Aug 20 2010

More Evidence that Acupuncture is a Placebo

I have written numerous blogs both here and on SBM about the acupuncture literature, which clearly shows that acupuncture, for any indication, is nothing but an elaborate placebo. Rigorous studies of acupuncture that actually try to isolate variables have shown that it does not matter where you stick the needles or even if you stick the needles – those variable do not have any specific effect. Acupuncture points and meridians are an illusion – nothing but superstition.

But there does appear to be a significant placebo effect, in addition to non-specific effects from relaxation and therapeutic attention, to the ritual of acupuncture. Does this mean “fake acupuncture works?” No – it means acupuncture does not work, but there are known placebo effects from the process of getting treated.

Now we have yet another study that supports the conclusion that acupuncture is just a placebo – but with an added element that is very interesting. Researchers compared traditional Chinese acupuncture (TCA) with sham acupuncture (non acupuncture points, shallow needle insertion) and another control group with no treatment for knee osteoarthritis. The researchers also did one very interesting thing, and one very sloppy and annoying thing (in my opinion). The sloppy thing was to use “electroacupunture” – which isn’t pure acupuncture. It’s acupuncture plus transcutaneous electrical nerve stimulation, which is an already proven modality for pain. In the TCA group they gave full “electroacupuncture” and in the sham group they gave less stimulation – enough to serve as an active placebo but not enough to have any effect.

This is problematic because it adds a variable that was unnecessary. In addition, I am disturbed by how stealthy this was – you can read the abstract and have only the slightest clue that this is a study of “electroacupuncture”. The authors clearly are equating “electroacupuncture” with acupuncture, which is not reasonable. It conflates two treatments, and makes for bad research.

In this particular study the TCA and sham groups had the same effect, and both were greater than the no intervention group. From this we can conclude that either both treatment groups had no physiological effect on pain, and the entire effect was due to placebo effects. Or we can conclude that both groups had placebo effects plus a similar effect from the electrical stimulation, despite the difference in intensity. There is no way to tell, which is why the addition of the electrical stimulation modality was a mistake.

However, the researchers also did something very interesting which I think is independent of their mistake with the electrical stimulation – they also altered another variable, the interaction between the acupuncturist and the subject. They split the treatment groups into “high” or “neutral” expectations. Essentially, either the acupuncturist said they had great success with this treatment, or they said that the treatment may or may not work. This is what they found:

455 patients who received treatment (TCA or sham) and 72 controls were included. No statistically significant differences were observed between TCA or sham acupuncture, but both groups had significant reductions in J-MAP and WOMAC pain compared to the waiting group (-1.1, -1.0, and -0.1, p<0.001; -13.7, -14, -1.7, p<0.001). Statistically significant differences were observed in J-MAP pain reduction and satisfaction, favoring the ‘high’ expectations group. Fifty-two percent and 43% in the TCA and sham groups thought they had received TCA (kappa=0.05), suggesting successful blinding.

So acupuncture does not work, but a positive attitude on the part of the acupuncturist does. This supports the conclusion that a significant portion of the effect measured in this trial is due to placebo effects.

This is not a perfect study, but it adds to the growing literature that is consistently showing that acupuncture does not work and is nothing more than an elaborate placebo. As we saw in a recent NEJM review, however, this kind of research is being used to recommend TCA – complete with acupuncture points and all the superstition that has been repeatedly disproven by research. At this point we can comfortably dispose of any notion of acupuncture points – they do not exist. They never had any scientific plausibility, and it is not surprising that research cannot demonstrate their existence (let alone clinical utility), and in fact strongly shows that they do not exist.

Further, clinical research also shows that it does not matter if you stick needles through the skin – simulating acupuncture by poking the skin with toothpicks or dull needles is sufficient. What does matter is the interaction between the acupuncturist and the patient. This is something that has been known for decades – that a positive therapeutic interaction will have non-specific placebo effects in the patient’s perception of their symptoms and well-being. That is precisely why we do placebo-controlled trials.

But proponents of acupuncture and other “alternative” modalities want to live in a bizarro world where placebo effects are used to justify a specific treatment. New York Times writer Tara Parker-Pope covered this article, writing:

The results don’t mean acupuncture doesn’t work, but they do suggest that the benefits of both real and fake acupuncture may have something to do with the way the body transmits or processes pain signals. Other studies have suggested that the prick of a needle around the area of injury or pain could create a “super-placebo” effect that alters the way the brain perceives and responds to pain.

Actually, these results do mean that acupuncture does not work. When you compare a treatment to a placebo treatment and there is no difference, that means the treatment does not work – it is not adding anything to the placebo.  The math is simple: treatment = treatment effect + placebo. Sham treatment = placebo.  Then you simply subtract the latter from the former: treatment effect + placebo – placebo =  treatment effect, which in this case = zero, zip, nada – no treatment effect, which means that acupuncture does not work.

CAM proponents have pulled a Jedi mind trick on the media and large portions of the public, however, and convinced them that no treatment effect = an effective treatment. This nonsense is infiltrating academia and medicine because proponents were never very good at thinking scientifically (apparently, since they think magical superstition works) and most everyone else is a shruggie who is content to look the other way because they don’t see a threat. The public then thinks it’s all legitimate because they read about it in the NY Times referencing articles in the New England Journal of Medicine.

At least I can rant about it whenever it happens.

Share

41 responses so far

41 Responses to “More Evidence that Acupuncture is a Placebo”

  1. johncon 20 Aug 2010 at 9:08 am

    Steven,

    Your argument seems to be that that if the placebo shows the same effects as the real treatment, then the real treatment is a sham.

    Your logic is quite faulty. If it were true, then we could apply it to cancer treatments, flu medication, anti-depressants and many other medical practices that have shown equivalent efficacy when a placebo has been used.

    Certainly my own experiences with acupuncture fall well outside the remit of your opinion on this one. Though if it is just a placebo, then the power of my mind over my body is far more effective than any conventional treatments, which raises doubtful questions about conventional therapy, and positive questions about alternative therapy.

  2. bluedevilRAon 20 Aug 2010 at 9:27 am

    Thanks for taking this apart. I wonder to what degree patient expectation played into the Tai Chi Fibromyalgia study that just came out in the NEJM. At least news outlets showed a bit more journalistic integrity by saying “Tai Chi may help fibromyalgia.”

  3. SARAon 20 Aug 2010 at 9:35 am

    I continue to be startled by placebo. I think we should be spending a great deal more research on studying what is happening to our brains and bodies when placebo effect kicks in. Because there is clearly some benefits we can garner from placebo, if we can just get a clearer understanding of the physiological process. The body has the ability to somethings manage things for itself, we just need to figure out how and why it does it, so we can start mining that knowledge for more effective treatments.

    We downgrade placebo, but clearly its important and my guess is that it would help us to develop better drugs if we understood it.

  4. bluedevilRAon 20 Aug 2010 at 9:39 am

    johnc, anecdotes are welcomed by readers so long as they are educational or humorous. As data, they mean nothing. So your experience with alternative therapies is erroneous and you should reflect on that. Do not fall into the trap of “it works for me.” That is illogical.

    As for your assertion about placebo, you’re flat out wrong. The definition of an effective treatment is one that performs better than placebo. If interventions perform the same as placebo, then they are considered ineffective and should probably not be administered (barring any confounding factors). Why give a treatment that could potentially do harm when there is no benefit?

    The data for antidepressants are quite clear that they work for severe depression. It is less clear if antidepressants work beyong placebo for moderate and mild depression. This is a debate that psychiatrists are having right now. Is it better to treat mild depression with talk therapy or with medication that has the potential for side effects? I believe that Dr. Novella is merely calling for this same critical debate to be applied to alternative medicine like acupuncture. There is too much blind acceptance of alt med by people like you or shruggie attitude with complete disregard for the science.

  5. bluedevilRAon 20 Aug 2010 at 9:48 am

    SARA, I agree it is a powerful and amazing thing. However, I do not see the potential for it to be used very effectively. First thing to consider is the ethical issue. Should a doctor knowingly lie to a patient that he/she is receiving a real treatment, when in fact it is placebo? Keep in mind that in randomized-controlled trials, the doctor should be ignorant of treatment group and the risk of placebo vs. treatment is explained to the patient. The subject in the trial knows that he or she has a random chance of being on placebo. This is part of the consent process when the subject is recruited into the study.

    Secondly, I would expect that the placebo effect is more prominent in a percentage of the population. The more optimistic? The more gullible? I don’t really know. So how do we tailor it to individuals?

    We should probably research placebos more with advances in brain imaging, but I would hesistate to throw too much money into it, mainly because of the ethical issues.

  6. johncon 20 Aug 2010 at 10:00 am

    bluedevilRA,

    Practical medicine only uses existing data as a statistical starting point, after that everything is based on patient response, ie anecdotal evidence, which in the context of any individual case trumps any research or data to the contrary.

    In this, single, case, acupuncture worked and conventional treatment did not.

  7. Watcheron 20 Aug 2010 at 10:26 am

    John,

    If there’s no difference between the treatment and sham groups, it means that the same thing is occurring in both groups, ie. any altered variables show no difference towards affecting the outcome. So if the only difference in vairable between treatment and sham is the actual treatment, yet the outcome remains the same, how can you say that the treatment is effective? I don’t see the logic in your argument …

  8. bluedevilRAon 20 Aug 2010 at 10:29 am

    johnc, you’re right to a point, but then you’re also just wrong. For example, in the treatment of hypertension there are many medication options. Initial treatment is usually a diuretic (based on statistical data showing that they are effective first line agents). If the blood pressure is not effectively lowered, an ACE inhibitor, additional diuretic, beta blocker, etc can added. Medications may be taken away if the patient is experiencing unpleasant side effects. So yes, medication can be tailored to an individual patient based on what works best.

    However, it is not anecdotal that some patients have resistant hypertension or nasty side effects. This is data that has been well documented. That is why a hierarchy of drugs exist (first line, second line, etc). The research evaluates efficacy vs. potential for side effects. Thus, a proper evidence-based physician uses the best available data.

    That is not to say that a great many doctors do not always adhere to this paradigm. It is regretable, but we are all susceptible to confirmation bias. Doctors observe trends among their patients and sometimes extrapolate from that. Observations can lead to proper research and that is a good thing. Observations without statistical backing is meaningless though.

  9. Watcheron 20 Aug 2010 at 10:31 am

    If it were true, then we could apply it to cancer treatments, flu medication, anti-depressants and many other medical practices that have shown equivalent efficacy when a placebo has been used.

    Emphasis mine

    Also, just as an aside, I know that flu medication and AD’s are a bit murky when it comes to placebo vs. treatment, but cancer? Could you point me to a study or two that shows sham treatment efficacy on par with that of chemo and/or excision?

  10. DrMonkeyon 20 Aug 2010 at 10:55 am

    SARA it sounds like you’re suggesting that placebo is people actually getting better. If I’m not mistaken, the placebo effect is primarily just people’s *perception* of getting better (which means the more leeway there is in perceiving symptoms, the more of an effect there can be). Of course there could be some effects related hormonally to reduction in depression and stress that run in parallel, however I haven’t read of any study correlating the placebo effect to actual improvements in objective disease indicators for non-psychological conditions (I am interested to hear if anyone has read of such effects).
    So basically skewed perception is the major factor here, not any actual healing.

  11. eeanon 20 Aug 2010 at 11:04 am

    @Watcher
    My thoughts exactly, if there was a sham cancer treatment it would hopefully get booted out. There’s been plenty of treatments that have been stopped because further research has shown it to be no better then placebo.

    Its obvious that acupuncture is just a form of massage therapy. People benefit from the attention. So if I could wave a magic wand, I would just group them together with massage therapists and require the same level of certification for both. But clearly requiring extensive training is just silly and having special acupuncture licenses is a sham.

  12. Allyon 20 Aug 2010 at 11:04 am

    Specific research on popular treatments, such as acupuncture, is something I strongly support. A treatment that is a waste of time and money should be exposed (especially if it very expensive).

    Conversely, an effective alternative to invasive surgery and the use of pain medication should be promoted, as surgery and strong pain medication can cause issues. Mainly due to the onus of responsibility on the patient to take medication regularly and responsibly, and/or to take proper care of the surgery sight post-op to prevent infection.

    Bowen Therapy is another interesting popular treatment. I know someone who was charged $80 per session for treatment of his back. He is usually very frugal with money, but because of his desperation to end the pain went ahead with the treatment without much prior investigation.

    It’s good to have a look at the evidence regarding “alternative therapies” as soon as possible to avoid basing your decision on desperation to relieve your condition should something go wrong medically down the track.

    In an ideal world reliable information would be the norm and biased fluff would be the exception. Ah well. No use crying over spilled milk.

  13. stompsfrogson 20 Aug 2010 at 11:41 am

    @ Ally:
    Is $80 a bad price for a massage? I don’t think it is.

    I went to PubMed for some studies on Bowen Therapy and I didn’t find any. The Wikipedia entry on it only has three, two poorly done studies and one well done study. The Wiki entry on it was clearly written by someone who was against the treatment, but the only well-done study they cited was positive, although you had to click the source link to find that out!

    http://en.wikipedia.org/wiki/Bowen_Technique#Research

    This is like, so outside of the normal formula for alt med crap peddlers. I’m baffled. Initially it looks bad, but a little more googling and it just looks unstudied. And, you know, massages are nice. And Bowen doesn’t recommend any other therapies in conjunction with Bowen therapy, i.e. he doesn’t like SBM but he’s not a homecrapathist either. Probably just the humble beginnings of another woo-pusher, but still…

  14. Steven Novellaon 20 Aug 2010 at 12:05 pm

    You should not confuse individualization of therapy with using anecdotal evidence to support a therapy. It is OK to individualize treatments that are already proven to work, but it is not OK (i.e scientific) to use only anecdotes to justify a treatment, especially one that is implausible or does not work under controlled conditions.

    Further – the relationship between patient feedback and treatment depends upon context. For symptomatic treatment, patient feedback is critical. For prevention of disease, it is meaningless except for tolerance. For example, we treat patients with aspirin to prevent heart attacks and strokes based entirely on objective evidence. There is no basis for individualization based upon feedback (as long as the patients tolerate the aspirin).

    johnc, like most people, appears to grossly underestimate the potential for self-deception.

  15. dodomanldson 20 Aug 2010 at 1:56 pm

    Mr. Novella

    I encourage you to read the latest scientific studies that actually show that acupuncture’s effects can be mapped in a functional MRI in real time, showing activation of different brain areas according to different acupuncture points, thus leading to a map of neuronal pathways that are stimulated with acupuncture, this pathways in turn regulate different systems like sympathetic neurotransmitters, immune and pain regulatory citokynes, hormones and endocrine systems and so on and so forth. Here are some articles you might find interesting, for a change:

    Brain encoding of acupuncture sensation — Coupling on-line rating with fMRI
    Vitaly Napadow, et al. NeuroImage 47 (2009) 1055–1065

    Time-variant fMRI activity in the brainstem and higher structures in response to acupuncture
    Vitaly Napadow, et. al, NeuroImage 47 (2009) 289–301

    The Status and Future of Acupuncture Mechanism Research
    Vitaly Napadow. THE JOURNAL OF ALTERNATIVE AND COMPLEMENTARY MEDICINE
    Volume 14, Number 7, 2008, pp. 861–869

    Acupuncture mobilizes the brain’s default mode and its
    anti-correlated network in healthy subjects
    Kathleen K.S. Hui.Brain Res. (2009), doi:10.1016/j.brainres.2009.06.061

    Spatiotemporal Mapping the Neural Correlates of
    Acupuncture with MEG
    Rupali P. Dhond, Ph.D., et al.THE JOURNAL OF ALTERNATIVE AND COMPLEMENTARY MEDICINE
    Volume 14, Number 6, 2008, pp. 679–688

    Dhond RP et al., Do the neural correlates of acupuncture and placebo effects differ?, Pain (2007), doi:10.1016/j.pain.2007.01.001

    Once you’ve read some of the above mentioned papers you might be on a better position to talk about acupuncture and placebo, and maybe, answer some of the following questions:

    Why is it so difficult for you to understand that energy is part of our bodies just as the liver cells are? Don’t we have used an EKG to measure the heart cell’s electric behavior for decades? Is it too far fetched then to understand that manipulating electric impulses in our body can have a regulatory effect in a distant part of it? Can the insertion of metal in the myofascial network influence it’s electrical and ionic polarity and can this insertion be a sufficient stimulus to conduct an electric impulse towards the CNS?
    What do you know about the myofascial tissues as a big network where interstitial fluids and ions are in fact connecting distant areas of our bodies? What do you know about the propioceptor neurons that are located in the joints along myofascial tissues and that respond to the insertion of an acupuncture needle with an electric impulse that, now we know, activates or deactivates different brain areas, affecting sympathetic, endocrine and citokine responses to environmental stimulus?

    Acupuncture is a complex intervention, and the role of placebo in complex therapeutic interventions is far more prominent that in, lets say, pharmacotherapy, but it still is present in drug therapy. In fact, it is present in any therapeutic act, as long as the patient has some expectations about getting better. This doesn’t deny the value of new acupuncture research that shows precise anatomical and neurological pathways to explain acupuncture effects. All in all, we celebrate the debate and hope for more research that shows how things work where our narrow minds don’t see anything.

  16. CJon 20 Aug 2010 at 4:29 pm

    I think the final point mentioned there about self deception is a pretty key one.
    I have a lot of friends who are proponents/users of alternative medicines, and some how have beliefs in ESP, Psychic powers etc etc.
    One thing that links them all is that they all think that their memory and recall of events is much better than it actually is.

  17. sonicon 20 Aug 2010 at 4:44 pm

    The study doesn’t study what it says it studies, the therapists are unblinded, the treatments results are not measured in any objective manner–
    If the conclusions were different, what would the reliability of the study be?

  18. CivilUnreston 20 Aug 2010 at 5:03 pm

    I know that there is lots of testimonial (ie: anecdotal/unusable) evidence for acupuncture (and various other forms of woo) alleviating some really debilitating symptoms. The “case study” from the recent NEJM article comes to mind.

    Anyhow, I’m curious as to what actually happens when someone comes in with self-reported debilitating pain and then, after acupuncture, reports an ability to resume normal functioning. Was their pain psychogenic from the beginning? Were they intentionally or unintentionally exaggerating their pain? I guess it’s just as likely that the “effect” is temporary and fades rapidly until the next session.

    I suppose I’m asking the same question as SARA — do non-specific effects induce real physiological changes? If so, perhaps we could figure out how to harness those physiological changes without resorting to woo.

  19. colluvialon 20 Aug 2010 at 6:04 pm

    @SARA: “I think we should be spending a great deal more research on studying what is happening to our brains and bodies when placebo effect kicks in. Because there is clearly some benefits we can garner from placebo, if we can just get a clearer understanding of the physiological process.”

    @bluedevilRA: “I do not see the potential for it to be used very effectively. First thing to consider is the ethical issue. Should a doctor knowingly lie to a patient that he/she is receiving a real treatment, when in fact it is placebo?”

    I think the first step in sizing up the placebo effect is to catalog all the situations in which it has had some effect. Once we have a firm grasp on where it effectiveness lies, from “it improved my sense of well-being” to “it sped my recovery” to “it sent my cancer into remission”, we’d know how important the effect is and how much time we want to invest in figuring out how it works.

    Regarding the ethics of administering placebos, lying to a patient may not be necessary. If all a placebo does is make someone feel better, we’re all taking placebos all the time, in the form of pleasurable activities.

  20. elmer mccurdyon 20 Aug 2010 at 7:53 pm

    I’m curious what the evidence is TENS is. Not having access to expensive medical databases, I did a quick Google Scholar search came up with some (very selective) systematic reviews showing no effect beyond placebo. Since I own a TENS unit, and it works for me, I just shrugged it off (in fact I had some thoughts about why those reviews might be flawed, but never mind); however, I am curious as to what Dr. Novella’s evidence is in this case.

  21. elmer mccurdyon 20 Aug 2010 at 7:54 pm

    evidence *for* TENS

  22. Steven Novellaon 21 Aug 2010 at 3:59 pm

    dodomanlds – I am very aware of that research. It is not impressive. All it shows is that there is a non-specific physiologic response to poking a needle into tissue. There are also many small and poorly controlled studies mixed in there, with very sloppy use of fMRI (which is very tricky to do well).

    So far – I am not impressed at all. This research does not establish that there is a specific physiological response to acupuncture, or anything special about acupuncture points. The scientific research does not establish the existence of acupuncture points.

    And you can’t just vaguely invoke “energy” and unkown myofacial networks. There is no known anatomy that corresponds to acupuncture points – they are imaginary.

    Also – if you know anything about the history of acupuncture it is implausible in the extreme that it’s development represents knowledge about any actual anatomy or physiology.

  23. Steven Novellaon 21 Aug 2010 at 4:06 pm

    elmer – TENS does not work better than placebo for chronic lower back pain, which is not surprising given that this is a multifactorial pain syndrome.

    However, reviews of TENS for other indications – neck pain, neuropathic pain – are positive but not definitive. We need better studies, but existing evidence suggests an effect.

    Reviews for osteoarthritis are inconclusive.

  24. Rob4on 21 Aug 2010 at 7:57 pm

    Hi Steven,

    In the UK Professor Edzard Ernst is a very public criticizer of CAM and branded “ the scourge of alternative medicine”. His conclusion in 2007 of the evidence base for acupuncture was that:

    “The effectiveness of acupuncture remains a controversial issue. … The results indicate that the evidence base has increased for 13 of the 26 conditions included in this comparison. For 7 indications it has become more positive (i.e. favoring acupuncture) and for 6 it had changed in the opposite direction. It is concluded, that acupuncture research is active. The emerging clinical evidence seems to imply that acupuncture is effective for some but not all conditions.”

    (http://www.ncbi.nlm.nih.gov/pubmed/17265547)

    What are your thoughts on his conclusions?

  25. ceekon 21 Aug 2010 at 10:24 pm

    What about NEJM study comparing Tai Chi vs wellness-education-stretching for fibromyalgia…. Is this NCCAM funded woo?

  26. elmer mccurdyon 21 Aug 2010 at 11:13 pm

    Well, that’s pretty consistent with my experience. I manage to tweak my back every 5-10 years and found the TENS to be of no help for that, though I find it very helpful for some of my other, more chronic pains.

  27. werdnaon 21 Aug 2010 at 11:15 pm

    “Why is it so difficult for you to understand that energy is part of our bodies just as the liver cells are?”

    For me it’s because you don’t seem to know what the heck you’re talking about. The only formal definition of “energy” that I know of is the “potential to do work”. This is an abstract concept – it might be reified in electrons, mixtures of chemicals in a battery or even the position of a heavy rock on a high hill – but energy itself is an abstract concept. My body is, as far as I can tell a physical thing – so no I can’t see how an abstract concept is “part” of my body in any reasonable usage of the term.

    Want an interesting experiment try replacing “energy” with the exact physical mechanism behind it. I expect you’ll find yourself in one of two places:

    i) You don’t know – which should tell you that you’re kind of blowing smoke.
    ii) The system you mention is easily falsifiable.

    (More frequently I find “energy” people choose “option 3″ replace “energy” with an even more opaque term. I recall this great conversation I had with people who wanted to tell me about different physical realms co-existing in the same points in space-time. I asked them why can’t we see them and they responded that the molecules were vibrating at different frequencies. I told them that we already had a word for that: “Heat”)

  28. Paul N.on 22 Aug 2010 at 12:46 am

    Just a remark from the other side of the pond.

    If you have a centralized Health Care System, which is as I suppose you are going to have soon, you will see that this system pays for all types of quackery but cuts expenses for real treatment (http://healthcarebubble.wordpress.com/2010/04/09/acupuncture-fully-covered-by-german-health-insurances/)

    There are some reasons why this develops inevitable:
    1.All Health Care systems face limited resources
    2.A centralized Health Care System makes a democratic decision how to share. Most people are pro-acupuncture as such type of medicine gets the best media coverage.
    3.By central order then, physicians get well paid for it but hardly for real medicine as administrators know if physicians were allowed to allocate the res sources, they would drop all kind quackery first.

    The primary goal of a centralized Health Care system is not to care for patients but to preserve itself.

  29. elmer mccurdyon 22 Aug 2010 at 2:54 am

    Paul, I suppose it’s best to ignore that, but I’m bored.

    1. You’re in the UK, no? Odd that instead of citing your own country, you went for Germany, as if that country were typical in its level of support for acupuncture.

    2. I know that here in California, the major private insurers cover acupuncture.

    3. The idea that govt. is overly responsive to popular demand is, well, perhaps not the craziest libertarian argument I’ve ever heard, but it’s certainly unusual.

    But we digress…

  30. Paisleyon 22 Aug 2010 at 3:24 am

    Steven Novella: “This is something that has been known for decades – that a positive therapeutic interaction will have non-specific placebo effects in the patient’s perception of their symptoms and well-being. That is precisely why we do placebo-controlled trials.

    It seems to me that you (as a practicing physician) are failing to recognize something very significant here – namely, the power of belief and its effect on health. The placebo-effect (which is based on the positive belief of the patient) provides proof-positive that faith heals. On the other hand, the nocebo-effect (which is based on the negative belief or doubt of the patient) provides proof-positive that skepticism is actually detrimental to one’s health.

  31. Al Morrisonon 22 Aug 2010 at 8:11 am

    @Paisley
    @Sara
    @BluedevilRA
    @Colluvial

    While the placebo effect is interesting, it is part of a larger issue in research: Confounding variables more specifically referred to as demand features or demand characteristics. If you are, in your heart, interested in understanding the placebo effect, you owe it to yourselves to look into the broader topic of experimental design and the experimental method.

    The placebo effect is not objective, but comes from subjective self-reports, usually involving pain and perception.

    If mom/dad kisses you knee, or bandages it, or twists a toothpick against a “meridian point” and tells you “all will be better,” your knee will feel better. Of course, we all have experienced this. It is interesting, and exemplifies how social an animal we are. This “demand characteristic” or placebo effect can take on many forms and need only work a very short time until the point where the pain would have abated anyway.

    The real test for the placebo effect, faith healing, mom or dad interventions (on so on) is long term and objective. Objective measurements, such as blood pressure or tumor size, do not demonstrate the same efficacy as treatment. Meta-analysis of more than one hundred placebo-controlled studies shows no statistically significant effect for the placebo across all disease types. This suggests the placebo effect is not universal, and that the placebo effect is an artifact of the research paradigm.

    For example, clinical trials in which participants are aware they may receive either the treatment or a placebo self-report reduced efficacy regardless of whether they received the actual treatment or placebo.

    Another example. Research where self-report questions are asked in person by an authority figure or given as a questionnaire demonstrates an expectancy (placebo) effect for those in the latter group. That is, depending upon how an experimenter measures for a placebo effect, there may or may not be an effect.

    The placebo is an important tool. Research shows the effects of placebo are not as strong as treatment (see Steve Novella’s equation, TREATMENT + PLACEBO – PLACEBO = TREATMENT EFFECT). Therefore, a hypothesized treatment with an effect equal to or below that of the placebo is (if the experimental paradigm is sound) is rejected. For this reason, understanding the placebo and its role in experimentation is very important to understanding sham vs. actual treatments. Unfortunately, sometimes you must have access to, and understand, the research to determine what if any effect (placebo or otherwise) is present. Often, the effect is an artifact of the research design and nothing more.

  32. elmer mccurdyon 22 Aug 2010 at 10:45 am

    If you’re interested in research on placebo effect itself, the disgruntled phd blog is slowly making its way through a very interesting series on the subject.

  33. tmac57on 22 Aug 2010 at 11:43 am

    Can an experimenter’s bias be properly considered as part of the placebo effect? I have heard it argued both ways. Do we have a firm definition of what constitutes the placebo effect?

  34. Paisleyon 22 Aug 2010 at 12:28 pm

    Al Morrison: “The placebo effect is not objective, but comes from subjective self-reports, usually involving pain and perception.

    Reducing or eliminating pain is not a trivial thing, especially for the individual who is presently experiencing it.

    Al Morrison: “Meta-analysis of more than one hundred placebo-controlled studies shows no statistically significant effect for the placebo across all disease types. This suggests the placebo effect is not universal, and that the placebo effect is an artifact of the research paradigm.

    This is not entirely true. You are referring to the Hróbjartsson and Peter Gøtzsche study . This research has been criticized for the following reason.

    Placebos also do not work as strongly in clinical trials because the subjects do not know whether they might be getting a real treatment or a sham one. Where studies are made of placebos in which people think they are receiving actual treatment (rather than merely its possibility) the placebo effect has been observed.[95] Other writers have argued that the placebo effect can be reliably demonstrated under appropriate conditions.[96]

    (source: Wikipedia: Placebo)

    Al Morrison: “The placebo is an important tool. Research shows the effects of placebo are not as strong as treatment (see Steve Novella’s equation, TREATMENT + PLACEBO – PLACEBO = TREATMENT EFFECT). Therefore, a hypothesized treatment with an effect equal to or below that of the placebo is (if the experimental paradigm is sound) is rejected. For this reason, understanding the placebo and its role in experimentation is very important to understanding sham vs. actual treatments.

    How many medical doctors prescribe antidepressants to their patients? There is evidence that placebos perform better than antidepressants.

    A 2002 article in The Washington Post titled “Against Depression, a Sugar Pill Is Hard to Beat” summarized research as follows: “In the majority of trials conducted by drug companies in recent decades, sugar pills have done as well as — or better than — antidepressants. Companies have had to conduct numerous trials to get two that show a positive result, which is the Food and Drug Administration’s minimum for approval. The makers of Prozac had to run five trials to obtain two that were positive, and the makers of Paxil and Zoloft had to run even more”.[34]

    (source: Wikipedia: Placebo)

    (This is not a trivial issue. The antidepressant drug industry is a huge business with political clout.)

  35. Al Morrisonon 22 Aug 2010 at 2:06 pm

    @Paisley,

    You are correct, pain is not trivial. When I write “The placebo effect is not objective, but comes from subjective self-reports, usually involving pain and perception,” I mean exactly what I have written: it is subjective (not able to be objectively measured) and requires the subject to report the intensity and any change in that intensity. For research paradigms with these measures there are typically placebo effects. When all placebo research is considered, the effect is lost.

    This does not suggest (1) that pain is trivial or (2) that the placebo effect does not occur, or is not real. What it does suggest is the placebo effect is a function of the research paradigm. Sometimes we can take advantage of the effect, by incorporating placebo-control groups (especially where subjective measures are used). Sometimes we want to try to eliminate the placebo effect by using other measures.

    One Washington Post article does not science make. Depression is a very poorly defined disorder. The real numbers shake out (not perfectly, certainly) like this: Major depression responds to both a statistically and clinically significant level to contemporary antidepressants. However, the definition of depression is poor. Where you seem to feel this is evidence for offering placebo treatment (correct me if I am mistaken), I see an issue with defining and diagnosing that has always plagued psychology and psychiatry. Again, we need to resolve the terms of the discussion (the operational definition of the measurement) before our research can progress.

    However, let us conduct a little thought experiment. We will assume antidepressants do nothing above placebo, ever. This does not mean the placebo has done anything. It means antidepressants have done nothing. That is the power of the placebo effect; the placebo can confirm or support the null hypothesis. That is it. No special healing powers. Just the power to negate. It is, though, a very important power; not one that should be trivialized.

    Everyone should keep in mind, when it comes to research, we are looking at a single, very specific social interaction. The placebo in research is not different than other possible research variables. Furthermore, such interactions are not limited to research; we see similar effects in society writ large. The power of suggestion, authority, belief, and perception (and so forth) are important in- and out-side the clinical or laboratory setting. That is why skepticism and critical thinking are important not just for the professional scientist, but all of us.

  36. Eric Thomsonon 22 Aug 2010 at 10:39 pm

    Another interesting piece of the puzzle. This study’s use of clinician confidence was a great idea, and I’m surprised it hasn’t been done before. It has the same problems as other studies in that the “sham” case may be producing adenosine release just in lower amounts to the “real” thing, so it could be like saying tylenol doesn’t work for pain b/c aspiring works better.

    Second, it could be that placebo and adenosine release actually interact, so placebo + acupuncture (adenosine release) together relieves more pain that placebo alone and adenosine release considered individually. That is, nonlinear interaction between placebo and treatment is not eliminated.

    So while the paper isn’t conclusive, depsite Steven’s funeral call, it is a step in the right direction, and I think with more science we’ll be able to disentangle all these factors in an evidence-driven way.

  37. Calli Arcaleon 23 Aug 2010 at 12:26 pm

    Paisley:

    It seems to me that you (as a practicing physician) are failing to recognize something very significant here – namely, the power of belief and its effect on health. The placebo-effect (which is based on the positive belief of the patient) provides proof-positive that faith heals. On the other hand, the nocebo-effect (which is based on the negative belief or doubt of the patient) provides proof-positive that skepticism is actually detrimental to one’s health.

    First off, the “nocebo” effect is hardly limited to skeptics. Like the placebo effect, it is universal, and likely accounts for a significant (but difficult to quantify) proportion of the adverse event reports for various interventions. Especially the vague ones, like “fatigue”, which have been reported even for sugar pills as the control arm of a pharmaceutical trial. It’s a real phenomenon, and really it’s the same phenomenon as the placebo effect — expectations affect perception.

    And that’s the key here. The placebo effect isn’t about actual improvement. It’s about how expectations affect perception, and yes, as one person pointed out upthread, it does affect the experimenter as well. (This, incidentally, is why animals seem to experience the placebo effect. Broadly speaking, they’re not actually getting better — their owners just think they are.)

    There isn’t a “faith/skepticism” dichotomy, exactly, and even if there was, it wouldn’t map to “faith heals, skepticism hurts”. The placebo effect doesn’t care about your religious or spiritual perspective on life. It’s purely about your expectations, and these can come from anywhere. Skepticism can help a bit, but skeptics are fooling themselves if they think they are immune to it. Which, in fact, is precisely why science exists — we are all vulnerable to the placebo effect and other forms of observer bias, so we need to be very methodical to rule out the possibility that we’ve just deceived ourselves.

  38. daijiyobuon 23 Aug 2010 at 6:44 pm

    Dr. N., according to ‘Not a Doctor Maloney Naturopath’ [of the naturopathick], you are closed minded and the acu. studies out of China rock!

    ( see http://alternativendhealth.wordpress.com/2010/08/22/acupuncture-novella-acts-as-judge-and-jury/ ).

    I call this the ‘reverse sectarian accusation,’ whereby, the ‘dogma-bound sectarian pseudomedicalist’ accuses the ‘person following the evidence provided by medical science’ of being predetermined / dogmatic in mindset.

    I’m trying to figure out the line of thinking [I'm being generous here] that ‘NAD Maloney ND’ provides:

    “the only way to continue to ignore the positive acupuncture outcomes is to completely disregard everything coming out of China. Since we now outsource much of our medical production to China, it seems like we should accept them as legitimate.”

    I read it this way: since China manufactures medical product materials, therefore studies coming out of China must, a priori, be of high quality and just-as-actual-as-those-materials.

    As if.

    -r.c.

  39. elmer mccurdyon 23 Aug 2010 at 9:54 pm

    tmac57 – “Do we have a firm definition of what constitutes the placebo effect?”
    See here: http://disgruntledphd.blogspot.com/2010/07/placebos-all-you-never-wanted-to-know.html
    As I said before, I’m a little bit hesitant to link, since he doesn’t seem to be in the habit of doing so (plus he’s already gotten a visit from the sort of individual who sneeringly compares the placebo affect to the invasion of Iraq, of all things), but it does look like it’s turning into a useful resource regarding the question of the extent to which the placebo effect is real vs. just being a matter of reporting, etc.

  40. Rob4on 24 Aug 2010 at 7:06 am

    The placebo effect as a neurobiological phenomenon has been most researched by Fabrizio Benedetti. His book ‘Placebo effects’ is brilliant. He notes that you can now legally dope athletes through the conditioning response of giving morphine repeatedly with a sugar pill – take the morphine out and when given, the body will mimic the effects of the morphine simply in response to the sugar pill. There are a variety of different placebo effects that being one of a conditioned response. But the placebo response in an expectation context is a real response by the body with a variety of physiological correlates

  41. Enzoon 31 Aug 2010 at 2:05 pm

    It is interesting to speak with people of the mentality that placebo effect constitutes a legitimate treatment. I suspect using people’s faith in a treatment against a disease/condition has no practical value in a clinic once the novelty or mysticism behind a treatment wears out. And I’m willing to bet that kind of deception by a physician is dangerous…

    “My doctor put me on this new [sugar] pill for my hypertension. You should stop taking your prescription and get on this. It works!”

Trackback URI | Comments RSS

Leave a Reply

You must be logged in to post a comment.