Acupuncture and back pain: some interesting background references

September 25th, 2007 by Ben Goldacre in acupuncture, bad science, placebo | 97 Comments »

I was just on Radio 4’s PM program talking about the acupuncture study that’s in the news today, you can listen to it here (37 minutes in to the programme):

Here are some references and background bits and bobs.

The paper itself was very interesting. It took 1200 people, with an average of 8 years back pain each: we can assume not been helped by biomedical treatments. They were split into three groups: one group had medical treatment; one group had proper, real, bells and whistles, needles in the “meridiens” acupuncture; and one group were treated with pretend acupuncture.

The results were fascinating: they set a threshold for “improved” (which was either a 33% improvement in 3 aspects of one score, or 12% improvement on another measuring scale). They found that people having acupuncture were almost twice as likely to reach this level of improvement in back pain as people on medical treatment (which had already let them down for 8 years of course). But even more interestingly, the pretend acupuncture group, where they just bunged needles in any old place with a bit of ceremony, did just as well as the people having proper, posh, theatrical, genuine acupuncture.

The press release is here, and the full academic paper is here:

I’ll try tomorrow to encourage them to make it free as always.

[edit 26/9/07: the American Medical Association have basically said no I think.]

I’ve written about the placebo before, how it about much more than a pill, but is about the cultural meaning of a treatment, our expectations, and more. It’s worth seeing these fascinating acupuncture results in that context:

Here’s what I wrote just last week on the subject:

There’s nothing inherently wrong with the idea of giving out sugar pills. The placebo effect can be very powerful, because it’s not just about the pill, it’s about the cultural meaning of the treatment: so we know from research that four placebo sugar pills a day are more effective than two for eradicating gastric ulcers (and that’s not subjective, you measure ulcers by putting a camera into your stomach); we know that salt water injections are a more effective treatment for pain than sugar pills, not because salt water injections are medically active, but because injections are a more dramatic intervention; we know that green sugar pills are a more effective anxiety treatment than red ones, not because of any biomechanical effect of the dyes, but because of the cultural meanings of the colours green and red. We even know that packaging can be beneficial.

It’s also really worth checking out this fascinating paper, comparing two different placeboes: a placebo pill, up against an elaborate placebo ritual involving a medical device (it was modelled on acupuncture). They found that the more elaborate placebo ritual was more effective than the tablet.

More even than that, I think back pain is incredibly interesting: in 90% no cause is found, and we know that things like psychosocial stressors, work problems, bed rest and depressive symptoms are significant risk factors for moving from a twinge to chronic enduring pain. There is no doubt that eg brief educational interventions (don’t avoid exercise, do avoid rest, do avoid specific exercises) are helpful.

Back pain is clearly a problem which requires more than simply pharmaceutical pills. The question is whether an elaborate, expensive, gimmicky and theatrical placebo ritual is an effective use of money, or whether other, cheaper, more pragmatic, honest psychosocial interventions might be more appropriate and cost effective.

Here is an excellent review of causes and treatments for back pain, setting out the clear evidence for keeping active, avoiding rest, and the role of psychosocial factors in the move from acute to chronic. I’ve pulled out some key quotes for people with lazy clicking fingers.

What are the most important prognostic indicators for chronicity?

 Early identification of patients with low back pain at risk for long term disability and sick leave is theoretically and practically important because early and specific interventions may be developed and used in this subgroup of patients. This is of special importance because recovery for people who develop chronic low back pain and disability is increasingly less likely the longer the problems persist.

The transition from acute to chronic low back pain seems complicated, and many individual, psychosocial, and workplace associated factors may play a part. In this respect, increasing evidence indicates the importance of psychosocial factors.w7 A recently published systematic review of prospective cohort studies found that distress, depressive mood, and somatisation are associated with an increased risk of chronic low back pain.13

Table 1 shows a list of individual, psychosocial, and occupational factors, which have been identified as risk factors either for the occurrence of low back pain or for the development of chronicity. “Yellow flags” have been developed for the identification of patients at risk of chronic pain and disability. A screening instrument based on these yellow flags has been validated for use in clinical practice.14 The predictive value of the yellow flags and the screening instrument need to be further evaluated in clinical practice and research.

How effective are treatments in acute low back pain?

 The evidence that non-steroidal anti-inflammatory drugs relieve pain better than placebo is strong. Advice to stay active speeds up recovery and reduces chronic disability. Muscle relaxants relieve pain more than placebo, strong evidence also shows, but side effects such as drowsiness may occur. Conversely, strong evidence shows that bed rest and specific back exercises (strengthening, flexibility, stretching, flexion, and extension exercises) are not effective. These interventions mentioned were equally as effective as a variety of placebo, sham, or as no treatment at all. Moderate evidence shows that spinal manipulation, behavioural treatment, and multidisciplinary treatment (for subacute low back pain) are effective for pain relief. Finally, no evidence shows that other interventions (for example, lumbar supports, traction, massage, or acupuncture) are effective for acute low back pain.3

Box 3: Recommendations in the European clinical guidelines for diagnosis and treatment of chronic low back pain22


  • Diagnostic triage to exclude specific pathology and nerve root pain
  • Assessment of prognostic factors (yellow flags) such as work related factors, psychosocial distress, depressive mood, severity of pain and functional impact, prior episodes of low back pain, extreme symptom reporting, and patient’s expectations
  • Imaging is not recommended unless a specific cause is strongly suspected
  • Magnetic resonance imaging is best option for radicular symptoms, discitis, or neoplasm
  • Plain radiography is best option for structural deformities


Recommended—Cognitive behaviour therapy, supervised exercise therapy, brief educational interventions, and multidisciplinary (biopsychosocial) treatment, short term use of non-steroidal anti-inflammatory drugs and weak opioids.

If you like what I do, and you want me to do more, you can: buy my books Bad Science and Bad Pharma, give them to your friends, put them on your reading list, employ me to do a talk, or tweet this article to your friends. Thanks! ++++++++++++++++++++++++++++++++++++++++++

97 Responses

  1. coracle said,

    September 25, 2007 at 6:40 pm

    Nice one, let’s hope the journal comes up with the goods.

  2. RS said,

    September 25, 2007 at 6:44 pm

    I was ranting about the sort of people who get chronic back pain being just the sort who might be susceptible to placebo this morning as soon as I heard the study.

    You might wonder why physio was so much less effective thought.

  3. drchancer said,

    September 25, 2007 at 6:48 pm

    This post previously contained the full text of the paper. I contacted the American Medical Association to ask if they would consider making the paper freely available to the public, since it has been press released and discussed so widely, and public discussion of the study is hindered without access to the research paper describing the research, on which they own full intellectual property. They have asked me to remove this copy of the paper. I have deleted it.

    They will consider making the paper openly available, but they think the answer will be no.

  4. Quixotematic said,

    September 25, 2007 at 6:54 pm

    “I’ve written about the placebo before, how it about much more than a pill, but is about the cultural meaning of a treatment”

    Sometimes termed a ‘Therapeutic Encounter’.

    An acupuncturist friend, who once sat on the committee of the RCHM, told me years ago that it was already well established that it matters not a jot where the needles go.

    As an hypnotherapist myself, I can see that a valuable function of the needles is to get your client to lie still and shut up for 30 minutes. In some cases this is all the therapy they need.

  5. mrzero said,

    September 25, 2007 at 7:05 pm

    The radio broadcast is at 36 minutes

  6. ips said,

    September 25, 2007 at 7:08 pm

    you might find this of interest …..

    I am a NHS Physiotherpist with an interest to the thing that is most problematical pain . The physiology and education of pain has greatly improved see this for an example

    I am doing my MA dissertation on placebo and find the subject fascinating as it straddles brain science and culture . David Morris book the Culture of pain is highly recommended.
    Biomedicine is obsessed with biomechanics and structural dysfunction and perhaps acupuncture is a lot better than (often) futlile back surgery. I think much is made of the bio psych model of pain but it is seldom a base for clinical practice other than in some decent pain clinics .
    People as Ilich suggests are programmed into having things done to or for them so active treatements are often not very popular . I think acupuncture and other touch therapies are useful if education re graded return of fitness and pain sensitisation is given. Otherwise people are usually subjected to a treadmill of futile and expensive treatments .
    I don’t think there is that much difference between acupuncture and hypnotherapy they both work and influence the CNS ……the art is creating an encounter that will faciliate a physiological change (hopefully a lasting one)

  7. Despard said,

    September 25, 2007 at 7:32 pm

    Very nice paper.

    drchancer, thanks for posting the full text, but are you not concerned that the publisher might take legal action? At least wait until Ben has (hopefully) persuaded them to release it into the public domain.

  8. warhelmet said,

    September 25, 2007 at 7:57 pm


    You don’t think that there is much of a difference between acupuncture and hypnotherapy? I can sort of see where you are coming from, but my understanding is that there is some evidence to suggest that hypnotherapy is less effective in general.

    Common sense suggests that a physical intervention will do better with a percieved physical problem on the placebo effect.

  9. ips said,

    September 25, 2007 at 8:17 pm

    I was trying to suggest that there is much ‘central’ re oganisation in chronic pain and perhaps the main effect of the stimuation occurs not at the ‘physical’ spine but at the homoncular representation …..

    neuroscience seems a good way of reducing dualism in medicine?

  10. Will said,

    September 25, 2007 at 8:45 pm

    Isn’t the problem with placebo that if you are a rationally minded individual, as soon as you know that the effects are placebo, the effects will diminish or disappear? How can you ever say that it is worth funding these treatments when the public who are likely to receive them will also be likely to find out about their mode of action (or at least the arguments for and against them) – thereby rendering them useless?

  11. BobP said,

    September 25, 2007 at 8:48 pm

    Not at all impressed by David Tredinnick MP in the interview – but a good debate, with an excellent introduction – proves that the BBC is worth listening to sometimes.

  12. misterjohn said,

    September 25, 2007 at 9:22 pm

    The original article is available free.
    Benno Brinkhaus; Claudia M. Witt; Susanne Jena; Klaus Linde; Andrea Streng; Stefan Wagenpfeil; Dominik Irnich; Heinz-Ulrich Walther; Dieter Melchart; Stefan N. Willich
    Acupuncture in Patients With Chronic Low Back Pain: A Randomized Controlled Trial
    Archives of Internal Medicine, Feb 2006; 166: 450 – 457.
    ……Original Investigation 16 376 Acupuncture in Patients With Chronic Low Back PainA…Switzerland (Dr Melchart). Background: Acupuncture is widely used by patients with low…unclear. We investigated the efficacy of acupuncture compared with minimal acupuncture and……

    It says that there is little difference between real acupuncture and “fake” acupuncture, but both appear to be better than no acupuncture.

    The abstract of the new article is on this link

    These are both free so there’s no problem getting to read them. You won’t get Ben into trouble; and we wouldn’t want that, would we?

  13. evidencebasedeating said,

    September 25, 2007 at 10:07 pm

    Isn’t Trouble his middle name?

  14. junypharmd said,

    September 25, 2007 at 10:24 pm

    “The question is whether an elaborate, expensive, gimmicky and theatrical placebo ritual is an effective use of money, or whether other, cheaper, more pragmatic, honest psychosocial interventions might be more appropriate and cost effective.”

    Why not? Are “honest” psychosocial interventions more effective or just morally superior? If there are cheaper and simpler and honest interventions, great, but we all know such magical solutions are hard to come by.

    Between a lengthy and expensive course of honest psychotherapy lasting about 3 to 6 months and a shorter and elaborate course of placebo ritual — although less honest — is the professional inclination to the former really about effectiveness for the patient or about honesty and morality for the practitioner?

  15. ControlFreak said,

    September 25, 2007 at 10:55 pm

    It was surprising the range in media reports on this press release. Most reported reasonably accurately but the Independent’s online report was particularly poor and pro acupuncture, not mentioning that sham acupuncture was ‘nearly’ as good until quite a way in to it.

    None of the reports picked up that the p value for the two way sham vs “real” acupuncture was 0.39 at 95%, which most would class as no significant difference.

    To me this paper shows that all the acupuncture “theory” is likely to be complete nonsense ( no real surprise!).

    Are we likely to now see a doubling of all the CAM variants with “sham” methods springing up ? How could you tell the difference between real and sham homeopathy ?

  16. misterjohn said,

    September 25, 2007 at 11:30 pm

    Real homeopathy is a diluted version of sham homopathy.

  17. misterjohn said,

    September 25, 2007 at 11:31 pm

    Real homeopathy is a diluted version of sham homeopathy, is what I meant to say.

  18. Emilyb said,

    September 25, 2007 at 11:54 pm

    Two comments:
    I worked in a chronic pain clinic for many years. Many patients benefited from trigger point injections into the painful muscles. Low back acupuncture (sham or real) seems like it could work on a similar mechanism. The physicians all used different medications in the injections – I remember the director once telling me it didn’t really matter what you used, the benefit probably came from the action of the needle and the relaxation of the muscle in response to removal of the needle. Something like that…

    Secondly, I will read the entire article at work tomorrow, but I’m wondering…was the acupuncture (sham or real) new to these patients? Many pain patients are disillusioned with treatments they have tried in the past that offered no lasting benefit (ie nsaids, pt, etc). If it is the case that the acupuncture was the only new treatment, would it follow they would have more of a placebo response? It wears off eventually, no?

  19. Geoff_S said,

    September 26, 2007 at 1:44 am

    I heard the broadcast and appreciated Ben’s contribution. I have some personal experience of accupuncture and found it effective, which surprised me as I was very sceptical. Can the placebo effect work for a sceptic?

    In 1990 I had a cycle accident where I was thrown over the handlebars onto my head when a cat ran into my front wheel. Despite wearing a helmet I suffered spinal shock which totally paralysed me for several days. I have recovered to a large extent but still experience restricted movement and referred pain in my limbs. I can no longer cycle long distances as I used to (200 miles+ in a day) because of severe neck pain.

    One of the symptoms is leg pain which causes me to be restless whilst asleep (it sometimes wakes me). As ‘normal’ medicine offers no help I tried a lot of things (including reflexology, I’m ashamed to confess) and eventually accupuncture. Whilst I felt it hadn’t worked my wife assured me that she’d had the best night’s sleep for some time (no kicking!) and it also made my legs less restless whilst awake. During the treatment I certainly felt a tingling, like mild electric shock, when some of the needles were inserted.

    None of the other alternative therapies I tried had any effect at all. It’s surprising what you’ll try when there’s nothing on offer from conventional medicine.


  20. JLHarris said,

    September 26, 2007 at 3:10 am

    As a physical therapist, I’m just waiting for my patients to come running to me with this article, as told by mainstream media, asking why they shouldn’t get acupuncture instead.

    The problem is, this study is wholly unreproducible. The “western” medicine group (as MSNBC calls it in their coverage of this study) is an uncontrolled hodge-podge of treatments.

    We know heat, massage, and muscle relaxors don’t effectively treat LBP. And what was the phsysiotherapy treatment?

    “physician or physiotherapist who administered physiotherapy, exercise, and such.”

    And such, wow, don’t overwhelm us. Research by Childs, Dellito, Wainner, Erhard, and Cleland (to name a few) is showing us that subgrouping LBP patients (instead of lumping all – sprain/strain, disc, DDD, instability, etc – together) is demonstrating significant outcomes. This includes specific exercise.

    I worry the cultural definition of “physiotherapy” (which seems to not include exercise in Germany) will skew and inflate the results of this single, unreproducible study.

  21. csrster said,

    September 26, 2007 at 8:51 am

    Have there actually ever been any studies on whether placebos are ineffective if the patient knows they are placebos?

  22. apgaylard said,

    September 26, 2007 at 9:24 am

    Picking up on ControlFreak’s point on significance. The paper points out that the verum vs sham difference was 3.4% with a 95% CI including zero (3.7% to 10.3%). I’d say that’s good evidence of statistical insignificance.

  23. apgaylard said,

    September 26, 2007 at 9:26 am

    whoops, missed the minus sign out: 95% CI: –3.7% to 10.3%.

  24. FlammableFlower said,

    September 26, 2007 at 10:24 am

    I wonder if these guys have any opinions?

  25. Norbury said,

    September 26, 2007 at 12:17 pm

    I delayed calling my wife yesterday because that came on the radio. On the point of the dog with the arthritis, has any research been done to show how effective placebos are for animals? People seem to like pointing to homeopathy and now acupuncture working in animals as if that’s proof that they aren’t placebos, but that makes an assumption about the psychology of animals not being susceptible to placebo. Just a suggestion: at the primitive emotional level dogs aren’t necessarily that different from us so why shouldn’t they benefit from a placebo? Or am I talking bollocks?

  26. nekomatic said,

    September 26, 2007 at 12:28 pm

    “Have there actually ever been any studies on whether placebos are ineffective if the patient knows they are placebos?”

    That’s not really the question here though is it – it’s “if we treat the patient by sticking pins in them, does it matter if they think we’re sticking the pins in places prescribed by traditional Chinese teachings or not?” Either way the pins are being stuck in the patient, which is a real physical intervention even if only by (as noted above) making him/her lie still and shut up for half an hour.

  27. BSM said,

    September 26, 2007 at 1:27 pm

    “Or am I talking bollocks?”

    Speaking as a veterinary professional, you are talking bollocks.

    My exposure to the purveyors of the various modes of placebo therpay to animals has made it evident to me that there is such a large element of misinterpretation, misrepresentation and outright fibbing that I don’t see much space left for what one might call a ‘real’ placebo effect.

    But that’s just an opinion, I’m unaware of any trials ever to have tested the hypothesis.

    More interestingly, I think this raises the semantic issue of the definition of a placebo. Does it implicitly require conscious awareness by the patient that they are receiving a supposedly effective treatment?

    If, for instance, you told me that animals recovered better from surgery and showed less objectively recordable clinical signs of pain or distress if they are cared for gently and attentively compared with neglected controls, I would not be surprised, but I’m not sure I’d be happy to define that as a placebo response.

    However, those who advocate the idea that various sCAM modalities might work in animals ‘if only by a placebo effect’ usually go on to invoke the animal’s supposed awareness of the care it is receiving as constituting that placebo.

    Should we see sCAM placebos as having two elements. One is the ritual rigmarole associated with the administration of the specific therapy, the other is the therapy-specific placebo effect of the modality itself?

    Thus homeopathy has both elements and we sceptics aim our main criticism at the second element. We are prepared to acknowledge, and would find it uncontroversial, though interesting, that the ritual-element may be beneficial.

    Does the homeopathic pill do more than an blank sugar pill and does either do better than the going through the ritual without a pill being given at the end. Acupuncture can be seen to have the same elements- 1. ritual and 2. the therapeutic technique itself.

    Conversely, psychotherapy consists only of ritual and lacks the therapy-specific technical intervention.

    If psychotherapy works better than controls who rceieve no psychotherapy is that a placebo response? Or, do we need a new word?

    I’ve only just thought this through ‘on the fly’, so please feel free to pick holes in it, i.e. “am I talking bollocks?”

  28. BSM said,

    September 26, 2007 at 1:35 pm

    “Either way the pins are being stuck in the patient, which is a real physical intervention even if only by (as noted above) making him/her lie still and shut up for half an hour.”

    So, what you need is a sham sham-acupuncture to distinguish some putative positive effect from having needles non-specifically stuck into you versus merely thinking you are having needles stuck into you.

    Whenever I am led to think about this topic I am always struck that the clever bit of any trial is not designing the verum arm but defining the controls well enough to answer the question you really want answered.

  29. Despard said,

    September 26, 2007 at 2:09 pm

    #29: “Whenever I am led to think about this topic I am always struck that the clever bit of any trial is not designing the verum arm but defining the controls well enough to answer the question you really want answered.”

    You’re damn right. Designing experiments correctly is not an art, it’s… a science. :-)

  30. jack said,

    September 26, 2007 at 2:12 pm

    I’ve got back ache.

    Let me get this right.

    Its a 50/50 chance of success and it does nlt mater where or how deep I stick the pins?

  31. muscleman said,

    September 26, 2007 at 2:56 pm

    In my case being sedentary is most definitely a factor in back pain. I am unable to lie in bed of a morning reading, I dare not.

    the one that is mid thoracic and has immobilised me at times definitely benefits both from pressups and me running. I think it is that you throw your shoulders back to clear the top of your lungs when running and thus you strengthen your back muscles.

    The lumbar problem, more a general ache, benefits from the hamstring stretches and situps I do routinely as part of my running routine.

    When I run I have no back pain, easy. I shudder to think of how much it will hurt when I’m 70. Assuming I live that long I suppose.

    I have had physio manipulation of the thoracic spine and that did help, at least to the extent of giving me a window where I could initiate the exercise and therein lies the rub: when your back is sore the last thing you want to do is exercise.

    So a placebo combined with psychosocial advice would seem to the be the way to go. The placebo to give the window of symptom relief and the psychosocial to deal with the cause(s).

  32. muscleman said,

    September 26, 2007 at 3:00 pm

    emilypk: “With animals it is likely to be placebo-by-proxy. After all, does the *dog* tell you it feels better?”

    Not acupuncture and probably not placebo. But when the dog we had when I was a teenager had problems with rheumatoid arthritis a green lipped mussel extract pill hidden in his jellymeat transformed him from a stiff old dog who just wanted to sit in in the sun into a dog who gambolled and ran like a 3year old. The pills were expensive so he only had them when his symptoms warranted it. If nobody could tell if an animal was sick or in pain Vets would all be charlatans. Oh and if he noticed the pill it was left on the side of his bowl, licked clean. So no possibility of a placebo effect.

  33. jonecc said,

    September 26, 2007 at 3:36 pm

    Placebo used to be the vespers sung for the dead.

    In Latin, placebo means ‘I will please’. Vespers for the dead began ‘Placebo Domino in regione vivorum’, meaning ‘I will please the Lord in the land of the living’, and were known as the Placebo, as you might speak of the Te Deum or the Ave Maria.

    Other English words from the same root are placid, placate, placable and ultimately please.

  34. doris said,

    September 26, 2007 at 3:41 pm

    My final year research project at Surrey University examined the efficacy of acupunture-induced analgesia in a group of patients suffering chronic pain.
    The clinic was run by a consultant anaesthetist at St Luke’s hospital.
    The project involved monitoring the patients’ responses to the treatment and the use of a pain scale and a quality of life questionnaire.
    The results were not conclusive(typical under grad. study),but it was an extremely interesting experience.
    The anaesthetist offered the treatment as a last resort to his ‘hardcore’ patients,many of whom were extremely depressed and isolated.
    He had a pragmatic approach and felt that the treatment sessions offered some degree of solace and relief to individuals whose experience of mainstream treatments had been unsatisfactory.
    The placebo effect was no doubt a major component of the efficacy of the relief which several patients claimed.
    The doctor involved had no particular adherence to acupuncture per se,as I recall.
    This was in 1990,so my recollections are a little hazy.

  35. emilypk said,

    September 26, 2007 at 9:21 pm

    I am not suggesting all alt treatments for dogs are placebos. Just that humans can provide treatments as part of their overall husbandry and interpret behavior consistent with their expectations–thus producing what is effectively a placebo effect without the dog needing to comprehend the nature of medicine him- or herself.

  36. jodyaberdein said,

    September 27, 2007 at 8:43 am

    Regarding the ‘unplacebo’ effect, I’m not sure but by no means have I read all the literature. I do recall reading that there is the so called nocebo effect, for example placebo malaria prophylaxis results in adverse side effects.

  37. Norbury said,

    September 27, 2007 at 9:48 am

    “when the dog we had when I was a teenager had problems with rheumatoid arthritis a green lipped mussel extract pill hidden in his jellymeat transformed him from a stiff old dog who just wanted to sit in in the sun into a dog who gambolled and ran like a 3year old”
    That’s not quite what we’re discussing is it? The ‘alternative’ medicine arguement is seperate from homeopathy. There could well be something in green-lipped mussels that relieves the symptoms of arthritis, just as there is aspirin in willow bark, and lots of other ‘herbal’ remedies do contain active ingredients (not sure you could call a shellfish ‘herbal’ but u no wot i mene). The issue there is more to do with regulation, testing and control of potency.

  38. misterjohn said,

    September 27, 2007 at 11:42 am

    “for example placebo malaria prophylaxis results in adverse side effects.”
    Yes, malaria being the most common result. Placebo malaria prophylaxis is the same as none.

    My mother-in-law has recently recovered from malaria, and is now convinced of the benefits of malaria prophylaxis.

  39. ayupmeduck said,

    September 27, 2007 at 12:14 pm

    There is a paper here that seems almost identical:

  40. megachicken(b) said,

    September 27, 2007 at 1:00 pm

    Going back to a point made well back in the thread:

    “Have there actually ever been any studies on whether placebos are ineffective if the patient knows they are placebos?”

    Theoretically placebos might still work even if you know they are placebos. It depends on the mechanism governing the specific placebo reaction you are talking about.

    Some placebos work because of a conditioned responses. E.g. if you always get a genuine painkilling effect from a blue pill, then you learn to associate the analgesia (the unconditioned response) with the blueness of the pull (the conditioned stimulus). If you then take a blue pill you might still get an attenuated placebo response as a result of the conditioning, even if you know that this time it is a placebo pill.

    I’ve experienced the opposite with a nocebo response. I once had food poisoning after eating sausage and mash. After that, the smell of sausage and mash made me feel extremely quesy for weeks afterwards. I knew it was a psychological effect, but that didn’t stop the conditioned response from occuring.

  41. ayupmeduck said,

    September 27, 2007 at 1:06 pm

    I don’t understand this sentence from the study:

    “Nevertheless, the effectiveness
    of acupuncture cannot be attributed merely to
    a placebo effect because there is no reason to believe that
    the action mechanism of conventional therapy is the result solely of the placebo effect.”

    It seems to be saying that they do not in fact think it’s placebo effect, but from the wording, I can’t understand why. Can anybody clarify for me?

  42. Despard said,

    September 27, 2007 at 1:06 pm

    #45: that’s known as a learned taste aversion, and is well documented in pretty much all mammals. It’s one reason rats are so hard to kill with poison: they nibble a tiny bit, and if they get ill, they never touch any of that kind of food ever again.

    It’s also why you should never eat your favourite food while undergoing chemotherapy.

  43. megachicken(b) said,

    September 27, 2007 at 1:13 pm


    Hi Despard,

    Did you know you can also kill rats using pretty much the same effect – pair an immunosuppresant with flavoured water, then feed half of them flavoured water only and half unflavoured water. The ones still getting the flavoured water die in much larger numbers.


  44. ShatterFace said,

    September 27, 2007 at 1:21 pm

    “Have there actually ever been any studies on whether placebos are ineffective if the patient knows they are placebos?”

    There are a lot of optical and auditory illusions that work even though you know they are illusions, like the Müller-Lyer illusion with which you are probably familiar:


    Or the way the Moon looks enormous when it’s close to the horizon.

    These illusions are ‘cognitively impenetrable’ because our lower-level perceptual activities are modular.

    Maybe placebos can have some effect even though we are consciously aware that they *are* placebos?

  45. kim said,

    September 27, 2007 at 1:28 pm

    ayupmeduck – I think what this sentence, inelegant though it is, means is: “We can’t say the acupuncture patients got better because of a placebo effect, because some of the patients who received conventional treatment also got better. We think that the conventional treatment worked on its own merits, not as a placebo, therefore we have to assume that the acupuncture patients got better because of the merits of the treatment.”

    Implied, but not stated, is the following:
    “If there is a placebo effect, we would expect it to be the same in both cases, but actually, twice as many acupuncture patients got better as patients receiving conventional treatment.”

  46. ayupmeduck said,

    September 27, 2007 at 2:02 pm

    @Kim – Thanks. But logically, I still don’t get it. The first bit says:

    “Nevertheless, the effectiveness
    of acupuncture cannot be attributed merely to
    a placebo effect…”

    Ok, that’s pretty clear. They are saying that the positive effects of acupunture cannot be just placebo. The other half of the sentence is also pretty clear:

    “…there is no reason to believe that the action mechanism of conventional therapy is the result solely of the placebo effect.”

    The problem I have is the “because” that links the two clauses. I don’t see the logical connection. What rule is there to say that the placebo efffect of the acupunture has to be of a certain limited effect and cannot be greater that conventional therapy – especially when everybody agrees that conventional therapy in this area is not very effective.?

  47. buffalo66 said,

    September 27, 2007 at 2:36 pm

    @ ayupmeduck

    You’re right that the first half of this sentence doesn’t follow logically from the second half. Different placebos have different levels of effectiveness and, as you suggest, it could be that the placebo effect of acupuncture was bigger than the combined placebo and biomechanical effect of conventional therapy.

    Also, although it may be true that “there is no reason to believe that the action mechanism of conventional therapy is the result solely of the placebo effect”, it seems entirely possible that conventional therapy IS just a placebo.

    I think we can summarize their argument as:
    1. Assume that conventional therapy is not just a placebo.
    2. Assume that any placebo effect of conventional therapy would be as big as that of acupuncture.
    3. Given these assumptions (for which no evidence is presented), if acupuncture is only a placebo, it should be less effective than conventional therapy.

  48. buffalo66 said,

    September 27, 2007 at 2:52 pm

    @ShatterFace (#50)

    Just be thankful that you didn’t try to sketch the Hermann grid!

  49. ayupmeduck said,

    September 27, 2007 at 3:14 pm

    @buffalo: Thanks, so I think we both believe that the writers of the paper have come to a possibly false conclusion – i.e. that acuputure effects cannot be pure placebo.

    But to be fair to The Independent, David Tredinnick, etc. they would necessarily notice this error, if indeed it is one, and so their comments on this paper are not nearly as bizarre as some people here, including Ben, would claim.

  50. three tigers said,

    September 27, 2007 at 4:20 pm

    I don’t think the acupuncture or sham acupuncture is a placebo effect at all. It has all the hallmarks of the very real ‘counter irritation’ effect. Sticking needles (even fine ones) into skin and muscle would certainly cause local tissue damage and lots of needles = lots of counter irritation = less pain in the first site (in this case the lower back).

  51. kim said,

    September 27, 2007 at 4:32 pm

    buffalo66 – I think that’s a good summary of their arguments. And I think that the assumptions they make are made in the interests of fairness. If they were to say (and this, as I understand it, is Ben’s argument) “The placebo effects of acupuncture are stronger than the placebo effects of conventional treatment”, they would just be making another assumption, again without evidence.

    Of course, they could have said, “Maybe the placebo effects of acupuncture are stronger than those of conventional treatment, and maybe they’re not” but what scientist is going to admit to that degree of doubt in a published paper?

  52. buffalo66 said,

    September 27, 2007 at 4:55 pm

    kim – i don’t think it would have harmed them to say that they couldn’t determine whether the effects of acupuncture were pure placebo or not. They still had some very interesting findings – acupuncture is better than conventional therapy, but no better than having someone shove needles into you at random. Scientists always look silly when their conclusions go beyond what the data can support, e.g. the evolutionary claims in Hurlbert & Ling’s “girls prefer pink” paper:

  53. buffalo66 said,

    September 27, 2007 at 5:00 pm

    @ ayupmeduck (#55): “But to be fair to The Independent, David Tredinnick, etc. they would necessarily notice this error, if indeed it is one, and so their comments on this paper are not nearly as bizarre as some people here, including Ben, would claim.”

    That’s a good point – anyone who claims that this study shows acupuncture to be more than just a placebo is only reporting what the authors of the paper said. But I thought that David Tredinnick should have been more willing to discuss the fact that “real” acupuncture was no better than sham acupuncture – he didn’t seem interested in this finding at all.

  54. buffalo66 said,

    September 27, 2007 at 5:24 pm

    @ three tigers (#57) “I don’t think the acupuncture or sham acupuncture is a placebo effect at all. It has all the hallmarks of the very real ‘counter irritation’ effect.”

    But wouldn’t the counter irritation effect you describe cease after the source of irritation had been removed, rather than persisting for six months?

  55. emilypk said,

    September 27, 2007 at 5:34 pm

    I am not sure it shows accupuncture is better than conventional per se, but that it has additional benefit for the as yet uncured. If the group had been people who had tried only accupuncture for a period of time and still had symptoms, and they went to a real versus sham g.p. what would those results look like?

  56. buffalo66 said,

    September 27, 2007 at 6:22 pm

    good point, emilypk – novelty of the treatment to the patients does seem to be a confounding factor, as Emilyb pointed out earlier (#18).

  57. emilypk said,

    September 27, 2007 at 6:49 pm

    It is more than novelty per se. This is a non-random sample and so shouldn’t be over generalised. The study group excluded those for whom conventional treatment was successful; a group that might have responded less to accupuncture if they had received it instead of their effective mainstream care. And a group inclusive of those whose complaint has causes other then those effectively treated by mainstream medicine (potential including systemic and psychological factors).

    So it is not certain accupuncture (sham or otherwise) would be as effective as a first or simultaneous treatment regardless of the degree of novelty?

  58. Ben Goldacre said,

    September 27, 2007 at 7:01 pm

    lots of interesting points here.

    empilypk is absolutely right to point out that these were medical treatment failures of 8 years standing, not people passing through a surgery for the first time with back pain wondering “which treatment is best for me?”

    but most interesting is that odd interpretation by the authors, who seem perhaps to suggest that because the sham acupuncture group performed differently than the medical treatment group, the benefits of sham acupuncture (and real) cannot be solely due to the placebo effect.

    this is erroneous since it assumes that “placebo” is a unitary phenomenon. however there have been specific studies comparing one placebo with another showing that it is highly variable, related as it is to expectations, ritual, and meaning. there is even a paper (referenced in the post) comparing a placebo pill with a sham ritual involving a placebo device, modelled on acupuncture. i’ll have a look at the paper more carefully and might bung a reference-laden letter off to the journal on that, i think it’s well interesting.

  59. misterjohn said,

    September 27, 2007 at 11:04 pm

    This article sheds some more light on the situation, I think.

    Is acupuncture effective for the treatment of chronic pain? A systematic review.

    Ezzo J, Berman B, Hadhazy VA, Jadad AR, Lao L, Singh BB.

    Complementary Medicine Program at the University of Maryland School of Medicine, Kernan Hospital Mansion, 2200 Kernan Drive, Baltimore, MD 21207-6697, USA.

    Pain is the major complaint of the estimated one million U.S. consumers who use acupuncture each year. Although acupuncture is widely available in chronic pain clinics, the effectiveness of acupuncture for chronic pain remains in question. Our aim was to assess the effectiveness of acupuncture as a treatment for chronic pain within the context of the methodological quality of the studies. MEDLINE (1966-99), two complementary medicine databases, 69 conference proceedings, and the bibliographies of other articles and reviews were searched. Trials were included if they were randomized, had populations with pain longer than three months, used needles rather than surface electrodes, and were in English. Data were extracted by two independent reviewers using a validated instrument. Inter-rater disagreements were resolved by discussion. Fifty one studies met inclusion criteria. Clinical heterogeneity precluded statistical pooling. Results were positive in 21 studies, negative in 3 and neutral in 27. Three fourths of the studies received a low-quality score and low-quality trials were significantly associated with positive results (P=0.05). High-quality studies clustered in designs using sham acupuncture as the control group, where the risk of false negative (type II) errors is high due to large sample size requirements. Six or more acupuncture treatments were significantly associated with positive outcomes (P=0.03) even after adjusting for study quality. We conclude there is limited evidence that acupuncture is more effective than no treatment for chronic pain; and inconclusive evidence that acupuncture is more effective than placebo, sham acupuncture or standard care. However, we have found an important relationship between the methodology of the studies and their results that should guide future research.

  60. misterjohn said,

    September 27, 2007 at 11:28 pm

    A homeopath writes (to the bmj)in response to this article;-

    Effect of homoeopathy on pain and other events after acute trauma: placebo controlled trial with bilateral oral surgery
    Per Lokken, professor,a Per Atle Straumsheim, Dag Tveiten, homoeopath,b Per Skjelbred, chief surgeon,c Christian Fredrik Borchgrevink, professor d
    a Section of Dental Pharmacology and Pharmacotherapeutics, University of Oslo, 0316 Oslo, Norway, b Norwegian Institute for Natural Medicine, 1340 Bekkestua, Norway, c Section of Maxillofacial Surgery, Oslo City Hospital, 0407 Oslo, Norway, d Department of General Practice, University of Oslo, 0264 Oslo, Norway

    Saint Hanshaugen Medical Centre, 0168 Oslo, Norway Per Atle Straumshein, homoeopath. Correspondence to: Professor Borchgrevink, Fr. Stangs gt 11/13, 0264 Oslo, Norway.


    Objective: To examine whether homoeopathy has any effect on pain and other inflammatory events after surgery.
    Design: Randomised double blind, placebo controlled crossover trial with “identical” oral surgical procedures performed on two separate occasions in 24 patients.
    Interventions: Treatment started 3 hours after surgery with either homoeopathy or placebo.
    Main outcome measures: Postoperative pain and preference for postoperative course assessed by patients on visual analogue scales. Measurements of postoperative swelling and reduction in ability to open mouth. Assessment of bleeding after surgery.
    Results: Pain after surgery was essentially the same whether treated with homoeopathy or placebo. Postoperative swelling was not significantly affected by homoeopathy, but treatment tended to give less reduction in ability to open mouth. No noticeable difference was seen in postoperative bleeding, side effects, or complaints. Thirteen of the 24 patients preferred the postoperative course with placebo.
    Conclusions: No positive evidence was found for efficacy of homoeopathic treatment on pain and other inflammatory events after an acute soft tissue and bone injury inflicted by a surgical intervention. Differences in the order of 30% to 40% would have been needed to show significant effects.

    The homeopath responds;-

    Simon Rabinovich,
    Homeopathic physician, Homeopathic Consulting Co.
    Toronto, Ontario, Canada

    I would like to discuss two points of provided article. Firstly, I must draw your attention to the methodology of giving homeopathic drugs “3 hours AFTER the surgery”. Any experienced homeopath would tell us that according to homeopathic science any surgical intervention including extracting of teeth is a clear CONTRAINDICATION to homeopathic treatment, i.e. that treatment is doomed to be ineffective at all. The surgery stands in the row with such other contraindications to homeopathic treatment like excessive coffee drinking and consuming mint in any form including toothpaste using. Actually, if the researchers desired to see ANY effect of homeopathy they should have given homeopathic medications BEFORE the surgery. It is advisable also to repeat the treatment after the surgery although no sufficient evidence exists in all known clinical trials, however extensive clinical experience indicates clear benefits of such move.

    Secondly, the choice of homeopathic remedies and their dosage (potency) were, in my view, somehow inappropriate. The only homeopathic remedy that I would agree with giving to the patients in such circumstances was Arnica. Any “individualized homeopathic assessment” after such painful procedure would obviously lead to the almost similar “drug picture” as it was shown in your article and, accordingly, to the same drug – Arnica. Practically speaking, the results would have been remarkably better and indicating the obvious effectivness of homeopathic treatment if such remedy like Aconitum Nappelus was added to the treatment. The latter medicine is usually been prescribed in the cases of “great anxiety, mental shock and, sometimes, panic”. It is also known to any experienced homeopath that Aconitum is one of “homeopathic pain- killers” (because the feeling of pain is the expression of CNS’ condition). It is also known that Aconite demonstrates its extremely well effectiveness and works in almost every case, thus it is called the “universal” remedy just like Arnica. I presume, this attitude would have been totally denied by classical homeopaths you invited for participating in your trial.

    In conclusion, the use of correct homeopathic treatment in the case you presented would have been proved extremely effective, as it happened not once in the clinical experience of our clinic and number of dental surgeon’s clinics in Toronto.

  61. Squander Two said,

    September 28, 2007 at 10:32 am

    I find it interesting that almost no-one here is even discussing the possibility that this not be placebo. It’s hardly far-fetched to suggest that sticking needles into flesh can have an effect on that flesh, yet the entire discussion here seems to revolve around the certainty that the only way these patients could have got positive results is placebo.

    Sure, traditional acupuncturists talk a lot of bollocks about energy lines and such, and I’d heard long before this study that the evidence shows that the needles work just as well no matter where you stick them. A friend of mine is a chiropractor, and she uses needles to relax muscles, but she refuses to call it “acupuncture” on the grounds that acupuncture is quackery whereas what she’s doing is based on evidence. Another friend needed therapy after a rather nasty car crash, and his GP referred him to a specialist who used needles but, again, didn’t call it acupuncture, telling his patients frankly that all the stuff about acupressure points is bollocks. What he said was that the problem with most treatments for muscle problems is that they are confined to the surface; using needles allows you to apply heat to the muscle much deeper down — he used needles with heating elements. That theory is at least sensible enough to warrant proper consideration of the physical mechanisms, rather than to write it off out of hand as psychological.

    Meanwhile, my GP’s “treatment” of disabling back pain was to prescribe me ibuprofen, giving me the wonderful opportunity to buy for £6.50 some pills that would usually cost me 60p. If you talk to patients who have left the conventional system to have their back pain treated, you don’t find a large number of the anti-MMR, anti-chemicals, anti-pills brigade, but rather lots of people who believe in the benefits of evidence-based medicine and were extremely disappointed by their doctors’ ineffectuality in this area. When it comes to joints and muscles, many GPs won’t even touch the affected area, instead merely writing down the patient’s description of the pain, and their treatments tend to revolve around painkilling rather than isolating the cause.

    Conventional medicine is brilliant at surgery, the workings of internal organs, and biochemical effects, but has a bit of a blind spot when it comes to the physical mechanics of the skeleton and muscles. No idea why.

    On the subject of experiment design, I amaware that there have been proper scientific studies into the effects of enemas. I’ve often wondered about the placebo control group: how do you persuade someone that they’re receiving an enema when they’re not?

  62. wilsontown said,

    September 28, 2007 at 10:52 am

    To add to that, if the placebo effect varies for different placeboes, how can you perform placebo-controlled trials? They would only be controlled for one particular placebo. For example, if, as I would assume, homeopathy is basically an elaborate placebo, it might be a more effective placebo than a blank pill given without all of that homeopathic individualised treatment. Then, a ‘placebo-controlled’ trial would be bound to show an effect for homeopathy.

    It is confusing…(or maybe I’m just confused)…

  63. kim said,

    September 28, 2007 at 11:25 am

    Are people doing research on how placebos work? Presumably it’s not a question of thinking, “I’ve been given a pill and now I feel better.” I assume there’s some process going on where the brain is sending out chemical signals that somehow speed up the body’s own healing process. Likewise, when you’re depressed, as is often the case with chronic back pain sufferers, perhaps your brain is sending out chemical messages that block that process.

    Is it possible to test in any way what actually happens when someone receives a placebo?

  64. kim said,

    September 28, 2007 at 11:30 am

    I think Suw and SquanderTwo make very good points – to assume that the acupuncture is just acting as a “placebo” is to explain away an incovenient result in the way that alternative medical practitioners are often accused of doing.

    The other interesting thing about this study is that it couldn’t have been double-blind – the people applying the “fake” acupuncture must have known it was fake, yet the patients still got better. Interesting.

  65. buffalo66 said,

    September 28, 2007 at 12:19 pm

    @ Suw (#70). You’re right that it’s difficult to define what is meant by a placebo effect. If it’s defined as a purely “psychological” effect, then that means any psychological treatment, like cognitive behavioural therapy, is a placebo by definition. I think a placebo is a treatment that lacks something that is claimed to be an essential element (physical or psychological) of the treatment being assessed. If the essential element is claimed to be putting the needles in the right place, then the sham acupuncture in this study is a placebo. But if it’s claimed that it’s the physical effect of putting needles into someone that causes the improvement, and the exact location doesn’t matter, then the placebo treatment can’t involve inserting needles.

  66. ayupmeduck said,

    September 28, 2007 at 2:09 pm

    @Squander Two: I see what you are saying. Maybe I’m being too pedantic, but it seems to me that both in the discussion here and in the paper, the terms are not well defined.

    Look at it this way: The authors of the study have defined Acupuncture and have also defined “sham acupuncture”. To my mind the “sham acupuncture” is *NOT* acupuncture. It doesn’t use the rules that the normal acupuncture professionals practice. It’s just “random needle insertion” – I think this is a clearer term than “sham acupuncture”.

    In this sense they seem to have shown that Acupuncture actually has *NO* effect whatsoever.

    They have shown that “random needle insertion” does have an effect. The “random needle insertion” effect may be a placebo effect or it may not.

    Or is my own logic tying me in knots?

  67. emilypk said,

    September 28, 2007 at 2:50 pm

    Indeed. I would think that what makes it a placebo effect it that the sham inaccupuncturists stuck the needles in some harmless but random place. If where you stick them doesn’t matter at all then accupuncture is at the very least rather less complicated than one has been lead to believe.

  68. doris said,

    September 28, 2007 at 3:46 pm

    A fascinating discussion,although I don’t know whether I ‘ve fully followed all the logical arguments.
    However,more personal stuff to add.
    My other sister,who only has one functioning kidney,has been having regular weekly acupuncture treatments for the past couple of years.
    She swears by their eficacy:the lumbar pain and accompanying urinary difficulties which used to trouble her regularly,are now much improved:(howver,before objections are raised,she hasn’t had kidney function tests done for some time).
    She had a renogram a few years ago,and was offered a cystoscopy a while later.
    She has found the conventional investigations so distasteful and distressing that she has decided to rely on acupuncture.
    her argument is that,since she has relied on only one kidney for the past 36 years,she knows instinctively when things are going wrong for her.
    I really don’t know whether this can be attributed solely to a placebo effect,and as I said earlier,no clinical data is available to substantiate her assertions.
    However,I find her case fascinating,not least because she is my sister.
    Finally,to return to the far off days of my undergraduate study;the anaesthetist who provided the acupuncture for pain relief used the chinese system but did not adhere to the belief system.
    He said that if it seemed to work when all else had failed,that was fine by him.
    Again,any placebo effect was not separated out during my study.

  69. Joe Dunckley said,

    September 28, 2007 at 6:17 pm

    Apologies if anything like this has already been discussed, I haven’t had time for anything more than a glance over the thread so far.

    So, anyway, I just happened to catch the letters on PM. This bloke objects to having his back pain described as “imaginary”, does he? I didn’t hear Ben’s piece, so I don’t know whether he used the word “imaginary”, but if Ben didn’t, it’s interesting that the letter writer chose to use that word, for several reasons, starting with the ambiguity of the word.

    Does “imaginary” (and thus psychosomatic) mean to the letter writer “making it up”? If so, I think he’d be quite right to be furious (leaving aside the obvious).

    Or perhaps the letter writer really does understand psychosomatic and is scared of it: an activity of the brain that is unconcious and beyond his control. It starts sounding awfully like a mental illness. Feeling pain that isn’t really there is not a very big step from hearing voices that aren’t really there. Is it another case of the mental illness taboo? Mental illness is still seen more as an accusation than a diagnosis. Much better to be diagnosed with an injury than accused of having an imagination.

  70. Pepper said,

    September 28, 2007 at 8:33 pm

    Silly paper
    These results can appeared only due to NON-COMPETENCE OF AUTHORS IN ACUPUNCTURE! Or unsufficient competence.
    That is why – the paper presents almost the same data (%) both in real acupuncture and in sham acupuncture!
    Read the paper attentively.
    You are discussing here THE INCORRECT DATA!!
    And no more.

  71. joal said,

    September 29, 2007 at 6:04 pm

    Thanks for an excellent site.

    This is very interesting indeed as it confirms the history of mankind and our ancestors widespread use of voodoo like ceremonies. It works! The more elaborate the ceremony the more effective it will be. Mind over matter!

  72. Pepper said,

    September 29, 2007 at 9:56 pm

    It seems this paper about acupuncture is wrong.
    These results can appeared only due to NON-COMPETENCE OF AUTHORS IN ACUPUNCTURE! Or unsufficient competence.
    That is why – the paper presents almost the same data (%) both in real acupuncture and in sham acupuncture!
    Read the paper attentively, please.
    We are discussing here THE INCORRECT DATA.
    And no more.

    Are you wrong?

  73. Ben Goldacre said,

    September 30, 2007 at 10:41 pm

    psychosocial has absolutely nothing whatsoever to do with psychosomatic or psychogenic. psychosocial risk factors are very significant in back pain, that means psychological and social factors are very important in back pain. there was no ambiguity, even for people who can’t understand english, i gave clear examples, things like, whether you rest it, whether there is a low-strain job for you to go back to as an intermediate post, depression, brief educational interventions, public information programs, all the stuff i specifically mentioned. psychosocial risk factors and interventions, like i said.

    i didnt say back pain was psychosomatic. i didnt say back pain was psychotic. i didnt say back bain was psychopathic or sycophantic or cyclical. i said that psychosocial risk factors were significant in the development and treatment of back pain, and they are. very.

    in fact, amazingly, even the person whose letter they read out didnt say psychosomatic. its post 11 here:

    the letter said “I would take issue with your speaker, Ben Goldacre,) who, if I recall correctly, said that 90% of back problems are physcosymatic disorders.”

    so not only did this person seem to ahve a completely garbled understanding of what i said, that was, in fact, completely obvious from what they wrote, from the fact that the PM team had to go out of their way to make sense of their garbled comment in order to say i’d said something which i didnt say. if they’d read out someone angry that i’d said back pain was a “physcosymatic disorder” then i wouldn’t be so bothered.

    to be honest i’m amazed and appalled. this is why doctors and academics talk at length amongst themselves about what a bad idea it is to have anything to do with the media.

    i have never, and would never, say that back pain is psychosomatic. there are psychological and social risk factors for acute back pain becoming chronic, of which i gave many good, clear, evidence based examples in the program, and there are examples of psychosocial interventions which are probably better than acupuncture, as a psychosocial intervention, such as brief educational interventions, public education programs (eg the australian “back pain, dont take it lying down” project i describe in the other post), and so on.

    back pain is not physcosymatic. nor is it psychosomatic.

    i have never suggested back pain is psychosomatic.

    i would never suggest that back pain is psychosomatic.

    back pain is not psychosomatic.

  74. RS said,

    October 1, 2007 at 10:31 am

    Some back pain is psychosomatic surely Ben?

  75. buffalo66 said,

    October 1, 2007 at 10:34 am

    Ben – I don’t think you can really blame people for being confused about the difference between psychosocial and psychosomatic. The Concise Oxford Dictionary (8th edition) has these definitions:

    psychosocial: of or involving the influence of social factors or human interactive behaviour.

    psychosomatic: 1 (of an illness, etc.) caused by or aggravated by mental conflict, stress, etc. 2 of the mind and body together.

    Now, Table 1 in the BMJ paper you cited lists the following risk factors as “psychosocial”:
    Stress; anxiety; negative mood or emotions; poor cognitive functioning; pain behaviour.

    Number 1 in this list of “psychosocial” risk factors is “stress”, which is specifically mentioned in the Oxford dictionary’s definition of “psychosomatic”: according to this definition, an illness that is caused by or aggravated by stress is “psychosomatic”. I realise that there may be subtle differences between a “risk factor” and a “cause” or “aggravating factor”, but I don’t think you could blame anyone for thinking that the terms “psychosocial” and “psychosomatic” are virtually synonymous.

    Perhaps medical professionals like yourself use a different definition of “psychosomatic” from that used in everyday language typified by the Oxford Dictionary definition (like the way “significant” has a precise meaning to statisticians which can be at odds with the definition used in ordinary language)?

    It would be really helpful if you could explain how your definition of “psychosomatic” differs from “psychosocial”.

  76. buffalo66 said,

    October 1, 2007 at 11:57 am

    i posted this earlier but, bizarrely, it can’t be seen unless you log in as me. Here’s another try:

    Ben – I don’t think you can really blame people for being confused about the difference between psychosocial and psychosomatic. The Concise Oxford Dictionary (8th edition) has these definitions:

    psychosocial: of or involving the influence of social factors or human interactive behaviour.

    psychosomatic: 1 (of an illness, etc.) caused by or aggravated by mental conflict, stress, etc. 2 of the mind and body together.

    Now, Table 1 in the BMJ paper you cited lists the following risk factors as “psychosocial”:
    Stress; anxiety; negative mood or emotions; poor cognitive functioning; pain behaviour.

    Number 1 in this list of “psychosocial” risk factors is “stress”, which is specifically mentioned in the Oxford dictionary’s definition of “psychosomatic”: according to this definition, an illness that is caused by or aggravated by stress is “psychosomatic”. I realise that there may be subtle differences between a “risk factor” and a “cause” or “aggravating factor”, but I don’t think you could blame anyone for thinking that the terms “psychosocial” and “psychosomatic” are virtually synonymous.

    Perhaps medical professionals like yourself use a different definition of “psychosomatic” from that used in everyday language typified by the Oxford Dictionary definition (like the way “significant” has a precise meaning to statisticians which can be at odds with the definition used in ordinary language)?

    It would be really helpful if you could explain how your definition of “psychosomatic” differs from “psychosocial”.

  77. emilypk said,

    October 2, 2007 at 9:04 pm

    but can a placebo work while you are unconscious?

  78. RS said,

    October 3, 2007 at 9:05 am

    Since in animals the placebo effect is on the owners – anaesthetising the animal won’t make any difference.

  79. emilypk said,

    October 3, 2007 at 2:22 pm

    Since the animal is an exotic one on exhibit with minimal human contact I suspect it would. The only significant human contact would be during the procedure itself.

  80. RS said,

    October 3, 2007 at 6:26 pm

    emilypk – the whole point is who determines whether the animal has gained any benefit? You can’t ask it!

  81. emilypk said,

    October 3, 2007 at 7:30 pm

    You can however observe whether there is a return to normal movement, activites and socialisation as well as measure cortisol or other stress hormones in the faeces. On the whole speech is one of the less reliable pain indicators.

  82. RS said,

    October 3, 2007 at 8:31 pm

    kim, the same could be said for any medical procedure, veterinary or otherwise, but we know that without blinding doctors assessing the patients will ‘detect’ lower levels of pain (which is the whole problem with non-objective measurements).

    And I note that there is no evidence being presented here that acupuncture has helped these arthritic monkeys – so I don’t know what the talk of cortisol has to do with anything.

  83. emilypk said,

    October 3, 2007 at 9:15 pm

    It was an answer to the ‘you can’t ask’ statement which implies that pain can’t be diagnosed in the non-verbal subject. Cortisol typically increases in response to pain and has been used to assess lameness adn post-op pain in famr and zoo animals.

    I don’t know if they are doing a formal analysis of the effectiveness, but they certainly could.

  84. emilypk said,

    November 13, 2007 at 4:51 pm

    On the issue of vetwoo, thought this might interest-slash-amuse:

  85. emilypk said,

    November 16, 2007 at 6:29 pm

    I had an interesting thought today. What if these unsuccessfuly treated sufferers tended to have a history of using drugs that work via opiod receptors. Given that it is suggested that acupuncture operates via the same pathways they might be ‘immune’ to the potential benefits beyond placebo?

  86. malucachu said,

    December 13, 2008 at 12:10 pm

    Physiotherapy is a profession, not a treatment!

    Now that we have got that out of the way, let me explain my place in this circus. I am a physiotherapist with a particular interest in Low Back Pain (co-author of ‘The Management of Low Back Pain in Primary Care’ ISBN 0 7506 4787 6 – now out of print, but available in many medical libraries).

    IMHO, physiotherapy is simply about motivating patients to follow the best available guidelines and advice on self care. It’s that simple.

    However many physiotherapists have spent a great deal of time and money learning quasi-medical treatments and thus have a vested interest in making treatment care complicated and full of gobbledygook.

    The good news is that this way of thinking is thankfully changing, albeit slowly.

  87. RichardCarter said,

    May 24, 2009 at 10:45 pm

    …. and now we read, from The Observer’s Health Correspondent no less, that back sufferers to receive acupuncture on the NHS – and, worse, that NICE is apparently going to recommend this!

    Story here:

    Interestingly, The Guardian/Observer’s website had a story, on 12 May only, to the effect that simulated acupuncture using toothpicks pressed against skin was just as effective as “real” acupuncture:

    This appears to be based on a different BMJ paper than the 2006 one that Ben discusses above, but evidently with the same results. I wonder what NICE will have to say about this (if, indeed, The Observer report is accurate, not something I’d bet the farm on, given that paper’s previous form).

  88. RichardCarter said,

    May 26, 2009 at 11:53 pm

    I seem to have maligned The Observer: it turns out this ludicrous thing is true and NICE have, according to the BBC website anyway (, confirmed that they are recommending acupuncture (among other useless ‘remedies’) for the ‘treatment’ of back pain.

    The report says that NICE are saying that “anyone whose [back] pain persists for more than six weeks and up to a year should be given a choice of several treatments, because the evidence about which works best is uncertain.” I love he last bit!

  89. MJJMZ60 said,

    August 11, 2009 at 2:32 pm

    Pain is a difficult item to measure. What about papers where measurable things were measures, such as WBC?

    For example the following study (From Medline!)

    Lu W, Matulonis UA, Doherty-Gilman A, Lee H, Dean-Clower E, Rosulek A, Gibson C, Goodman A, Davis RB, Buring JE, Wayne PM, Rosenthal DS, Penson RT.

    Leonard P. Zakim Center for Integrative Therapies, Dana-Farber Cancer Institute, 44 Binney Street, Boston, MA 02115, USA.

    OBJECTIVES: The objective of this study was to investigate the effect of acupuncture administered during myelosuppressive chemotherapy on white blood cell (WBC) count and absolute neutrophil count (ANC) in patients with ovarian cancer. DESIGN: This study is a pilot, randomized, sham-controlled clinical trial. Patients received active acupuncture versus sham acupuncture while undergoing chemotherapy. A standardized acupuncture protocol was employed with manual and electrostimulation. The frequency of treatment was 2-3 times per week for a total of 10 sessions, starting 1 week before the second cycle of chemotherapy. SETTING: The setting was two outpatient academic centers for patients with cancer. SUBJECTS: Twenty-one (21) newly diagnosed and recurrent ovarian cancer patients were the subjects. OUTCOME MEASURES: WBC count, ANC, and plasma granulocyte colony-stimulating factor (G-CSF ) were assessed weekly. RESULTS: The median leukocyte value in the acupuncture arm at the first day of the third cycle of chemotherapy was significantly higher than in the control arm after adjusting for baseline value (8600 cells/microL, range: 4800-12,000 versus 4400 cell/microL, range: 2300-10,000) (p = 0.046). The incidence of grade 2-4 leukopenia was less in the acupuncture arm than in the sham arm (30% versus 90%; p = 0.02). However, the median leukocyte nadir, neutrophil nadir, and recovering ANC were all higher but not statistically significantly different (p = 0.116-0.16), after adjusting for baseline differences. There were no statistically significant differences in plasma G-CSF between the two groups. CONCLUSIONS: We observed clinically relevant trends of higher WBC values during one cycle of chemotherapy in patients with ovarian cancer, which suggests a potential myeloprotective effect of acupuncture. A larger trial is warranted to more definitively determine the efficacy of acupuncture on clinically important outcomes of chemotherapy-induced neutropenia.

  90. diudiu said,

    December 21, 2009 at 5:43 am

    ed hardy ed hardy
    ed hardy clothing ed hardy clothing
    ed hardy jeans ed hardy jeans
    christian audigier christian audigier
    ed hardy t shirts ed hardy t shirts
    ed hardy uk ed hardy uk
    ed hardy bags ed hardy bags
    ed hardy hoodies ed hardy hoodies
    ed hardy mens ed hardy mens
    ed hardy womens ed hardy womens
    ed hardy kids ed hardy kids ed hardy kids

  91. Snuggie said,

    February 5, 2010 at 8:52 am

    Snuggie blanket Snuggie blanket
    blanket with sleeves blanket with sleeves

  92. JustMe said,

    July 1, 2010 at 12:23 am

    It appears to me that people are not subjecting Ben Goldacre’s views to the same kind of critical appraisal he is advocating. I agree with much of what Dr Goldacre says in terms of EBM, but whilst it should form the foundation of research and medicine, it is not perfect. The views expressed by Dr Goldacre are all being supported with attempting to question whether there are in fact other reasons why this may actually be a positive result. This shouldn’t be viewed as a study which demonstrates that acupuncture is no more effective than sham acupuncture (which one person above comments “is NOT acupuncture”. This highlights the ignorance and lack of thought that is given when interpreting results, even in a community that believe themselves to be proficient in critical appraisal. This study actually shows pretty much NOTHING…..not one way or another. If it does show anything, it shows that the theory of meridians is probably flawed. However, it also shows that actually just putting several needles into the skin is beneficial. Placebo? Maybe yes, maybe no. It may well also be that the needling actually results in physiological changes (such as endorphin release) which benefit pain. It maybe that regularly doing this allows a physiological state of relaxation to be achieved…I wonder how many people have actually tried acupuncture? The lesson of Dr Goldacre’s views are certainly beneficial, we should all seek to be critical but it should not be “critical” in the sense that all positives should be dismissed as nonsense purely in a bid to sound like a “good scientist”. If we are to “appraise”, the correct way is to debate both sides of the argument and offer a fair interpretation. The comment at the top which says ” I was ranting about the sort of people who get chronic back pain being just the sort who might be susceptible to placebo this morning as soon as I heard the study” – sums this all up, a sweeping statement that implies that people who have chronic pain are just mad, depressed people who are suggestible enough to “fall” for the placebo effect (despite Dr Goldacres statement about the complexity of the placebo effect). Everyone of us is susceptible. Despite how much your ego may be telling you you are different, your mind stronger, you are not.

  93. Pierre de Lasteyrie said,

    January 26, 2011 at 11:45 pm


    I’ll try to share my doubts about acupuncture using a different perspective:

    If you want a therapy to be effective you need a balance between the diagnostic and treatment capabilities. In fact you should have more diagnostic tests that treatment techniques. Why is that ? Because for each “treatment technique” you need a diagnostic test to tell you that you need this technique applied to that area. Theoretically you even need more diagnostic techniques in order to delimit the capabilities of the therapy itself.

    If you don’t have enough diagnostic tests but too many treatment possibilities, your therapy will be like playing darts with a blind fold!

    For example in modern medecine you have blood test, BP, spirometer, ECG, EEG, MRI, Xray, orthopedic test, palpation…Each variation within the results will imply a different treatment and the range of treatment is pretty big too (pills, physio, casts, diet advice, dialysis, surgery…)

    What about acupuncture ?
    During a session you will “stab” about 10 needles and there is 350 points of acupuncture (2000 if we follow more recent trends). If we consider that each point could be stabbed (because this therapy is really holistic) it means that there is 350!/340! different possible treatments. If I am not wrong this is equals to 2.4 x 10^25 possible treatments or 24 000 000 000 000 000 000 000 000 (1,0 x 10^33 with the 2000 acupuncture points)!!! The number of different possible treatments is enormous.

    Now Let’s have a look at the diagnostic capabilities of acupuncture:
    4 pulses sites(?) with 3 different depths. 7 possible colours in the tongue, eyes and cheeks and 10 questions with 7 different answers possible.
    If we settle on these, this is 7.8 x 10^12 different diagnostic possibilities.

    You can add a few tests if you wish. But for the moment the ratio is 1 diagnostic test for 3 x 10^12 possible treatments: You will need a bit more than luck to get the right treatment…

  94. danleywolfe said,

    September 26, 2011 at 8:37 pm

    One of the wonderful things about blogs like this one is the range of responses and opinions that are given and that they are timeless or you might say cumulative over time … learning on learning. As a former acupuncture patient (treatment of ideopathic neuropathy) I can say that I found only incremental and temporal relief that is after treatment stops there was a return to the original condition. I am unable to judge and there is not enough evidence to make inferences as to whether the treatment caused an improvement or a placebo effect was at work. There is no question that electro – needling stimulates the nerve channels running from the neck to the pinkies. And little doubt that distal symmetric primarily axonal (rather than demyelinating ) sensory “dying back” neuropathy involves a disabling of nerve transmission. So the theory goes that acupuncture can juice up the nerves involved, at least temporarily. by the way the German paper now is available to the public free of charge at:

  95. Points of Life said,

    October 6, 2013 at 5:51 am

    Hello all, sorry I found this blog after so many years, but what’s a century or a two. I just want to add what Acupuncture is and what it is not. And believe it or not all of it is documentable if you read the associated textbooks.

  96. Vic said,

    November 17, 2013 at 11:41 pm

    Perhaps a lot of attacks on alternative medicine (I prefer the name complementary medicine) are from the conventional medical establishment who have not considered the following: if the complementary practitioner can cause the powerful “placebo” to help the patient, why poo-poo it?

    The “elaborate” one hour procedure results in the relief of symptom the patient experiences, so he/she continues the treatment – its their choice and their money, surely! The people who come to see me have exhausted what help the conventional medical can give them and about 80% of who I treat get the relief they are looking for. They would not come back if they did not value the results they get.

    Acupuncture is very difficult to trial with double-blind tests because putting a needle in a non-point cannot be guaranteed to have no effect! And choosing the point is not one done by mechanical computation but a complex, individual assessment that changes as the patient changes and gets better.

    So please – don’t knock procedures that help people, especially as the NHS is crippled by lack of resources and many hospitals are under investigation for incompetence and negligence.

  97. pogopatterson said,

    October 9, 2014 at 2:00 pm

    I undertook acupuncture for 5 years for stress following brain surgery, and although it didn’t yield all the benefits I hoped it would, I know that it should be considered to be a serious medical therapy for the neurological/emotional instability caused by damage to parts of the hypothalamus (possibly hypocretin neurons) following brain surgery. Not that surgery itself is classically seen as damaging in its own right by modern medicine- only the problem it was being used to overcome is ever mentioned. That in itself is a form of semantic quackery- but I digress.

    The study on back pain has been unscientifically used as a tool to bludgeon all acupuncture. I know people who have had no success for their back pain from acupuncture. Without the benefit of the study above, I didn’t conclude that I was merely undergoing a placebo affect, and these back pain cynics somehow saw through the ruse.

    Not unlike my own very specific stress related reaction to neurological damage, not all stress has the same cause, and therefore one presumes does not all have the same blanket cure. Equally all back pain cannot be assumed to have the same causes- and therapies which claim to alleviate that pain presumably have to tackle different roots of that problem in different ways.

    I’d suggest the only common denominator in the back pain study is the anatomical location of the problem, though even the back itself covers a fair amount of physical real estate. The causes of that pain could be manifold and the extent of damage also would need to be assessed independently of the patient’s own assessment. I suspect though that there are no means of accurately evaluating the extent of nerve and deep tissue damage, as well as bio-chemical messaging issues. Without an accurate objective baseline, the sham acupuncture study cannot be assumed to measure anything in a scientific way.

    It is quite possible even that the direct physiological source of all causes of back pain, have yet to be identified by modern medicine. The hypocretin peptide that I suspect has been compromised in my neurological damage was only discovered in 1998.

    Since most mainstream medical treatments for back pain simply block out the perception of pain, most observers are equating two very different approaches. The comparison is with a western medical approach which isn’t actually successful in any way other than masking the pain, with another technique which could actually be curing or at least reducing the underlying problem albeit in a minority of patients.

    It seems that the power of the scientific method is such that one vaguely focused study on the ambiguous subject of back pain can over turn the practical experience accumulated in the last 3,000 years of (billions of) satisfied customers across cultures for a variety of ailments from acupuncture.

    Except that counter factual scientific evidence such as the Georgetown University study doesn’t get reported in the same press to the same degree which shows measurable evidence of the therapeutic effect of acupuncture on Rats

    As a scientific endeavour I’d be more inclined to believe the Georgetown study with its measurable levels of changes in neuro-peptide Y, rather than the approach of the “Sham acupuncture” study, which relied on purely on human testimony. That acupuncture can increase the levels of a neuropeptide associated with relieving some forms of stress, does not mean acupuncture necessarily can cures all forms of back pain. But for all we know (though I wouldn’t argue the case necessarily) there were measurable improvements in the level of tissue/nerve/pain signal damage to the back pain patients backs- but this hadn’t reached the level of human consciousness or wasn’t yet enough to appreciably limit the experience of pain. Perhaps even the experience of pain was illusory and the attention received as a result of claiming to be in pain, is enough for the patient to willingly continue the subjective experience at some level. That’s all speculative but it gives me grounds to doubt the Sham acupuncture study and does encourage me to keep an open mind about where the true fraud lies.

    I’ve also read other respondents talking about placebo affects on animals as being a possible reason for acupunctures effectiveness on larger mammals. But if you want to suggest the same thing is happening with acupuncture and rats in the Georgetown University study then you must conclude that rats are open to the placebo affect from all the multifarious drugs they have been given for many years through various clinical trials. That would be a crushing blow for biological science everywhere, best not go there….

    For a site that claims to be debunking poor scientific method, the gaps in the poorly conceived back pain acupuncture study are obvious, but yet it seems to be naively trusted in the face of wider field experience. I’d suggest what is really at play here is a kind of western cultural imperialism which assumes that anything that doesn’t come in a shiny plastic container that beeps, must be an inferior technology. That can be the only reason why one study with a limited number of patients can be deemed to be greater than 3,000 years and a billion(?) patients.

    Moreover if the average western medical surgery was able to offer satisfactory treatments or explanations for the problems many of us patients have, then we wouldn’t have to be imaginative and go outside the NHS corral. We are aware of course that this does leave us open to alternative medicine charlatans. But is this really that different from our every day experience of western medical quackery? How many patients leave appointments with their GP having undergone no further investigation and been told they are imagining it? Or that whatever ails them is caused by your age, how you are managing stress and interpersonal relationships, that you’re not eating right or not exercising enough. Equally the solution posited is that it is mind over matter, or that immediately post surgery that the operation was a complete success. That’s just bland misdirection in a lot of cases- with no attempt to establish whether there is an underlying cause or sequelae. Since studies of what goes on in a private GP surgery are rarer than hen’s teeth, whether I am wantonly exaggerating the extent of the problem or have highlighted a major issue, is debatable.

    Nor are there any guidelines for the patient to gauge how well their doctor is performing. There are no measures of GP success, either in diagnosis or treatments successfully delivered. You are expected to trust their mastery of the subject based on paper exams, perhaps completed 20 years ago, and that in the meantime they have kept abreast of modern developments in medicine since graduation. The fact that the internet contains reputedly over 200 million medical facts, and GP’s can retain an impressive 2 million, has not seemingly opened the GP surgery to Information Technology as a diagnostic tool. This is despite the fact that the rest of society moved into the 21st century 14 years ago.

    More often than not, the doctor tells the patient that nothing can be done, and that is supposed to be taken as the definitive judgement on the subject. You are instructed to get a second opinion, if you are dissatisfied, as only a last resort of redress. But should for example it be proven that a therapy was available no censure is ever brought on the original doctor for failure to disclose an appropriate therapy. That is unless of course the absence of the appropriate therapy kills. No GP in the UK gets censured for a patient rash caused by prescription medicine, though if it had been caused by a Branded detergent, the manufacturers would be financially penalized. Failure to spot Alzheimer’s 18 months on average after their German counterparts also goes unchallenged let alone unpunished.

    Assuming that most problems a doctor has to confront are not immediately life threatening, and the drugs offered to counter problems equally non life threatening, outcomes other than mortality do not get measured, and a vast swathe of medical practice goes unchallenged in its effectiveness. That’s about as unscientific an approach to any human endeavour that I can imagine. Don’t attempt to scrutinize this human activity, because it can’t be measured. I can’t imagine that would be the credo of an Isaac Newtown or an Einstein.

    Worst of all, the “Surgery Appointment” the medium in which most initial diagnoses are made, is staged to give total control to the GP. I’ve been in consultations where the doctor sat in total silence because I believe he didn’t want to spend money on me, the patient. So, do I have to take the GP to court every time financial considerations hold sway? For example we do not have a system of patient advocates who could sit in on appointments and could intercede, where we suspect foul play has taken place, but where we don’t either want our medical problems made public and we suspect some additional pressure would open the GP’s wallet.

    In conclusion, and back to the acupuncture – a much better use of scientific enquiry would be to find out how acupuncture does work, in those situations where it has been scientifically proven to be effective. Again I would want more studies than the Georgetown experience, because one study (particularly the sham acupuncture study) like one swallow- does not make a summer. Reliance on one study to dismiss 3,000 years of experience is woefully unscientific- in fact shamefully arrogant. Such a general enquiry into acupuncture could open up a new understanding of the body, if Traditional Chinese medicine is to be believed on the subject of meridians. Opening up new fields of enquiry seems a much more stimulating course of action, than dismissing wholesale a relatively successful indigenous medical paradigm. It would bring I am sure ultimately benefits for all in Big Pharma and the medical establishment, with their history of subsuming other disciplines discoveries under their own paradigm.