Ernst '11 citation

edit

please don't edit this unless you're me -- thanks


Ernst '11 (PMID 21440191; full PDF here) is a review of reviews on acupuncture for pain (and of risks). As of January 28, 2014, it is cited at acupuncture to support this claim:

  • "A systematic review of systematic reviews found that for reducing pain, real acupuncture was no better than sham acupuncture and concluded that there is little evidence that acupuncture is an effective treatment for reducing pain."

This is not a "finding" of Ernst '11; the pertinent sentence is being read out of context. It's on the last page: Real and sham acupuncture were both more effective in reducing pain than no acupuncture at all, but real acupuncture was no better than sham. In context, it's referring to a particular, high-quality trial of acupuncture. He compares that trial's results with the mixed and contradictory findings of the reviews that are the subject of his paper.

Full context of Ernst's statement about real and sham

The statement "real acupuncture was no better than sham" appears in this paragraph from the Discussion, at the top of page 762:

These findings [i.e. the findings of the older reviews Ernst examined, which actually leaned positive but were mixed and contradictory -- see. p.761, §4 Discussion] should be seen in the light of recent results from high-quality randomized controlled trials. Cherkin et al. [14] have shown that, for chronic low back pain, individualized acupuncture is not better in reducing symptoms than formula acupuncture or sham acupuncture with a toothpick that does not penetrate the skin. All 3 forms of acupuncture, however, were more effective than usual care. The authors consider, therefore, that the benefits of acupuncture ‘‘resulted from nonspecific effects such as therapist conviction, patient enthusiasm, or receiving a treatment believed to be helpful’’ [14]. This view was further strengthened by a recent randomized controlled trial in patients with osteoarthritis examining the effects of acupuncturists’ communication style [128]. Real and sham acupuncture were both more effective in reducing pain than no acupuncture at all, but real acupuncture was no better than sham. Moreover, a communication style generating high expectations in patients resulted in improved outcomes compared to a normal style, regardless of the type of acupuncture administered. In the primary studies included in the systematic reviews evaluated above, the risk of bias was often considerable. Adequately controlling for nonspecific effects in future is likely to demonstrate that acupuncture has no or few specific effects on pain [89].

In context, it should be obvious that the statement "real acupuncture was no better than sham" refers specifically to one newer, better trial: ref. 128 (Suarez-Almazor et. al., 2010; it also accurately describes the "toothpick study" by Cherkin et. al., 2009). It doesn't refer to the older reviews, with which he is comparing and contrasting the newer trials. It's not a finding with respect to the literature any more than the sentence "A communication style generating high expectations in patients resulted in improved outcomes compared to a normal style" is.

Ernst's conclusions, re efficacy, are in fact that there is "little truly convincing evidence" and that "contradictions abound". He also says that "Adequately controlling for nonspecific effects in future is likely to demonstrate that acupuncture has no or few specific effects on pain". But he doesn't conclude that the literature as a whole shows that real acu = sham acu, nor do more recent reviews AFAIK, because not all trials have found that result.