Talk:Acupuncture/Archive 22

This is an archive of past discussions about Acupuncture. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 15

←

Archive 20

No single scientific consensus on efficacy for pain or nausea

Lately I've seen some comments suggesting there is a single scientific consensus on acupuncture's efficacy overall (i.e. for any condition). That stance isn't supportable; disagreement remains on pain and nausea (but not most other conditions, where consensus is "no good evidence for efficacy" and/or "good evidence for no efficacy". (There is of course a certain range beyond which there's no disagreement; no major MEDRS argues for a large effect size, AFAIK.)

See e.g. above from JzG/Guy (diff) and at Arbcom from Kww (diff). Both assert a general consensus covering all conditions, yet neither meets the burden of evidence, and can only show the existence of a significant view (which nobody ever doubted). That's because there are multiple excellent MEDRS's that disagree with one another. For pain, Vickers' review contradicts the Ernst's recent ones, and if anything Vickers is a stronger MEDRS than Ernst. It's the strongest type of the strongest MEDRS: a meta-analysis using individualized patient data (IPD), which is the most rigorous approach, the "gold standard", a way to find information other good reviews have missed [1]. Far from being generally discredited as skeptics tend to assume (from within the bubble of the skeptical blogosphere?), Vickers was accepted by other good sources, e.g. the well-respected Medscape [2].

For more, see my comments at Arbcom (diff) responding to Kww, and at WP:AE (diff 1; (diff 2). No hard feelings toward either editor, of course; apart from this misreading of the literature, they're both highly clueful (and I hope it's obvious that I raise the issue for its own sake, not to be vexatious). Anyway, I was able to provide sources at least as strong as the sources Guy and Kww did, proving that there is >1 significant view on nausea and pain. But again: there's no discernible consensus view for either condition (and I'm not even sure there's a discernible majority view). And there will be no consensus as long as excellent MEDRS's disagree with one another. --Middle 8 (contribs • COI) 22:10, 8 January 2015 (UTC)

There are two distinctions to be made. First, the claim that it does relieve pain or nausea would require a strong consensus, as it's a remarkable claim and there is no consensus as to exactly how it could do either of those things. Second, the claim that the both views are equally supported by "excellent MEDRS's" is questionable: NCCAM's reliablity is in question, and many of the other supporting studies come from China, where there is certainly a political pressure to find positive results. The bias of Chinese studies is also supported by reliable sources.—Kww(talk) 23:42, 8 January 2015 (UTC)

I agree with Kww partly. I consider it's even more likely that there are studies about the possible publication bias concerning NCCAM (a Federal Government's agency in the U.S.); the Chinese language itself might set certain barriers when it comes to studies conducted in Chinese language. I remember some publication bias studies concerning Chinese scientific literature in the field of economics; it was research called "Meta-analysis of China’s business cycle correlation"[3], and there was also studied whether the Chinese publications had any bias concerning the research. Well, as I said, the language-barrier might be a quite restrictive one; the study was carried out, thanks to the help of a Chinese research assistant.

However, claims on medical efficiency do require MEDRS compliant secondary sources. The scientific literature will discuss the consensus, that's not something we need to speculate. Cheers! Jayaguru-Shishya (talk) 17:21, 9 January 2015 (UTC)

@Kww - re your #1, it's simple; we just weight sources properly. If we have two reviews of comparable quality (good methodology, good journal), and one finds evidence for nausea and another doesn't, we just summarize and present them, and mechanism isn't really relevant (there are lots of possibilities, none of which require invoking qi). Most of the time, in this topic area, reviews will tend to be negative, but when they're not, we don't have to reinvent the NPOV wheel.

re #2 - you're right, it all depends on the source. For pain, Ernst and Vickers are both at the highest level, so again, we just present them side-by-side. --Middle 8 (contribs • COI) 20:16, 9 January 2015 (UTC)

No, when one journal is making an an extraordinary claim without plausible explanation and the other one doesn't, we go with the one that doesn't and dismiss or downplay the one that does. That's weighting sources properly. Your method highlights false positives because, unsurprisingly, it's those false positives that people are so eager to include.—Kww(talk) 22:10, 9 January 2015 (UTC)

Agree with KWW, per WP:REDFLAG and WP:GEVAL. Extraordinary claims require extraordinary sourcing, and in this case the burden of proof is on the side making the extraordinary claim, not the scientific default of non-effective. Dominus Vobisdu (talk) 22:53, 9 January 2015 (UTC)

Kww, as much as I appreciate your enthusiasm to the article, that's certainly something we should leave for the scientific research to decide. If we have a notable source, it is not our job in Wikipedia to speculate on the quality of their peer review process. If positive results do exist, then we will include within the range of proper weight naturally. Cheers! Jayaguru-Shishya (talk) 23:36, 9 January 2015 (UTC)

As above, Jayaguru-Shishya, no. Such an approach will invariably drift towards overemphasis of false positives. That's why WP:REDFLAG and similar concepts exist. Our role as editors is to compensate for source bias.—Kww(talk) 02:02, 10 January 2015 (UTC)

Incidentally, User:Middle 8, I would consider Vicker's "Although the data indicate that acupuncture is more than a placebo, the differences between true and sham acupuncture are relatively modest, suggesting that factors in addition to the specific effects of needling are important contributors to therapeutic effects" and Novella's "the benefits of acupuncture are likely nonexistent, or at best are too small and too transient to be of any clinical significance" to be in substantial alignment, differing primarily in the value judgement of whether a trivial impact is worth paying for, not in whether the impact is trivial. Vicker's judgment continues that "Even though on average these effects are small, the clinical decision made by physicians and patients is not between true and sham acupuncture but between a referral to an acupuncturist or avoiding such a referral". i.e. the value of acupuncture including the placebo effect is enough that he considers it worth the payment. There's no controversy that if there is an actual, non-placebo based benefit derived from acupuncture, it's small. There is a scientific consensus on that point. No reputable source is claiming that there is a substantial benefit relative to placebos.—Kww(talk) 02:36, 10 January 2015 (UTC)

That is a reasonable (and intelligent) observation. Saying that we should exclude Cochrane reviews whose conclusions we don't accept is not (reasonable). The benefit (if it is real and not an artifact of bias) is small or modest (like Advil or Zofran): nobody who is reality-based disputes that. What is debated is whether it is clinically relevant, and a "yes" conclusion in a meta-analysis is not to be treated as a Fortean phenomenon. --Middle 8 (contribs • COI) 09:03, 11 January 2015 (UTC) edited21:03, 11 January 2015 (UTC)

What I've been arguing is that this giant rack of reviews does nothing but obfuscate the issue. Our section on effectiveness should be clear, because consensus is clear. Something like

The physiological benefits of acupuncture are non-existent or trivial. Most, if not all, of the benefits are derived from the placebo effect, where ineffective treatments appear to have an impact because the patient believes it will have an impact.

We add citations to both the studies that think placebos are worth paying for and those that don't, but we don't bring in text that gives the false impression that acupuncture is effective. Then we kill off this giant list of studies that hint at trivial effect because they serve no purpose but to mislead the reader. —Kww(talk) 15:19, 11 January 2015 (UTC)

Kww -- We agree that the section on efficacy badly needs pruning (and I believe we agree that the Safety section also needs pruning). We disagree on weighting the positive conclusions. Your italicized sentence pretty much reflects the consensus for most conditions, and we know this because the best reviews are in agreement. But we can't infer that this consensus fully extends to pain and nausea, because the best reviews are not fully in agreement. Reviews are what indicate consensus, or lack thereof. What else would indicate consensus? --Middle 8 (contribs • COI) 21:59, 11 January 2015 (UTC)

What review indicates that there is a substantial benefit for pain and nausea that is not accounted for by the placebo effect? Vickers and Ernst both reach the conclusion that the primary benefit is placebo. You've agreed above that anyone "based in reality" agrees on that. There's some disagreement between "non-existent" and "trivial", which I covered with "non-existent or trivial". So where's the disagreeing study?—Kww(talk) 23:11, 11 January 2015 (UTC)

Here is what Vickers says: Acupuncture is effective for the treatment of chronic pain and is therefore a reasonable referral option. Significant differences between true and sham acupuncture indicate that acupuncture is more than a placebo."[4]

Kww, I already said it: I agree with your depiction of consensus for most conditions, "But we can't infer that this consensus fully extends to pain and nausea". I am being very reasonable about where we overlap, and you appear to be IDHT-ing over where we differ.

As an ICD meta-analysis [5], Vickers is the best of the best of MEDRS's. To add to that weight, Vickers is cited by e.g. Medscape and NHS. There's no way it doesn't weigh. So can you budge a little? --Middle 8 (contribs • COI) 00:31, 12 January 2015 (UTC)

P.S. FWIW, I've crunched the numbers the way Gorski did. The minimal clinically important difference comes from both placebo and verum; placebo alone doesn't quite get you there. Which (cf. what Jytdog said on my user page [6]) is what matters to a lot of docs clinically, even though that thinking is offensive to others, including yourself [7]. I respect that fact. We still have to cover all views. --Middle 8 (contribs • COI) 00:31, 12 January 2015 (UTC)

I still don't understand your objection: Vickers states "Significant differences between true and sham acupuncture indicate that acupuncture is more than a placebo. However, these differences are relatively modest" and " Even though on average these effects are small, the clinical decision made by physicians and patients is not between true and sham acupuncture but between a referral to an acupuncturist or avoiding such a referral. The total effects of acupuncture, as experienced by the patient in routine practice, include both the specific effects associated with correct needle insertion according to acupuncture theory, nonspecific physiologic effects of needling, and nonspecific psychological (placebo) effects related to the patient's belief that treatment will be effective." He's clear: the difference between sham and real is very small, and he is considering the placebo effect while determining the ultimate value of the therapy. His data clearly shows that. Is it my use of "trivial" as a synonym for "very small" that bothers you? Perhaps "The physiological benefits of acupuncture are non-existent or small. Most, if not all, of the benefits are derived from the placebo effect, where ineffective treatments appear to have an impact because the patient believes it will have an impact." would align with the way you read the study better? I'm quite serious, here: Vickers did not say that acupuncture supplies a substantial benefit beyond placebo, he said that some benefit appears to exist, but it was small, both in absolute terms and in relative terms. "Most, if not all, of the benefits are derived from the placebo effect..." accurately summarizes Vickers and, so far as I can determine, any competent review. —Kww(talk) 01:21, 12 January 2015 (UTC)

Kww Belatedly: You keep "rounding down to zero" when plainly I keep saying that sometimes nonzero is not seen as trivial; FWIW, please take a second to consider WP:OPPONENT. Yes, I do disagree with your use of "trivial" or "very small" in the case of pain and nausea, since there remains disagreement over whether the effect size -- though small -- exceeds placebo and is clinically relevant. For example, ibuprofen has a small effect size but is still clinically relevant, and acupuncture is about 2/3 as strong as ibuprofen, according to Vickers, and is almost certainly way safer. (That 2/3 number does not include placebo effects.) (a) Pain: Vickers is clear that specific effects exceed placebo, and that both specific and nonspecific effects contribute to clinical significance. (b) Nausea: Lee (from Cochrane) is clear that P6 stimulation is as good as anti-nausea drugs; Ernst doesn't necessarily disagree, but says that entiemetic drugs themselves are of dubious value and therefore so is P6. So there is no apparent consensus for, at minimum, pain and nausea.

Bit more: (c) There's also a significant view that the studies themselves aren't conclusive because designing controls remains a problem; just see the responses to Vickers, for example. (d) And there's the view (cf. Jytdog's comments recently [8]) that sorting out specific from general effects is irrelevant as long as it helps people (Ernst mentions this view). We need to cover all of this (cf. multiple MEDRS position statements); will propose wording later. --Middle 8 (contribs • COI) 01:43, 22 January 2015 (UTC) revised 05:32, 22 January 2015 (UTC)

I'm not the one "rounding down to zero". Some sources say it's zero. Some say it's "very small". But Vicker's clearly says that the component that can be traced to any physical effect from acupuncture is small in comparison to the placebo effect. Vickers clearly does not state that the physiological component exceeds placebo, he states that there is a physiological component in excess to the placebo effect. There is no reliable authority that is claiming that the delta between the total effect of acupuncture and the placebo effect of acupuncture is large: the strongest claim is that it is a small, yet clinically relevant effect. Jytdog's position (and those that agree with him) is irrelevant to the discussion of efficacy: any placebo will be more effective if more people recommend it. Indeed, the placebo effect relies upon deceiving the patient. That's not a valid argument for exaggerating a treatment's effectiveness. I'd be happy with "Acupuncture, if effective at all, is less effective than Tylenol and far more expensive", but I would suspect people would raise WP:OR objections. Still, whatever "wording" one comes up with needs to make it clear that the effects we are discussing here are so tiny that the people that study them argue about whether the things they sense are experimental error.—Kww(talk) 05:51, 22 January 2015 (UTC)

To simplify, my objection is to the dispute between whether the effects are "zero" or simply "small" being used to obfuscate the situation. The article should not use that discussion to leave it open that "oh, since scientists can't agree on the size of the effect, maybe it's really wondrous". There is a consensus that the physiological impact of acupuncture ranges from non-existent to small.—Kww(talk) 06:00, 22 January 2015 (UTC)

Kww -- That's a helpful simplification, thanks; and imo largely a reasonable one, based on which I don't think there's that much daylight between us. Some of it will boil down to WP:WTA. We may or may not disagree on how to weight the views that (to varying degrees) studies are inconclusive due to the old chestnut that "study designs are dicey" (cf. STRICTA -- which are actually quite sensible guidelines, but do not totally invalidate research to date; far from it). Such views aren't saying it might be wondrous, just that the jury's out, and it's not time to move on. That's a view I've seen coming from academic centers. Self-serving somewhat, sure (like much else that says "let's study this some more"), but it's still part of the mainstream. --Middle 8 (contribs • COI) 11:15, 26 January 2015 (UTC) edited 18:34, 26 January 2015 (UTC)

P.S. Note that costs comparisons with NSAIDS (not the same as Tylenol btw) should factor in dosage (one acu treatment = multiple doses of NSAIDS) and the costs of adverse effects from NSAIDS (it appears to be billions of dollars). --Middle 8 (contribs • COI) 18:34, 26 January 2015 (UTC)

Couple more comments, fwiw: (A) a caveat about sham, and (B) an example of a sympathetic view (the kind I just mentioned above). (A): Not all controls are placebos. Sources have noted that sham needling isn't inert, although it's obviously a valid control for point specificity. There is some analgesia at needling sites; it's called the "needling effect" (and Brangifer might know of sources). As you probably know -- but I'll say this anyway for others' sake -- needling at or near painful regions is common practice. That kind of treatment is analogous to massaging a tender area, and is a valid and ethical way to practice, as long as reasonable claims are made. A truly inert control, one that controls only for the "theatrical" and other placebo effects, would be faux laser stimulation, which was used by Hinman's (2014) study on knee arthritis [9]. Guess what, Hinman found that real needling, real laser and faux laser were all about the same. (B) Ernst noted that Hinman provides "more evidence to show that acupuncture is a ‘theatrical placebo’". Contrast that with Peter Wayne at Harvard Med, whose view is typical of proponents; he says that it's just one smallish study, we still don't know the whole story, and it's worthwhile to refer because of the clinically significant difference from no-treatment. Those are good examples of contrasting mainstream views, and may be worth mentioning as such. --Middle 8 (contribs • COI) 18:34, 26 January 2015 (UTC)

I want to add to something Middle 8 said regarding study design being flawed. Many researchers agree on that, and in fact, there is an ongoing movement towards improving research. I'm not a scientist by any means, but what I have learned is that some scientists often try measuring acupuncture identically to how they measure drugs. But acupuncture's effects are dependent on a much larger array of variables, and researchers agree that it takes time for research science to catch up. When measuring the effects of a pharmaceutical drug, there are few variables involved. Who administered the pill? When was it taken? With food, or without? That's about it. With acupuncture, you have all sorts of different styles. Everything from Japanese acupuncture to modern trigger point therapy. The size of needles used. The type of stimulation used. Needling depth. Practitioner credentials and experience. When measuring the placebo, you also have to question if the dummy point you're using is actually a point in another family lineage system actually used to treat the condition. Sometimes it's just the act of needling itself which causes physiological response. I saw a statistic once where acupuncture was only 10 percent more effective than the control for high blood pressure, but the control itself was powerful, something around 30 percent. Even the "placebo" outperformed some BP drugs. But what we do know is that there is no consensus on acupuncture being effective or ineffective for things like pain and nausea. My insurance covers both, in fact, and if it were proven ineffective there would be no way they would pay for it. Insurance companies absolutely rely on scientific consensus to make their decisions. To my knowledge, no governmental health body has stated there is consensus in acupuncture's efficacy, no major research bodies, just a few editors on Wikipedia who interpret the literature one way. In fact, our most prestigious health bodies, like the NIH or WHO, have positive things to say about acupuncture. To say otherwise is the very definition of original research and to make edits from that standpoint, instead of relying on what the statements and research actually say, isn't good for the encyclopedia. LesVegas (talk) 18:48, 22 January 2015 (UTC)

LesVegas, scientists are divided between only between whether acupuncture has no effect beyond placebo or little effect beyond placebo. I am trying to devise a summary of the research and consensus that does not mislead our readership into believing that research has left open the possibility that acupuncture has a substantial physiological effect. It doesn't: the dispute is between "zero" and "small", and as to whether it is ethical to pay for the placebo effect. That ethical dispute is a separate discussion from the nature of any physiological effects, but it is the basis for why insurance pays for it: they care only whether it makes people feel better in proportion to the dollars expended, not whether that benefit is based in delusion or reality.—Kww(talk) 22:00, 22 January 2015 (UTC)

When acupuncture has a 40% positive affect for blood pressure, insurance does not pay for that. Why? Because the treatment group only outperforms the control by 10% so it's not considered to have efficacy. It's considered unproven and only 10% effective. Kww, by your logic, insurance would still pay for it because it helps their customers. But they don't, at least not that I'm aware, because they rely on strong scientific data. Regarding your other statement, I'm all for summarizing "research and consensus" and hope that if you do this, you will post it here. Frankly, for some time, I've been meaning to work on translating some of the research from Chinese language journals for us to add to the article and it might make for a decent addition to what you're starting. Anyway, let me know when I need to get off my duff and get to work and add some of these. There's some of the richest data over there because they know the medicine very well and have been researching it with modern research methods for a very long time. Come to think of it, I also want to add some cool sinological and historical stuff to the article too. Hopefully, I'll get more time and energy real soon. Only so many hours in the day! LesVegas (talk) 01:21, 23 January 2015 (UTC)

Outdenting, and replying to Middle 8's "clarification" ( you really shouldn't edit comments after I've replied to them: it makes it look as if I have replied to something I didn't get a chance to read): again, the study you point at as pro says "Keep in mind that the differences were quite small—like a 1-point reduction in pain on a scale of 0 to 10". Your source completely supports my statement: scientific consensus is that the effect of acupuncture ranges from none to trivial/small/"quite small"/"very small". There is no reliable source that says that acupuncture has a substantial effect, only that it may have a measurable effect. Using a study like that to try to obfuscate the fact that there is agreement that acupuncture does not have a large effect on pain and nausea is just that: obfuscation. Choose what synonym you want for "trivially small", and let's put a plain English summary of the scientific consensus in place of this long laundry list of studies: "The physiological benefits of acupuncture are non-existent or small. Most, if not all, of the benefits are derived from the placebo effect, where ineffective treatments appear to have an impact because the patient believes it will have an impact." Even your pro acupuncture studies align with that summary.—Kww(talk) 06:01, 28 January 2015 (UTC)

@Kww: Hang on -- I didn't point to any study as pro, and I'm NOT saying that a 1-point reduction on a 10 scale is big or even modest (nor that it's clinically relevant). I said that Peter Wayne's comments here are a good example of an acu-sympathetic POV: that we don't know enough yet, which doesn't square with your proposed summary. I also don't think an effect size comparable to Zofran or Advil is trivially small. (I don't believe I edited any comments after you replied, but which are you referring to?) --Middle 8 (contribs • COI) 09:03, 28 January 2015 (UTC)

I'm quoting Pendick (yes, editorial, not study) as saying the effects are small: "quite small" is a direct lift from the editorial point. Wayne is a heavily biased source: chairman of a pro-acupuncture society being quoted in a blog. And look at what he says:

“This is a small study that replicates what we already know”: Yes, that's true: we know acupuncture has little to no physiological effect on pain, and this study confirmed that. It falls right in line with the conclusions of Ernst and Vickers.
“When you compare acupuncture to no treatment, there seems to be clinically meaningful differences for many pain conditions, including back pain and knee pain.” Also not disputed by anyone: when the placebo effect is included in the total comparison, there's a clinically meaningful response.
Based on this pragmatic comparison, if I were deciding whether to send a family member or friend for a pain-related acupuncture treatment, I would say ‘yes’.” Editorializing about his personal belief about the ethics of prescribing placebos, not a statement about the physiological effects of acupuncture.

Again, there's nothing here that would contradict "The physiological benefits of acupuncture are non-existent or small. Most, if not all, of the benefits are derived from the placebo effect, where ineffective treatments appear to have an impact because the patient believes it will have an impact" as a summary. BTW, this clarification occurred after my last contribution to this thread.—Kww(talk) 14:40, 28 January 2015 (UTC)

Here's what else Wayne says, and can't be assumed to be consistent with your proposal of consensus: “I would be careful saying acupuncture doesn’t work for all pain conditions and no one should do it; we simply do not know enough yet”. I'm surprised you didn't catch this, since I've at least twice mentioned the view that we don't know enough yet (about efficacy). That view is consistent with both (a) the qualifiers that most reviews have (re: conclusions being limited because further research being needed), and (b) doubts that some scientists express about study design.

Your proposed wording is virtually identical to the invited "anti" editorial by Colquhoun and Novella. Do you see the problem with using one side of an invited pair of editorials as your source for consensus? Novella says that extant reviews are consistent with the null result and at best indicate tiny, clinically insignificant effects. Novella says we can draw definite conclusions from extant reviews. But some scientists don't buy that; they argue that the studies (on which those same reviews are based) are themselves conclusive. Surely you can see this, even if you vehemently disagree with it and believe that it obfuscates the truth. I'm not saying I much agree with that view -- on the contrary, I don't think it's very reality based -- but it's significant, and it doesn't accord with your proposed summary. The likes of Wayne weigh and are mainstream sources: Wayne is with an academic center at Harvard. Almost by definition, "quackademics" are mainstream sources, because they're academics at major universities. Or are you saying that pro-acupuncture sources aren't part of the mainstream just because they're pro-acupuncture? If so, that's putting the cart before the horse and contravenes NPOV.

BTW, I'm just not grokking your objection to the timing of my clarification; I posted at your talk page. Happy editing! --Middle 8 (contribs • COI) 16:27, 28 January 2015 (UTC)

So propose something that captures that fact that even the pro-acupuncture side is saying there is only a small effect, You are right that Novella is arguing that the effect is non-existent. Others are arguing that it exists, but concede that it is "small", "quite small", "small in relation to the placebo", or any of a number of qualifiers. The fact remains that all sources are either saying the physiological benefit is non-existent or small. That's there some argument between "none" and "small" doesn't mean that there isn't a scientific consensus that it isn't "large", "overwhelming", "gigantic", or some such. The only argument is whether the difficulty of controlling acupuncture studies is creating an illusion of effectiveness or whether there is actually some small effect. I won't go so far as to say that people that practice acupuncture are completely unreliable, but at this point, anyone proposing that acupuncture has any effect beyond "small" is getting into fringe territory, and people that head pro-acupuncture groups are certainly biased, even if they are not 100% wrong. That's why we rely on studies like Ernst and Vickers in the first place: to try to weed out flawed studies put on by biased groups.—Kww(talk) 16:49, 28 January 2015 (UTC)

The more I think about it, the more I see your statement as more of a significant view than a consensus, because it goes further than many reviews do (more below). I've already said that I don't think your statement is consistent with pain or nausea.

Pain not "small": Vickers' numbers show that specific effects (verum) range in size from 40% - 70% of the size of general effects (sham). Not bigger than sham, but hardly teeny-tiny in comparison either.
Nausea not "small": Lee & Fan [10] don't say it's small, they say it relieves PONV, and is comparable to anti-emetics. Ernst [11] appears to contradict this, which means we cite both, not just Ernst.
Reviews in general hedge: your statement is more definite than the reviews themselves, which frequently hedge by saying conclusions are uncertain because of low trial quality, blah blah (example: [12]). Your statement doesn't reflect this hedging, but rather Novella's and Colquhoun's interpretive attempt to cut through it. Do you not yet see the problem with relying so heavily on the "anti" editorial of an invited pro & anti pair? A good deal of the "obfuscation" you're complaining about in this article actually reflects the reviews themselves. The views of academics like Wayne, and Ted Kaptchuk and others -- whether optimists or pragmatists -- reflect and emphasize the "wiggle room" in reviews. I doubt if they'd argue the effect size is likely, with further research, to turn out to be huge, but I see no indication they'd go so far as to call it tiny or trivial -- and I'm sure you realize that the burden lies with you here since you're arguing for a consensus.

If reviews hedge, then we do too, even if it annoys Novella and people who agree with him. His dismissal of the hedging typical in Cochrane reviews is a sig view, but not a consensus, just as other academics' rather credulous "embrace" of that hedging is also sig view.

Are there any sources meeting WP:RS from which we can discern consensus, or speak in WP's voice? Medical textbooks? Note the discussion at WT:MEDRS on how to infer scientific consensus; some lean toward Cochrane while others lean toward position statements by expert bodies, so we should look to those as well. The NHS is an example [13]. Happy editing.... --Middle 8 (contribs • COI) 08:53, 29 January 2015 (UTC) edited11:32, 29 January 2015 (UTC)

Ernst, Vickers, and all quality reviews that we quote use "small" or a synonym to describe the actual physiological effects of acupuncture. While they hedge, it's a hedge in both directions: the studies that do the best job of simulating all the placebo effects with none of the actual practice find the least residual effect, so an ideal study is more likely to find no effect than the ones we have so far. I'm not proposing language from Novella, I'm simply providing a layman summarization of Ernst and Vickers. "Non-existent or small" covers that hedge quite neutrally, saying the the con side might be found wrong, and there is an actual effect for some inexplicable reason, while "non-existent" indicates the pro-side might be wrong and there is no actual effect at all. It specifically uses the neutral terms "placebo" and "believes", as opposed to a harsher formulation such as "works primarily by deceiving the patient into believing something useful has been done". If you wanted to add a sentence such as "Studies continue in order to quantify precisely what, if any, the effects of acupuncture are and to clarify whether the studies are returning evidence of genuine benefit or are simply returning false positives", I could live with that.—Kww(talk) 15:03, 29 January 2015 (UTC)

I think its generally best to simply quote the conclusions of the systematic reviews, there will always be disagreements about how we color them when paraphrasing or summarizing. Herbxue (talk) 16:45, 29 January 2015 (UTC)

The primary issue is to eliminate the impression that there is some significant controversy about the effectiveness of acupuncture. Listing technical jargon extracted from hundreds of studies serves only to overwhelm the reader with words that most of them don't understand. What part of my summary do you believe fails to adequately translate technical jargon into lay English?—Kww(talk) 17:22, 29 January 2015 (UTC)

The primary issue is to properly represent reliable sources. To imply that there is consensus on the question of efficacy goes beyond what the sources say. Herbxue (talk) 19:34, 29 January 2015 (UTC)

But there is consensus that the physiological effects of acupuncture are small and less than the placebo effect. Can you point to any source with the reliability of either Ernst or Vickers that disagrees, Herbxue?—Kww(talk) 22:31, 29 January 2015 (UTC)

"less than placebo"? What source says that? Vickers says "Acupuncture is effective for the treatment of chronic pain and is therefore a reasonable referral option. Significant differences between true and sham acupuncture indicate that acupuncture is more than a placebo." - it goes on to say the difference is "modest" - ok, quote that. They still concluded a. It works and b. there are specific effects above and beyond placebo. It literally says both of those things. Now, we don't use that to over-generalize ("acupuncture is proven to work for pain" for example), even if it would technically adhere to the meaning of a sentence in the source, it would be stretching the source to make a broader statement than the source makes. Our job is to faithfully report on reliable sources, not to do original research. Herbxue (talk) 15:40, 30 January 2015 (UTC)

Vickers does say that the actual effects of acupuncture are less than the placebo effect. "On average these effects are small" and "these differences are relatively modest". "These differences are relatively modest" explicitly categorizes the difference (i.e. the physiological effect) as being small relative to the effect of the placebo (sham acupuncture). Why would you have us complicate the summary by including a pile of quoted sentence fragments? Our job is to summarize, not to quote.—Kww(talk) 15:52, 30 January 2015 (UTC)

@ Kww - Answering in order of my points above:

Agree Vickers is consistent with "small" or "relatively modest" (Vickers uses both terms), as long as we're clear that verum's contribution is smaller than sham but not small relative to it. We should say sham rather than placebo because general effects aren't always the same as placebo: some kinds of sham that involve needling aren't inert. See e.g. NHS (§ "Some positive evidence", 4th para; also § "Assessing the evidence", final para).
I didn't see you respond re Lee & Fan (Cochrane) for nausea; we can't take their "equivalent to anti-nausea drugs" as meaning they also think it's equivalent to "small" or "smaller than sham". They don't (IIRC) say.
Re reviews' hedging, the issue is how that hedging is taken by scientists. Here the caveat is the same as with nausea: When scientists say we don't know, we cannot assume they'd agree with your statement. I know you (and Novella) believe that acu is not taken seriously or widely studied, but it in fact is. We know this because of the plenitude of Cochrane reviews that are always being updated, and Vickers and Ernst, and stuff like this from Stanford. I don't think it would be so widely studied if your statement were widely embraced.

Finally -- and particularly in light of the thread at WT:MEDRS about how we determine sci consensus -- I would be a lot more comfortable with your formulation if we had consensus statements from MEDRS's that said substantially the same thing (which isn't a matter of my having to prove they don't; that burden doesn't fall to me). The WT:MEDRS thread says that such statements are at least equal to systematic reviews as sources for sci consensus. Some examples: The National Institute for Health and Care Excellence (NICE) recommends a course of acu for low back pain, headache and migraine. [14][15][16] Mayo Clinic [17] enumerates multiple conditions, as does NHS. [18] (§ "Some positive evidence"). These statements are not necessarily either consistent or inconsistent with your statement; we cannot assume either. It's more than reasonable -- if anything it's necessary -- to ask whether such sources have themselves made the translation that you have (and isn't that a SYN??). I'm not seeing that they have, which is why I think your summary is probably not a statement of consensus, but rather a sig view. (Addendum: I've said before that it's entirely possible, and maybe likely, that such expert statements will eventually converge toward summaries like yours and Novella's. But global consensus is that we lag rather than lead these kinds of sources.) --Middle 8 (contribs • COI) 09:17, 31 January 2015 (UTC) edits (prior to any replies): copyedit for clarity, add parenthetical at end 18:53, 31 January 2015 (UTC). Add sources for NICE 18:58, 31 January 2015 (UTC). Minor copyedit 23:19, 31 January 2015 (UTC)

Middle 8, why do you keep bringing up irrelevant objections about people recommending acupuncture as a treatment? I'm not arguing, nor have I argued, that there's a consensus that people shouldn't recommend acupuncture for the treatment of pain. What there is a consensus for is that the strongest effect of acupuncture is due to its placebo effects, and that the physiological component, if it exists at all, is small in relation to the placebo effect. Why do you think the fact that many people are comfortable recommending a treatment that is primarily effective due to the placebo effect somehow contradicts the consensus that it is primarily a placebo? They are logically completely unrelated topics.—Kww(talk) 00:16, 1 February 2015 (UTC)

Kww, a few things: first, I notice that you are making an argument for "scientific consensus" based on just two sources: Ernst and Vickers. But we have so many more sources. Really, if scientific consensus is to ever exist on something like acupuncture, it's going to have to be a large scientific body to make a statement like that anyway. As highly as I regard everyone's opinion, I have to admit that Wikipedians arguing on an online talk page will never be qualified to make consensus statements about anything, much less something like this. One way or the other. And you especially can't do it when, out of the hundreds to thousands of good sources on pain, you are arguing for a scientific consensus opinion based on two of them, and one of those two doesn't say what you're wanting it to say so you're editorializing it and OR'ing it.

Actually, Vickers's point about the decision being to refer or not refer is a good one, and I think there's a larger point implied in it. In China, acupuncture studies never use placebo controls. They believe it is unethical to give a group of human beings no treatment at all in the name of science. So all their studies in acupuncture or herbs or whatever, are tested versus some sort of pharmaceutical or pain patch or whathaveyou. They still use controls, but the overall effect is what's most important. Placebo response in acupuncture is difficult to measure because of a high amount of variables anyway. Based on my knowledge of Chinese medical history, this is likely something the Chinese figured out a long time ago. In China and in most of Asia, the decision to treat with acupuncture is made when considering: 1) performance versus the alternative, and 2) known or potential side effects of the alternative. So you routinely see studies where acupuncture outperforms hydrocodone in both subjective pain assessment as well as in things like objective anti-inflammatory markers, and for those conditions acupuncture is recommended. In other conditions, let's say choleocystitis, an acupuncture protocol might only have a 60% cure rate so surgery is always recommended to be scheduled and acupuncture is done up until then to try obtaining a resolution. And if acupuncture has alleviated the symptoms, then ultrasound or whatever is recommended to be done, to see if the stones are actually gone before cancelling the surgery, things like that. Because they understand acupuncture and use it in modern integrative settings already, their literature is much more "real world" in that regard, and I really need to start posting some of it here. I think it's much more of a help to Wikipedia's readership to read about acupuncture's efficacy vs something else than 'acupuncture did X vs the placebo in this study and did Y vs the placebo in this other study.' Really, I can't imagine being a reader and reading this article, it's abstract, contradicting and confusing. So I see why you want to streamline all of it, but in my opinion, OR and overlooking many other sources is the wrong way to go about it. LesVegas (talk) 00:45, 2 February 2015 (UTC)

Kww: practice guidelines are "irrelevant"?!?! They're are about as good for sci consensus as sources get! See WP:MEDSCI. Your assumption that the NICE, et. al., are recommending acu for its placebo effects is just that, an assumption, and it's up to you to demonstrate it's correct, not up to others to prove otherwise. If they were recommending acu as a placebo, why don't they recommend it for a huge range of conditions? Where else on their site are they recommending placebos? --Middle 8 (contribs • COI) 02:10, 2 February 2015 (UTC)

You two are infuriating ... I am not making a single assumption, I am simply reflecting the unpleasant reality of what the sources you insert into the article say: the medical benefits of acupuncture are small relative to its placebo effect. As for Chinese studies, LesVegas, I'm well aware that they don't follow standard practice and don't use reliable controls. That's why they aren't particularly relevant and are frequently disregarded by other medical journals..—Kww(talk) 02:24, 2 February 2015 (UTC)

To continue, Middle 8, I don't have any objection to a sentence that indicates that it is frequently prescribed and funded, so long as that is not used to pretend that it has any substantial effect.—Kww(talk) 02:27, 2 February 2015 (UTC)

@ Kww -- That's a good addition, but the fundamental problem with sci consensus remains. If you find it "infuriating" that I argue from practice guidelines (of all things), then consider WP:TIGER. I think it applies, given your comments at ArbCom case requests, where you made the surprising and extreme assertion that acupuncture's effectiveness isn't under wide and serious study. But of course it is, which means no single consensus exists in some areas, efficacy for a few conditions being one of them. That's a very mild position, only an extreme one if you believe (as you obviously do) that Colquhoun and Novella's "anti" editorial coincides with sci consensus.

Sci consensus is shown not only in reviews, but in how reviews are taken by scientists (hence practice guidelines), and that's what you're missing. If you can prove that the NICE et. al. are recommending a treatment that they believe is primarily a placebo, then yes, we can be sure they're consistent with your summary. Not otherwise. I'm pretty sure the NICE doesn't recommend placabos, but who cares what I think? You're the one making the generalization, so it's about what you can prove. --Middle 8 (contribs • COI) 07:20, 2 February 2015 (UTC)

P.S.: Correction: Above I mentioned a thread at WT:MEDRS that proposed a revision to MEDSCI; however, my comments in this thread rely on MEDSCI in its present form. Still it's an interesting thread that underscores how important practice guidelines are. --Middle 8 (contribs • COI) 07:20, 2 February 2015 (UTC)

Practice and science are unrelated topics: it is you that are committing original research by conflating the two, especially given the explicit statements by the sources that are actually attempting to analyze actual effects vs. placebo effects (which NICE does not make any explicit statements about).—Kww(talk) 13:51, 2 February 2015 (UTC)

"Practice and science are unrelated topics." WOW. Yes, that's why there's no such thing as evidence-based medicine. Come to think of it, I've heard that before: from Randy in Boise. And since there's no EBM, there's no Cochrane, and no widespread, serious study of acupuncture. Yes, it all makes sense now. No wonder editors who argue otherwise need to be topic-banned: this will "help provide an environment that will allow our scientifically-minded editors to prevail". --Middle 8 (contribs • COI) 23:34, 2 February 2015 (UTC)

Don't accuse me of incompetence. Practice is a social issue, one that combines a number of factors. "Unrelated" was strong, but there's no cause and effect: the fact that something is practiced doesn't permit the conclusion that there's a scientific foundation, nor does something having a sound scientific foundation necessarily lead to it being practiced. Evidence of one cannot lead to the conclusion that the other is true. Vickers explicitly states that acupuncture has a strong placebo component, which leads to his recommendation that it be practiced. That in no way refutes the concept that it is primarily a placebo. You are conflating two separate concepts. "It's widely practice" in no way invalidates the summary of ""The physiological benefits of acupuncture are non-existent or small. Most, if not all, of the benefits are derived from the placebo effect": that accurately summarizes Ersnt, Vickers, and you have yet to provide a source that attempts to quantify the proportion that refutes it. You simply assert that these bodies don't recommend any procedure that has a strong placebo component while providing no evidence to support that assertion beyond you being "pretty sure that's true".—Kww(talk) 23:52, 2 February 2015 (UTC)

Evidence → practice guidelines (that's what EBM is about) and practice guidelines are indeed sci-consensus sources (MEDSCI). --Middle 8 (contribs • COI) 03:05, 3 February 2015 (UTC)

Any evidence that the practice guidelines assiduously reject all placebo effects before making their decisions? That they reject Vickers's concept of evaluating the total effect, and proceed as Novella would, looking only at the delta between the placebo treatment and the total effect? That's a pretty strong claim, and its pretty hard to see how any competent authority could recommend acupuncture on that basis. It would seem that only Vickers's approach of evaluating the total effect could ever lead to endorsing acupuncture.—Kww(talk) 13:59, 3 February 2015 (UTC)

Let me ask you a question, Kww. What can you tell me about acupuncture placebo? How many more variables do you think acupuncture placebos contend with versus, say, the sugar pills our readers think about when seeing the word "placebo"? The Chinese don't even use it in their research. You think that makes them less of scientists? Wow, I really can't believe what I'm hearing from you. You said earlier that this makes them irrelevant and disregarded by other medical journals. However, some other medical journals simply have publishing policies which pre-exclude any non-placebo controlled studies from publication because that's the ethical and scientific standard adopted by the Occident only. We could get into a long debate about that, but what matters is that they do use controls, and that's more relevant for clinical settings anyway. If acupuncture outperforms hydrocodone for lumbar pain relief, I suppose, by your rationale, that makes hydrocodone "worse than a placebo". I say stick to the sources and let them speak for themselves. But if we are to do any editorializing, it needs to be away from making judgements about whether or not something is equivalent to a placebo (since half the world's scientists disregard that anyway) and more towards its clinical practice. LesVegas (talk) 18:26, 3 February 2015 (UTC)

I am sticking to the sources, including the widespread skepticism about Chinese investigations into aspects of Chinese Traditional Medicine.—Kww(talk) 19:38, 3 February 2015 (UTC)

Listen, Kww, you're saying you stick to the sources, yet just a few weeks ago, you said We shouldn't report it until multiple studies report the same conclusion I agree with Middle 8 that WP:Tiger describes this situation perfectly. It seems to me that you are a well-meaning editor who wants to help here, but you do perceive your biases as neutral POV. I'm starting to believe that my own personal editing philosophy is to AGF until there is no other possibility but to assume bad faith. But when you called for me and other editors to be topic banned, I still made an effort to AGF, but I could only assume good faith in the worst way: that you are overcome with such strong biases that you assume anyone arguing with you to be violating NPOV. Of course, you received no support at ArbCom, and many protested against you, but that didn't seem to entice you to stop and reflect for a moment. That's the real source of the conflict, and until that's resolved, I believe that arguing along these lines will be fruitless and talk page arguments will extend ad infinitum. Seriously, until you stop to consider how your biases might be getting in your way I don't think we will make any progress at improving the status of this article. LesVegas (talk) 23:20, 3 February 2015 (UTC)

I perceive my bias as the scientific POV, which is the only one that matters in relation to this particular discussion. If I was trying to alter the article to say that all acupuncturists and herbologists should be jailed for fraud, or anything like that, your comment would have merit, as that's an an issue where other POVs do have weight. At this point, all studies that address the issue of the proportion of acupuncture's effect that is due to placebo effects and the proportion that is due to an actual physiological effect come to the conclusion that the physiological effects are small in relation to the placebo effect, so small that there is widespread argument that they don't actually exist. Even Vickers, the most pro-acupuncture general review anyone can point at, agrees that the non-placebo effects are small. You have yet to demonstrate any review that comes to a different conclusion, supplying only a vague handwave at studies that don't meet any accepted standard for controls. Can you demonstrate any actual controversy with my proposed summary in terms of legitimate reviews of placebo-controlled medical studies that are of a higher quality than (or even equivalent to) Ernst and Vickers?—Kww(talk) 01:32, 4 February 2015 (UTC)

I perceive my bias as the scientific POV, which is the only one that matters in relation to this particular discussion.

That is not a scientific POV, but a pseudoscientific POV.

If I was trying to alter the article to say that all acupuncturists and herbologists should be jailed for fraud, or anything like that, your comment would have merit, as that's an an issue where other POVs do have weight.

If you want to send a large proportion of practicing physicians to jail, go ahead. Just don't complaint if you ever fall ill.

At this point, all studies that address the issue of the proportion of acupuncture's effect that is due to placebo effects and the proportion that is due to an actual physiological effect come to the conclusion that the physiological effects are small in relation to the placebo effect, so small that there is widespread argument that they don't actually exist.

Regarding the the physiological effects of acupuncture, the American Academy of Otolaryngology – Head and Neck Surgery says: "The mechanism of action of acupuncture in the treatment of AR is unknown. Studies suggest that acupuncture inhibits cytokine synthesis, such as interleukin-10 in patients with AR and interleukin-6 and interleukin-10 in patients with asthma; however, it remains unclear whether these findings correlate with clinical effect." (http://oto.sagepub.com/content/152/1_suppl/S1.full)

There is no mention of the placebo effect whatsoever. If the placebo effect is indeed relevant in the treatment of allergic rhinitis using acupuncture, surely it would have deserved a brief mention?

You have yet to demonstrate any review that comes to a different conclusion, supplying only a vague handwave at studies that don't meet any accepted standard for controls. Can you demonstrate any actual controversy with my proposed summary in terms of legitimate reviews of placebo-controlled medical studies that are of a higher quality than (or even equivalent to) Ernst and Vickers?

See PMID 20359961 and PMID 20070551

-A1candidate 14:26, 4 February 2015 (UTC)

Thank you A1Candidate for those excellent points and sources. I particularly think the Hopton/Macpherson source makes an excellent point: if acupuncture is more effective than placebo shouldn't the research towards shifting research priorities away from asking placebo-related questions and towards asking more practical questions about whether the overall benefit is clinically meaningful and cost-effective? Perhaps we Wikipedia editors should be doing the same? LesVegas (talk) 20:34, 4 February 2015 (UTC)

If "acupuncture is more effective than placebo", there would be studies showing that. There aren't. Acupuncture has been shown to be effective, but only in non-placebo-controlled studies. — Arthur Rubin (talk) 02:00, 5 February 2015 (UTC)

That overstates things, Arthur Rubin. Reliable studies have found a delta. There are reasonable arguments that they are statistical artifacts or that they are the result of small study flaws, but they are there.—Kww(talk) 02:29, 5 February 2015 (UTC)

Perhaps. But were the studies triple-blinded (in addition to the patient and the researcher, the the person inserting the needles should not know whether it's "real" acupuncture or "sham" acupuncture)? If not, there's still a significant uncontrolled placebo effect. I haven't noticed any such remarks. But I haven't read all the literature, even that available without cost (or through JSTOR).) — Arthur Rubin (talk) 02:36, 5 February 2015 (UTC)

Regarding See PMID 20359961, it confirms my summary, User:A1candidate: "In general, effect sizes (standardized mean differences) were found to be relatively small." As for PMID 20070551 the abstracts do not reveal any finding of the relative magnitude of the response vs. the placebo controls, only that the total is larger than placebo (something that I am not disputing). Can you share a statement out of the study that asserts that assessed separately, the physiological benefit of acupuncture isn't smaller than the placebo effects?—Kww(talk) 02:10, 5 February 2015 (UTC)

Regarding See PMID 20359961, it confirms my summary, User:A1candidate: "In general, effect sizes (standardized mean differences) were found to be relatively small."

This quote is not taken from PMID 20359961 but PMID 20070551. The authors summarized the effect sizes in the abstract as "relatively small" (which is not the same as "small"). A closer look at the full text would reveal what relatively small actually means:

"The collated results indicate that in the short term, acupuncture provided statistically significant effective pain relief compared with sham controls in low back pain, chronic osteoarthritis of the knee (with the caveat that this holds provided outcomes were measured after treatment was completed—one inconclusive outcome was based on a 4-week time point, well before trial treatments ended), and headache. These differences remained statistically significant in the longer term at 6 to 12 months-, for knee pain and chronic headache (with the caveat that outcomes were based on completed treatments—one inconclusive outcome was before the end of treatment). The differences between acupuncture and sham for low back pain in the longer term were inconsistent; one of these reviews found a statistically significant effect of pain relief maintained at 6 to 12 months,16 while another was inconclusive. Overall, the effect sizes were small to moderate"

As for PMID 20070551 the abstracts do not reveal any finding of the relative magnitude of the response vs. the placebo controls, only that the total is larger than placebo (something that I am not disputing).

Have you mistakenly mixed up the papers? PMID 20070551 clearly discuses the magnitude of the response vs. the placebo (sham) controls in the abstract.

-A1candidate 09:21, 5 February 2015 (UTC)

A1candidate, it states that there is a significant effect difference. There's nothing about the relative sizes of the effect in the abstract, and that is what this enormous discussion has been about. There isn't consensus that the physiological effects of acupuncture are insignificant, but there is a consensus that the effects of acupuncture that are not attributable to placebo effects are small. There's a big difference between being statistically significant and being large.—Kww(talk) 14:03, 5 February 2015 (UTC)

(e/c) @ Kww: I first mentioned expert statements on Jan. 31 [19]: These statements are not necessarily either consistent or inconsistent with your statement; we cannot assume either. It's more than reasonable -- if anything it's necessary -- to ask whether such sources have themselves made the translation that you have (and isn't that a SYN??). Five replies later, you're still trying to shift the burden.[20][21][22][23][24]

My point, cf. WP:MEDSCI, remains: for sci consensus, we look not only to reviews, but to how they are taken. You need to produce evidence that the practice guidelines (and all the other sig POV's mentioned) specifically agree with you: silence doesn't necessarily mean agreement.

Sounds like we're headed for an RfC/A or something like that. I'm pretty busy for the next month. Whoever handles this, let's keep the wording neutral, please. (How not to do it: [25][26])

And quite apart from everything I just wrote about practice guidelines, you still haven't answered re nausea. It doesn't appear Lee and Fan's review found the same as Vickers did in terms of effect size (w.r.t. sham). They found P6 stimulation as effective as anti-nausea drugs. Ernst later said neither P6 nor drugs were clinically effective, but that's a sig view, not a consensus. Any reply on that?

And there's the meta-level problem: your plain-language proposal appears to be the first of its kind. We should be working from plain-language summaries from consensus-level sources: following sources, IOW, not leading them. --Middle 8 (contribs • COI) 11:11, 5 February 2015 (UTC)

Middle 8, the problem is that you keep pointing at things that do not address the issue at hand, say "these don't address the issue", and then think it misses something. Every source that compares the size of the placebo effect to the difference between the total effect and the placebo effect says that delta is small. Some say is it small but significant, some say it is small but insignificant, some say it is so small as to be imaginary, but they all say small. That some study authors said nothing means just that: they said nothing. Practice guidelines take the totality of effect into account: placebo and actual together, and the claim that they completely exclude placebo effects in their analysis is the extraordinary one, because virtually all medical treatments include some placebo component: that's why we have to have blind trials. Silence means just that: silence.—Kww(talk) 14:03, 5 February 2015 (UTC)

As for the RFC, probably about 10 days from now: I have vacation looming. Perhaps an RFC will attract some discussion from editors without profound conflicts of interest and we can get somewhere.—Kww(talk) 14:18, 5 February 2015 (UTC)

A1candidate, it states that there is a significant effect difference. There's nothing about the relative sizes of the effect in the abstract, and that is what this enormous discussion has been about.

If you don't have access to the full text, ask for it, please. The full text says "Overall, the effect sizes [between acupuncture vs. sham] were small to moderate".

Every source that compares the size of the placebo effect to the difference between the total effect and the placebo effect says that delta is small. Some say is it small but significant, some say it is small but insignificant, some say it is so small as to be imaginary, but they all say small.

Read PMID 20359961 again. Since you don't have access to the full text, I'll quote the part that is relevant here:

"In a recent 11C-carfentanil PET study with fibromyalgia patients (Harris et al., 2009), acupuncture therapy but not sham acupuncture (at non-acupuncture points) elicited significant activation of mu-opioid receptor binding capacity in typical areas of the “pain matrix”, the cingulate, the caudate, the thalamus and the amygdala both short-term (after one session) as well as long-term (after 4 weeks) while with sham acupuncture, small deactivations of this matrix was noted, an effect that has been seen also with placebo analgesia (Zubieta et al., 2005)"

The differences are found in the relative magnitude and nature (activation vs. deactivation) of the effects.

That some study authors said nothing means just that: they said nothing.

No, it means you didn't read the paper and you shouldn't pretend that you've read them.

-A1candidate 14:39, 5 February 2015 (UTC)

A1candidate, I was upfront that I only had access to the abstracts, stated I couldn't find what you were claiming in the abstract, and then asked you for a statement from the full text. I don't understand why you think I "pretended" anything. OK, small to moderate: pretty wishy-washy, still covered by "small" in my judgement. Your second study appears to state that there is one effect triggered only by actual acupuncture in fibromyalgia patients (interesting study group that, as people don't seem to be able to gain consensus as to whether it is a physical or mental disorder). Does it make any assertion as to the amount of overall relief this effect provides? Most of these studies ultimately include some review where patients rate their pain before and after various treatments, and the acupuncture group reports a slightly larger relief than the control. Remember, I am not trying to claim that there is a consensus that acupuncture has no actual effect, because that certainly isn't true. What I'm not saying is that the reviews all agree that the effect beyond placebo is small.—Kww(talk) 15:02, 5 February 2015 (UTC)

Small to moderate means just that: Small to moderate. The second study (Harris et al., 2009) cited by the review says "Treatment differences were attributable largely to increases in MOR BP [mu-opioid receptor binding potential] following TA [traditional acupuncture] whereas SA [sham acupuncture = placebo ] evoked either a small decrease in MOR BP or resulted in no change". (PMID 19501658) According to Harris et al., patients rated their pain using the short version of the McGill Pain Questionnaire. -A1candidate 15:54, 5 February 2015 (UTC)

I'll have to take that as "no, the study made no statement whatsoever about the relative magnitude of the treatement differences. That surprises me, both because it seems illogical for a study not to reach a conclusion about that and because I can't see why you would think it was relative to the discussion without it, but I've directly asked multiple times and you haven't provided a statement as to relative magnitude from the study.—Kww(talk) 18:40, 5 February 2015 (UTC)

The magnitude is not significant ("reduction in clinical pain was similar between groups") because the placebo controls also worked, albeit in the opposite manner and resulted in a reduction (rather than an increase) in MOP binding potential. I don't strongly object to describing the overall magnitude difference as small to moderate, but the significant and opposing biological effects of true vs. sham acupuncture would of course have to be highlighted separately. -A1candidate 19:02, 5 February 2015 (UTC)

Kww, I have notified you numerous times: in Wikipedia, we do not care about "what you think", we care about the sources! Still you are pushing your own opinion instead of sticking to source. If a source says "small to moderate", then small is small and moderate is moderate, no matter what personal opinions you might have. I hope that from now on you stick to the results and conclusions mentioned in the source, and not make your own conclusions. Should you not have the whole article? Please ask an editor who has an access to it. I am sure they are more than willing to help. Cheers! Jayaguru-Shishya (talk) 23:04, 5 February 2015 (UTC)

Jayaguru-Shishya, I am attempting to summarize the sources. That's what editors do. The chief problem with this article is that it is a giant laundry list of disjointed quotes. So far, the summary of "The physiological benefits of acupuncture are non-existent or small. Most, if not all, of the benefits are derived from the placebo effect, where ineffective treatments appear to have an impact because the patient believes it will have an impact" apppears to summarize the sources quite well, as this study found that the magnitude of relief from placebo and actual acupuncture were not significantly different. Their effort to establish a mechanism may be worthy of separate mention.—Kww(talk) 23:21, 5 February 2015 (UTC)

You'll also have to forgive me that if we were discussing a five-year-old individual study that had concluded that acupuncture was completely ineffective, the discussion would be all about how we couldn't trust it because later superior reviews had contradicted it. In that sense, 19501658 doesn't even begin to meet my request of being a review of equivalent quality to Ernst or Vickers.—Kww(talk) 23:27, 5 February 2015 (UTC)

The review article that cited PMID 19501658 is PMID 20359961, which falls within the last 5 years and therefore satisfies WP:MEDDATE. -A1candidate 23:46, 5 February 2015 (UTC)

Maybe I've misunderstood, but PMID 19501658 (Harris) is a WP:PRIMARY source and therefore you can forget about using it's findings in this article. PMID 20359961 (Enck) is a review that refers to the Harris paper. Now, I don't have full access to that review, but it seems from the above discussion that certain editors here are arguing that the findings of Harris et al's experiment should be included because they appear in a review (if I have misinterpreted anyone's intention then I apologise). This seems rather clear cut to me. Harris cannot be cited directly here as it is a primary source. Similarly, what Enck et al say about the Harris study also cannot be used as a way to sneak the Harris findings in. Enck's conclusions could be included, but they seem (from the abstract at least) to be a bit more-research-is-needed-y, so what would be the point? Currently, this reference (Enck) is used to support the following text in the article: "A 2010 review found real acupuncture and sham acupuncture produce similar improvements, which can only be accepted as evidence against the efficacy of acupuncture.[79] The same review found limited evidence that real acupuncture and sham acupuncture appear to produce biological differences despite similar effects.[79]" Now, of course sticking a real needle in someone vs. sticking a fake needle in someone will probably "produce biological differences", the primary one being a hole in their skin! This doesn't say anything about efficacy as a treatment and the section in which this citation appears is called Effectiveness not Mechanism or How placebo trials should be carried out or Biological differences between real and sham acupuncture, etc. The reference to producing "biological differences" seems to be a slightly mischievous way to say "acupuncture does something". Get rid. Famous _dog (c) 08:05, 6 February 2015 (UTC)

Those articles were brought up to prevent an editor from misinterpreting medical literature, not to include any of them in the text. I thought this was fairly obvious from the above discussion. -A1candidate 13:25, 6 February 2015 (UTC)

No, it is not obvious in the slightest. The preceding discussion is extremely long and confused/ing. I think everybody needs to remember that this talk page is for discussing edits to the article and is WP:NOTAFORUM. Famous _dog (c) 14:04, 6 February 2015 (UTC)

To make it more obvious then, Famousdog, I believe that our long list of citations to individual studies confuses the reader into believing that acupuncture has more effect than it does, kind of a "Wow! If it gets studied so much and so many people use so many big words, it must do something!" effect. It also suffers from source bias, as our tendency to include every study with any slightly positive result distorts our perspective. I would like to replace our "Effectiveness" section with the statement "The physiological benefits of acupuncture are non-existent or small. Most, if not all, of the benefits are derived from the placebo effect, where ineffective treatments appear to have an impact because the patient believes it will have an impact" and cite that to Ernst and Vicker, being the latest and most thoroughly cited reviews. It's an accurate summary: while there is no consensus as to whether the effects of acupuncture exist, there is a consensus that if they do exist, they are small relative to the placebo effects of the treatment.—Kww(talk) 14:16, 6 February 2015 (UTC)