placebo” as if it had an extra-verbal referent is, it seems, to entertain the Fallacy of Misplaced Concreteness.
The Hanover Functional Ability Questionnaire (the functional outcome of Haake) uses language, but uses it to refer to things in the extra-verbal world. After receiving the needle-based intervention, more patients were able to reach for books on high shelves, carry a 10 kg suitcase at least 10 meters, wash their hair over a washbasin, pick a crumpled piece of paper up off the floor, stand for 30 minutes without interruption, lie in bed on their backs, and run 100 meters to catch a departing bus, than were able to do the same things with conventional therapy. To measure and compare these things is to engage in Appropriately Placed Concreteness.
@Ed Whitney,
Placebo “lacks a concrete referent in the world of extra-verbal reality”?!
If you want to play word games, I’ll reword my question to “Why not conclude that the reason both interventions were better than standard care was that the nonspecific treatment factors in both resulted in patients reporting less pain than when they received standard care? That is the most parsimonious and logical explanation of all the observed data.”
I agree that objective endpoints are more meaningful than subjective reports of pain intensity. I have criticized studies that reported pain relief but didn’t result in patients using fewer rescue pain pills. But does that mean we can disregard all the studies that have used patient reports as endpoints? I don’t think so.
ack! sorry for the mistake in formatting:
argh. OK. I give up. sorry for the formatting. Hopefully it is still readable.
Damn! I think we may be getting close!
I have been saying that Brinkhaus, Cherkin, and Haake (henceforth to be known as BCH) have shown something interesting: by poking someone, they can hasten a person’s ability to lift a jug of water from the floor to a table with ease. Further, they have shown that there is more than one way to poke a cat. I insist that it has not been shown that any old kind of poking will do; that was what I meant by all that stuff about higher order equations having more than one solution, but not just any solution.
You have been trying to get me to see that the effective intervention is not acupuncture. I think I am pretty close to agreeing with that, at least insofar as it is an integral part of a system of thought we have been calling TCM. You call it woo, I call it folderol, but we are not all that far apart on this issue.
In saying that BCH were going to get serious pushback from the TCM community, I was in essence saying what I think you are saying when telling me that the BCH intervention is not acupuncture. I am not sure that BCH actually realize the significance of what they have done. They may be on to something entirely new, thinking that they are defending something old.
The dividing point between us may be how we respond to the phenomenon we are looking at. In calling “placebo” an explanatory principle, I meant that an explanation of a phenomenon is a way of setting the mind at rest. A guy can hardly put on his socks without great difficulty, Haake tweaks him with a needle, and soon he can put them on with ease, sooner than he could have if he had just kept taking Aleve and doing his pelvic tilt exercises. “Placebo” is an explanation that places this into the category of “what we already know.” Further inquiry is truncated. This is not a word game, but has real consequences for what comes next.
But I want to say, “These guys are on to something! I wonder what it is?” Further inquiry then begins. Maybe anyone can poke the guy with the same favorable result, but I will insist on good evidence for that. Someone may make that claim, but until that is proven, to me they will sound like the yahoo at the art museum who says his kid can paint a Jackson Pollock. BCH are all trained acupuncturists, and what they are doing has my full respect.
If low back pain were a rare condition that had no effects on work absenteeism, chewed up a trivial amount of physician time, and had no impact on health care system costs, I would be less inclined to propose that the BCH phenomenon is worthy of further research. This, judging from most of the comments on this thread, is not the case; further research is wasted effort because the effectiveness of acupuncture is a question that has been settled. You may as well waste your research resources on the design of a perpetual motion machine. That is what I have been getting from the SBM community so far.
If I were unfortunate enough to be a health care system administrator, and there was a possibility that a new intervention could have an effect on my total system costs, I would be motivated to give BCH a grant to develop an intervention that may save the system money and get me a giant bonus.
Are we any closer to common ground?
“by poking someone, they can hasten a person’s ability to lift a jug of water from the floor to a table with ease.”
Maybe they are only hastening the person’s willingness to try harder to lift a jug.
Could be. Or to change their answer to “Können Sie einen schweren Gegenstand vom Boden auf den Tisch stellen?”
Question is: are you curious to know more?
Sure I’m curious to understand more about the placebo response, but I don’t think studying variations of acupuncture and claiming there is some kind of specific effect is a fruitful approach. It is the placebo response itself that should be studied, as Benedetti’s group is doing.
A litte closer perhaps. The problem is that you are stipulating there is something to the “tweaking with a needle” that is helping this person. That is clear. As Dr. Hall has said, it may be that the person is just willing to try harder to lift the jug. That is hardly a trivial point. However, when you are looking at the utility and implementation of an intervention, as we have said, you cannot take a large amount of data and try and connect some dots. You must develop a hypothesis and test it and determine its sigficance. So what do we have?
Needles aren’t necessary, toothpicks work as well.
Specific points as dictated by TCM aren’t necessary, any poking will do.
It doesn’t seem to work on non-self limiting processes – a person with paraplegia won’t walk again and it won’t cure diabetes.
So what are we left with? What hypothesis would you put forth to test and develop a treatment that isn’t TCM acupuncture (since you draw that delineation) but is BCH acupuncture? What is it that separates them?
The SBM crowd is dismissing these because we advocate a Bayesian, a priori, basic sciences metric to be added to our analyses and experiments. Since the needles aren’t necessary, the points are not defined, and there is no plausible mechanism for the acupuncture that you and I are both (apparently) agreeing to dismiss, then what kind of experiment could you design that would shed new light on this? Countertraction/irritation? Been done. Distraction? Been done. Placebo? Understood to be able to motivate and have measurable effect. So what is left? Focus on that question and try and frame a hypothesis from the beginning and come up with an experiment. Don’t start with your conclusion and work backwards.
Furthermore, the phenomenon you describe is actually described, noticed, taught, and becoming more and more understood. The chronicity, social factors, personal factors, placebo effect, motivation, etc etc vis-a-vis back pain are understood to be a complex milieu. Thus when all the facets of acupuncture can be changed with no change in effect, we are left looking for the parsimonious conclusion. Dr. Hall said it quite well: “Maybe they are only hastening the person’s willingness to try harder to lift a jug.” It has been shown that even when painful, pushing through the pain to regain mobility is the best way to alleviate mechanical back pain. I have personally experienced this myself. But people are usually scared to do so – back pain hurts a lot. Give them a placebo to motivate them and voila! You have your explanation and it further shows that there is nothing to acupuncture.
Nothing unique anyways. There are many ways to poke a cat, you say. There are many ways to placebo a cat as well. In medicine, when there is no difference like that, you must always choose the method with the least amount of side effects. And when you know your treatment is reduced to nothing but placebo, you cannot ethically use it on your patients. That, Ed, is why the SBM crowd is dismissing acupuncture.
Let’s try this.
For non-specific effects to do “something”, there must be physiology that couples the non-specific effects to what ever “something” is happening. We may not know what that physiology is, but I don’t think that anyone at SBM would deny that there must be physiology to produce the “something” that is observed.
nybgrus is correct. The only way that adherents of SBM can use placebos is if they occur only as a result of actual physiological relevant treatment. A good bedside manner is a placebo, it is an ethical placebo, it is actually unethical to not use a good bedside manner.
If there was a pharmacological treatment that did trigger the placebo effect over and above the placebo effect of a good bedside manner, then it would be ethical to use that pharmacological treatment provided the risk was low enough.
What should be looked at in acupuncture research is the comparison of acupuncture with an equivalent time of “good bedside manner” instead of “standard care”.
I think that spending equivalent time with a real doctor going over the patient’s symptoms, taking history, blood pressure, discussing diet, exercise, stress reduction, that the “effects” would be equivalent to the acupuncture and the sham acupuncture groups, or maybe even superior because the good bedside manner also includes advice on other things which could help in addition to its placebo effects.
No. You keep saying “acupuncture.” But there is no acupuncture, just nonspecific stuff.
Moreover, you miss the big picture. The whole point of “research” in the alt med world is to keep the wolves of real science away from the front door. The alties have already built their schools, award their degrees, vertically integrated their manufacturing and distribution streams across international borders, and formed their professional societies and accreditation committees. The “studies” are just part of their marketing budget.
The alt med schools, degrees, and products must be scrapped. Too many young people are seduced into throwing their lives away on worthless degrees. They then see no escape, apart from perpetuating the lies that were told to them.
@ nybgrus
“A little closer perhaps. The problem is that you are stipulating there is something to the “tweaking with a needle” that is helping this person. That is clear. As Dr. Hall has said, it may be that the person is just willing to try harder to lift the jug.”
I am not at all certain that the tweaking of the needle was helping the person, only that the BCH intervention was causal. Basically, I made that causal inference based on having ruled out chance, confounding, and bias, and having satisfied myself that the observed difference in response rate between BCH and standard care did not arise from post-hoc cherry picking of data (the Texas sharpshooter problem). The causal inference applies to the entire BCH and I am not ready to defend attribution to a single component.
“…any poking will do.”
Not at all proven; the sham poking in BCH has all been planned in advance and done according to protocols.
“So what are we left with? What hypothesis would you put forth to test and develop a treatment that isn’t TCM acupuncture (since you draw that delineation) but is BCH acupuncture? What is it that separates them? ”
I have been wondering the same thing and have nothing approaching an answer. Serves me right for knowing next to nothing about classical TCM acupuncture, never having been much attracted to it. All I know is that in their sham interventions, they inform me that they are departing from their customary procedures in various ways. They place the needles away from acu-points, but I do not know what that means. Are they placing them in the same dermatome? So I would postpone making hypotheses until I was preparing to design an original study of BCH.
“Dr. Hall said it quite well: “Maybe they are only hastening the person’s willingness to try harder to lift a jug.” It has been shown that even when painful, pushing through the pain to regain mobility is the best way to alleviate mechanical back pain.”
This is likely to be an important effect of BCH; fear-avoidance is a major part of what moves back pain from a self-limiting to a chronic condition. Getting up and about and doing things is the cure of nonspecific back pain, and all other interventions are justified insofar as they contribute to that end. Passive modalities, whether ultrasound, massage, or BCH, are only adjuncts to active therapy. Cognitive-behavioral therapy and exercises was just as good as fusion with pedicle screws (Brox, Ann Rheum Dis 2010;69:1643–1648) for chronic low back pain in one well-done study.
“It has been shown that even when painful, pushing through the pain to regain mobility is the best way to alleviate mechanical back pain. I have personally experienced this myself. But people are usually scared to do so – back pain hurts a lot”
Ditto. See above.
“Give them a placebo to motivate them and voila! You have your explanation and it further shows that there is nothing to acupuncture…the SBM crowd is dismissing acupuncture.”
Insofar as “explanation” truncates inquiry, it should not be indulged in until we know why BCH was a better placebo than the evidence-based guideline care with which it was compared. Dismissal suffocates curiosity. There is no justification at this juncture for saying “voila!” Enough for saying, “Mon dieu! Qu’est-ce que c’est?”
@ daedalus2u
“I think that spending equivalent time with a real doctor going over the patient’s symptoms…”
Even before managed care came along and strangled the doctor-patient relationship with its demands for “productivity,” the time for talking and listening was all too limited. Better stop here before beginning to rant and rave about this real tragedy of our time. My doctor has advised me against it.
@ GLaDOS
“No. You keep saying acupuncture.’ But there is no acupuncture…”
Hence the switch to “BCH intervention.” Anyone who says “acupuncture” in the future will be sacked. And those responsible for sacking them will be sacked.
“…just nonspecific stuff.”
I have been tossing this term about and am in no position to cast stones, but “nonspecific” is a word without a scientific meaning. Like “idiopathic,” it is a confession of ignorance, and not an explanatory principle any more than “dormitive power” of opium in Le Malade Imaginere.” At least Moliere knew he was jesting.
“Too many young people are seduced into throwing their lives away on worthless degrees”
This does bring up an important reason for any placebo effect. We all know that expectations are an important part of the placebo response. Friendly and cordial cultural attitudes toward an intervention (e.g., everyone likes CAM) create the context in which positive expectations are formed. This increases the response rate in the placebo group, which makes it more difficult to detect specific effects, especially when the outcome is measured on a closed scale (like pain scores).
If a celebrity or two died after a visit to the acupuncturist, these cultural attitudes might have to be re-calibrated. In the USA, they all use aseptic technique and do not operate like knuckleheads, so this is not likely to happen.
@ the whole gang
Related questions for anyone who knows: in many states, physical therapists are doing dry needling of trigger points, which is within the scope of their practice after they complete an accredited training course. Obviously no lidocaine or Kenalog or anything can be injected. There was no evidence that the injectate added anything to the dry needling last I knew.
Are you seeing this done in your area?
Do you have any concerns about its becoming widely practiced?
Is this to be considered CAM or is it conventional practice?
@Ed Whitney
“Sham acupuncture better than usual care, and about as good as true acupuncture–strong evidence, as Cherkin sees it, that this pattern points to a real phenomenon.”
This is absolutely not strong evidence of anything. The problem in such three arm studies is that the usual care groups aren’t blinded like the true and sham acupuncture groups are. It’s highly problematic at best to compare one unblinded group with two blinded groups and draw any particularly useful or definitive conclusions on anything other than the relative strength of the placebo responses.
To say otherwise seems to require ignoring or undervaluing the value of blinding.
Also remember, since standard care is not particularly effective for the pain typically being studied here, you’re closer to comparing acupuncture to no treatment rather than seeing how it measures up to an established, effective treatment.
More precisely, the acupuncture and sham acupuncture groups are blinded to each other, but there is no blinding between these two groups and the usual care group.
You can say conclude that both acupuncture-like treatments produced greater effect than the usual care group, but since the usual care is not particularly effective for the type of pain studied (and very close to no treatment), you can’t say the results aren’t due to placebo response.
The subjects are likely well aware that usual care is generally ineffective (which is probably why they are in the study), and have little or no expectation of relief from usual care. They are, if anything, primed for a “reverse placebo effect”, and thus the lack of blinding between usual care and the acupuncture-like meta group is particularly problematic.
As a side note, instead of randopoking, I should refer to the potential clinical application of sham acupuncture as pseudopuncture.
What do you mean, “not a scientific meaning”? It’s very useful in science to say, “this part of the variance is down to God-knows-what.”
Confessions of ignorance are all the rage in science. The mistake happens when you use “we don’t know” to justify some positive claim. Example: “You can’t say acupuncture has absolutely no effect. Therefore doctors should offer acupuncture as an option to their patients.”
From Karl:
“The problem in such three arm studies is that the usual care groups aren’t blinded like the true and sham acupuncture groups are. It’s highly problematic at best to compare one unblinded group with two blinded groups and draw any particularly useful or definitive conclusions on anything other than the relative strength of the placebo responses.
To say otherwise seems to require ignoring or undervaluing the value of blinding. ”
One asks an RCT to control sources of bias that can reasonably be expected to be controlled. Often participants cannot be blinded, but outcome assessors can be; each of the three trials did this. Lack of blinding does leave open some susceptibility to assessment bias, and investigators can be asked to do what they can do to minimize this.
Even though blinding sometimes cannot be done, I am unable to think of any trial in which randomization cannot be done, and in which concealment of allocation cannot be done. All three trials did this. Selection bias was avoided.
To repudiate any trial in which blinding is incomplete is to create a novel standard for evaluating the validity of clinical trials. Consider the SPORT trials of surgical vs. nonsurgical treatment of back conditions; the links (free access) are below. The first is the trial of herniated disc treatment; the second is for lumbar spinal stenosis, and the third is for degenerative spondylolisthesis .
http://jama.ama-assn.org/content/296/20/2441.full.pdf+html
http://www.nejm.org/doi/pdf/10.1056/NEJMoa0707136
http://www.nejm.org/doi/pdf/10.1056/NEJMoa070302
In none of the three trials could the participants be blinded, yet these studies were widely cited and were influential on the thinking of the spine community on the issues that were studied.
Should they be repudiated because of lack of blinding? This seems unreasonable and unnecessary.
Basic point: selection bias (randomization and allocation concealment) is always avoidable, and it is a flaw of an RCT not to accomplish it. Lack of participant blinding is not necessarily a fatal flaw.
“Also remember, since standard care is not particularly effective for the pain typically being studied here, you’re closer to comparing acupuncture to no treatment rather than seeing how it measures up to an established, effective treatment.”
“An established, effective treatment” for chronic LBP is the Holy Grail. If there were, the volume of literature devoted to RCTs of various interventions would be reduced, If you find the Grail, the world will beat a path to your door.
The best surrogate is care from an evidence-based practice guideline, which was Haake’s third comparison group. It included 10 sessions with a physician or physiotherapist who administered PT, exercise, and such. They improved (the usual care group improved in each of the three trials), just not as much as the two comparison groups.
“The subjects are likely well aware that usual care is generally ineffective ”
Their awareness of its ineffectiveness should be seen as part of its ineffectiveness, and not as a confounder. That is, their awareness is part of the causal pathway for the ineffectiveness compared to the other two randomized groups.
“pseudopuncture” FANTASTIC! I will shamelessly plagiarize! “There is good evidence from three randomized trials that pseudopuncture was more effective than conventional care in blah blah blah. ”
From GLaDOS:
“Example: You can’t say acupuncture has absolutely no effect. Therefore doctors should offer acupuncture as an option to their patients.’”
But, as per the above, I am claiming that the a-word has an effect, and that doctors should be willing to offer it as an option to patients who express an interest. You can’t say it has absolutely no effect because patients who got it fared better than patients who did not.
As mentioned earlier, I think that Dr. Hall is correct to suspect that the effect of BCH intervention on hastening the ability of the patient to resume activities of daily living has to do with their willingness to push through the discomfort. A survey of a convenience sample of physicians specializing in occupational medicine and physical medicine (i.e., a coupla guys I know) suggested that patients who went for acupuncture were more willing to get up and get moving again than before that intervention.
It was not clear whether she was meaning this as a way of dismissing acupuncture, but since active movement is the cure for nonspecific LBP, this effect of acupuncture should be seen as a direct therapeutic effect, and not as a confounder. It is because they pushed through and moved again that they were able to reach up and get things off of tall shelves, pick things up off the floor, run for the bus, and all the other functions that were in Haake’s outcome.
This seems reasonable: poking the patients makes them more willing to do stuff in spite of pain, which leads to their recovery. I think this is where the discussion was leading earlier. The willingness is an intermediate in the pathway from the needle to the resolution of symptoms. It is not an extraneous factor; it is the essential factor.
To say, “well, it only made them better because it increased their willingness to push through the pain” seems like saying “the only reason the antibiotic resolved the pneumonia is because it killed all those bacteria, so that the patient’s immune system could eliminate the disease organisms.”
@Ed Whitney,
I was not “dismissing” acupuncture. I was pointing out that part of the placebo response is the expectation and that encouragement leads people to try harder on objective tests. Is this enough to justify knowingly prescribing placebos? The consensus of medical ethicists is that it doesn’t. If I gave the patient a sugar pill and told him it was a narcotic, he might feel better and try harder too, but I would be wrong to lie to him.
Almost there Ed…. you are getting closer to realizing what you need to do now. Toss the whole idea in the trash!
Ed Whitney
“To repudiate any trial in which blinding is incomplete is to create a novel standard for evaluating the validity of clinical trials.”
I never did any such thing. Of course, in some studies blinding is not always possible, even though it is pretty much always desired. I never disputed that, though you apparently inferred that I did.
My point was, in case you missed it, why bother blinding the acupuncture and pseudopuncture groups in regard to which of those two groups they are in if there is no blinding between the acupuncture like meta group and the usual care group?
The lengthy section about accessory blinding, randomization, and, selection bias in your reply was a response to a criticism that nobody made, but you glossed over the point that the study had inherently inconsistent participant blinding. Every participant knew whether they were in an acupuncture like group or the usual care group, but the people in the acupuncture like groups did not know if they were in the acupuncture or pseudopunture group. If you are going to use the usual care group as anything but a baseline for the measure of the strength of the placebo response, then it really should have as close to the same amount of blinding as the other two groups as possible.
Must kill italics!
Rats! It didn’t work.
Awwww! Just when I was on the verge of declaring victory on this thread!
Main reason to think it ethical to offer acupuncture as an option is the comparison of the three bar graphs from the three LBP studies, where the height of the bar represents therapeutic response. The three bars are labeled A, B, and C. A and B are higher than C in all three studies. The studies have done everything that can be asked of them to control sources of bias that can be controlled. Bars A and B are the same height, and both of them are higher than C.
If these bar graphs were presented to an observer without labels as to which interventions they represented, that observer would be expected to say that A and B were more effective than C. If the morbidities of A and B were no greater than C, that observer would say that it is ethical to offer A and B as an option to a patient whose condition had not resolved under intervention C. An observer who said the contrary, that C was ethical but A and B were not, would be defending a position which would require much additional explanation if most ethicists were to be convinced.
If the interventions under consideration were anything other than acupuncture, I would expect most ethical clinicians to be willing to offer A and B as options. If A and B were analgesics with acceptable side-effect profiles, I expect that most SBM folks would endorse their use.
To switch opinions after the labels are revealed to deal with acupuncture and pseudopuncture (a tip of the had to Karl), and to say that offering A is now unethical, seems to partake of the post hoc data re-evaluating that would be unacceptable under most circumstances. It will not do to reverse course after finding out that A and B are interventions you have always opposed, but to stay the course if you find that A and B are interventions you have always approved.
I was not aware of a consensus among ethicists about the use of placebos, but was aware of a consensus against deception. Not exactly equivalent. Finnis et al (Lancet 2010;375:686-95) looked at these three selfsame studies, and suggest saying this to a patient who has not been helped by standard therapy: “I recommend that you try
acupuncture. Several large studies have shown that traditional acupuncture is not better than fake acupuncture treatment, but that both of these produce substantially greater symptom improvement in patients with chronic low back pain compared with those
patients who receive no treatment or conventional medical therapy. Although the specific type of needling does not seem to make any difference, it is possible that acupuncture works by a psychological mechanism that promotes self-healing, known as the placebo effect.” They suggest that this disclosure is honest.
BTW, if it is my patient, I would be heavily influenced by a paper that the selfsame Cherkin published in 2001 (Arch Intern Med 2001;161:1081-1088), comparing traditional Chinese medical acupuncture with therapeutic massage and self-care education. Here, when massage was part of an RCT and there was a massage protocol (in which considerable room was left for variation as to exact massage technique), massage was effective, with long-lasting benefits. Traditional Chinese acupuncture was relatively ineffective. The massage group used the least medication and had the lowest cost of subsequent care. The big difference between this study and the other three (Cherkin, Brinkhaus, and Haake) looks to be in the nature of the comparison group. Rather than comparing acupuncture with a nebulous “usual care” or “standard care,” the comparison is with an intervention that was designed into the trial itself.
Having trained in massage myself at the Esalen Institute long, long ago, I would push massage before discussing acupuncture. Of course, the participants in this study were not blinded, but I trust that there will be no objections raised on that account in this instance!
@Ed Whitney,
You have fixated on those bar graphs. They don’t mean what you think they mean.
“The studies have done everything that can be asked of them to control sources of bias that can be controlled.”
No they haven’t. They didn’t control the bias between “treatment with impressive rigamarole” and “conventional boring treatment.”
@Ed Whitney,
“I would expect most ethical clinicians to be willing to offer A and B as options.”
Not if B was designed to be an inert control, and not necessarily if there was no blinding between the A-B meta group and group C. If B was designed to be an inert control, and the results showed it was just as effective as A, the natural conclusion would be that A was also inert.
One could speculated that B might not be inert as intended, but you’d need a plausible mechanism whereby it might be otherwise, and then you’d need to design a new study to attempt to determine if B really had a clinically significant effect beyond a control.
If you’re not going to treat the pseudopuncture group as a control, then it’s really an unblinded comparison between a meta-puncture group and a usual care group.
In that case, we’ve seen plenty of similar studies of acupuncture before, and this one just tells us that regardless of whether or not acupuncture is bogus, the invasive needling and points used don’t really matter.
@ Harriet:
No they haven’t. They didn’t control the bias between “treatment with impressive rigamarole” and “conventional boring treatment.”
Let me check quickly before attributing meaing you may not intend.
Patient in acupuncture or pseudopuncture group experiences paresthesias during the course of the session, and, because of expectations of benefit, attributes them as evidence of a powerful therapeutic effect of the puncture, feels that recovery is happening, and gets up and does what needs to be done, hastening functional recovery.
Patient in conventional care group experiences identical paresthesias during PT or other aspect of usual care, and, because of belief that conventional care sucks, interprets same paresthesias as exacerbations of problem, feels that recovery is not happening, and curtails activity, delaying functional recovery.
Is this in the balll park?
In that case, we’ve seen plenty of similar studies of acupuncture before, and this one just tells us that regardless of whether or not acupuncture is bogus, the invasive needling and points used don’t really matter.
This the usual interpretation, but we can’t quite say that. The patients still have to believe that they are being needled. Perhaps there is anticipatory release of endorphins and other neurotransmitters.
The same could apply to the needling of more sensitive areas.
What I meant was that it is not fair to compare usual treatment to two procedures that both have a much greater likelihood of eliciting a placebo response. The usual treatment will always come off looking worse.
Your paresthesia scenario is a possibility, but it is more likely that the patient will not develop any specific symptoms from the treatments. He will simply expect to feel better where his expectations have been influenced by nonspecific factors.
If that doesn’t kill them, I don’t know what will.
You are indeed fixated Ed. However you are incorrect.
No, the puncture isn’t the important thing. If it were, skeptics could enjoy some pain relief from being poked.
What is it then? The disrobing? The comfy massage table? The time spent relaxed while a nice professional paid attention?
No.
If you are very good and cause no serious trouble while I am away, later on this evening I may share with you an important insight. But you must brace yourself for transformation.
Groups A, B, and C have positive expectations of acupuncture. A and B get acupuncture and pseudopuncture while group C gets usual care. The comparison of groups A and B is unbiased. The comparison of A with C and of B with C is biased because the groups are unbalanced with respect to one very important predictor of the measured outcome: their expectations of their allocated treatment.
The randomization (assuming all that computer-generated allocation sequence, etc) results in a design that balances the groups with respect to their expectations of acupuncture. The study can get 9 out of 10 possible Cochrane Brownie points, and be assessed as a high-quality study, but still have an important bias that can influence the interpretation of the outcome.
The groups are balanced on one logical level (interest in acupuncture) but unbalanced on another logical level (interest in their allocated treatment). The latter imbalance creates a biased comparison of outcomes across the bar graphs. The measured differences between A vs. C and B vs. C arise from differences in expectation and not differences in specific efficacy of the interventions.
Sure this isn’t what you are driving at?
You seem to be completely avoiding what I am driving at Ed
@ nybgrus:
Sorry, I forgot to say “@ Harriet.”
I am trying to see if I am in the same ball park with her.
@ Harriet:
Maybe more succinctly: Validity means that a measurement actually measures what it purports to. A measurement of response to an expectation is an invalid measure of a response to an intervention. The comparisons of A and B with C measure the former and not the latter.
Any closer?
“This is higher order thinking that is required of us as physicians and goes well beyond graphs A, B, and C.
Now, let’s not be too smug about our supposed intellectual superiority.
There is much for us to be sheepish about if for any reason at all we are actively trying to deny patients potential benefits, especially if in clinical settings where science-based medicine lacks adequate answers.
I, too, think that Ed is going beyond present plausibilities if he believes that acupuncture may yet be shown to have any special physiological activity.
But nagging questions remain after consideration of all the relevant evidence.
The neutral observer, for example, may still want to know “why not give acupuncture the benefit of a little doubt?. There are many different ways in which it might benefit individual patients. I mean, what is medicine for?”
Also: “are ethicists being asked the right questions?” Do they understand what the science actually shows, i.e. substantial effect sizes from so-called placebos over standard medical care, even equivalent outcomes but with fewer side effects than with some drugs e.g. in the prophylaxis of migraine?
Are they fully aware of common contextual matters, including that these methods are mostly used for symptom relief when science-based medicine has no adequate answers?
These questions will only go away if you can show that the combination of placebo and non-specific influences that acupuncture and other regimes incorporate are of no value at all to users.
How do you think the evidence weighs up on that?
From pmoran:
“I, too, think that Ed is going beyond present plausibilities if he believes that acupuncture may yet be shown to have any special physiological activity.”
I gotta work on my English Composition more! I am not expressing a belief that acupuncture has a physiological activity which is relevant to low back pain.
I am not invested that much in acupuncture per se. I am greatly invested in sources of bias that remain in RCTs when they have crossed all the i’s and dotted all the t’s that go into making the Cochrane Risk of Bias Tables and the CONSORT 2010 reporting guidelines.
It was a while back that I alluded to the persistent bias with the acupuncture trials share: namely, that the patients who enter them are a special population, specifically, people who have an interest in acupuncture and are enticed to enter the trials in order to have access to it at the end if they are not originally randomized to it.
Usually we think of recruitment biases as affecting external validity. In the past few posts, I am looking at how they can threaten internal validity. Today A, B, and C stand for true acupuncture, sham acupuncture, and conventional therapy. Tomorrow, they may stand for something different.
Randomization may distribute expectations of a specific treatment uniformly across groups, but expectations of the allocated treatment differently. Percutaneous vertebroplasty could be studied in a three-arm trial with a true group, a sham group, and a group with usual care. The group labels A, B, and C would look the same as in acupuncture, but the analytical issues would be the same. If both vertebroplasty groups did better than usual care, we would be facing the same kinds of bias; the group which got the real thing and the group that got the placebo would be about equal, and both would be superior to usual care. If the patients all entered the trial with positive attitudes about the trial intervention, group comparisons could be biased even if the trialists did everything right. The less favorable measured outcomes in the usual care group could be measures of disappointed expectation and not of anything to do with an intervention on the vertebral body.
Vertebroplasty is a very hot potato issue right now. Acupuncture is chicken feed in comparison. More clinical trials are coming out in the future. The content will change, but the questions of process will persist. Many of the arguments will follow a similar form. It is the form, and not the content of the debates, that is of greatest interest to me.
“…not be too smug about our supposed intellectual superiority? ”
Good suggestion, pmoran. Tough to do, but always worth remembering.
Also, if ethics enter the discussion, your point about context markers sounds interesting. Maybe on another thread you can tell us more.
I have a glass of cranberry juice. I have a glass of milk. TWO THINGS for my patient, if kept separate. If integrated –as you recommend– my patient gets only ONE THING, and it does not taste very good.
You can’t mix pseudoscience with science in medicine. You just can’t. All you get for your effort are dull witted doctors skilled at equivocation, at best. Oh, and dead rhinos.
@Ed Whitney,
“Any closer?”
I have said what I meant, and I think my meaning is clear by now. I don’t see any point in your seeking some more ideal wording of your own to express what you think I meant.
I’ve asked about massage as a control for acupuncture before, but it turns out that it’s been compared to acupuncture and is better?
So why are we even talking about acupuncture/pseudo puncture at all?
@ Harriet:
I read your remarks about comparing high-placebo interventions with conventional treatment and agree with it.
This influences my interpretation of Haake and the other studies.
Haake 2007 had some regrettable omissions that would have shed light on these issues:
Patients were excluded if they had previous acupuncture for low back pain, but not if they had had it for other indications more than one year previously
The number of patients with previous acupuncture was not reported
The response rates of patients with and without previous acupuncture were not compared
Because some (including Dr. Ernst) think that its placebo effects may arise from a “novelty” effect (acupuncture-naïve patients more likely to respond), an opportunity to test this idea was lost
Expectations of acupuncture were reported in Table 2, and were high in all groups (7.7 on a scale from 0 to 10), but expectations of conventional treatment were not reported
The idea that conventional treatment came off worse because of much lower expectations is highly plausible, but not certain in the absence of data
Allison, I suspect that massage therapy is stigmatized in the US because sometimes massage treatments are euphemisms for prostitution and because people tend to like massage therapy even if they don’t “need it”.
I suspect that in the US, the puritanical idea that anything that people like but don’t “need” can’t be “good medicine”. Especially if that thing that people like can be conflated with prostitution.
Acupuncture has avoided being conflated with prostitution, and people don’t “like” acupuncture the way they like massage (and prostitution) and even though acupuncture doesn’t have the same positive effects that massage therapy does, there is a possibility that health insurance will pay for it.
There appears to be a rogue italics element in here. I shall close it:
Hopefully that worked, and further comments will not be italicized.
Drat — well, I tried.
Daedalus — I’m not sure massage is all that stigmatized. It’s more that it’s not primarily seen as medicine. It’s associated with spas and exotic vacations and rich people relaxing after a workout. It is also used as a cover for prostitution, yes, but most people think of massage as a luxury, as a way to treat oneself — like going out to a fine restaurant, or having your nails done, or buying a box of extra-special truffles and then scarfing the lot. It’s hard to equate it with real medicine when it’s primarily associated with luxury cosmetics, vacations, and, yes, romance and sex.
Sidenote: there was a bust in my hometown recently, where the masseuses were openly offering happy endings to the male clients. They apparently didn’t realize that a hand job for pay counts as prostitution here, and consequently were fairly open about what they were doing. The vice officers who arrested them said it was the easiest case they ever investigated — no need to be coy to get them to offer the extra service. Another service in a nearby city was a little more difficult, as the madam was more savvy to what is and is not legal. The investigation only started because of noise complaints from neighboring businesses; she tried to spin the moans as the results of “deep tissue massage” and was very good at figuring out who was a cop. Eventually they got evidence, though, and put an end to it.
@ Alison:
You are correct; massage was better than acupuncture in one well-done study referenced by me earlier.
The Cherkin 2001 article randomized patients to massage, acupuncture, and self-care. They measured symptoms and functional problems as their main outcomes. After follow-up at 4 week, 10 weeks, and 1 year, the massage group had lower symptom and disability scores than self-care, which in turn had lower scores than acupuncture.
In contrast to the other acupuncture studies we have been talking about, this did not use a sham acupuncture group. Consequently, there was much more variation in how acupuncture was done than in those randomized trials where it was done by a standardized protocol. That is, it included stuff like infrared heat, cupping, and electrostimulation of the needles. Doesn’t seem to have helped the patients much in this instance, since acupuncture came off inferior to the other two groups.
The massage group mostly got Swedish and deep tissue massage. Most massage therapists included “body awareness” techniques, which can make the clients aware of how they hold themselves, how they move, and how they can recognize early warning signals of injury. The massage practitioner had a choice of how to proceed, but the protocol specifically prohibited the use of so-called energy techniques like Reiki, and also prohibited stimulating meridian therapies like acupressure and Shiatsu.
The self-care got a book and videotape with instructions on controlling and preventing pain and improving quality of life.
Having a personal bias in favor of massage, especially combined with body awareness techniques, I had a favorable view of the study. When it comes to body awareness, for my money, Feldenkrais is the way to go.
The social acceptance of massage is another issue. It was once illegal in Modesto, California, to have a professional massage practitioner come to people’s homes; that law was only repealed in 1990. Massage is pretty mainstream stuff now, and has been for decades.
[...] and Chakras are related but different forms of vitalistic woo. As for acupuncture and Chi points, studies have shown that the location of needling in acupuncture is irrelevant. (They have also shown that [...]
[...] and Chakras are related but different forms of vitalistic woo. As for acupuncture and Chi points, studies have shown that the location of needling in acupuncture is irrelevant. (They have also shown that [...]
I didn’t realize that linking to this post would automatically generate a pingback/trackback that. Feel free to delete them as undesired clutter.
I was wondering what the heck those were. I have noticed them on a few posts over time here and was utterly confused. I am not so savvy to these things as I’d like to be.
[...] claims are bullshit. Acupuncture is pseudoscience with no medical value, and that has been well established in multiple studies. Also, this [...]
[...] akupunktura i akupresura beskorisni modaliteti alternativnog liječenja koji, jednostavno rečeno, ne djeluju, a u slučaju akupunkture znaju djelovati štetno (infekcija, sepsa…), pa i [...]
[...] Acupuncture Revisited [...]
[...] reports of 5 deaths and 90 other serious complications of acupuncture treatments. I wrote an accompanying commentary, “Acupuncture’s claims punctured: Not proven effective for pain, not [...]