and Medical Research Council (NHMRC) levels of evidence (I-IV) which define the source of the evidence, with Level I being ‘evidence obtained from a systematic review of all relevant randomised controlled trials’, and Level II being ‘evidence obtained from at least one properly designed randomised controlled trial’ (7). By this definition, this review examined mainly Level I evidence as systematic reviews and meta-analyses were prioritised. Level II evidence from individual randomised controlled trials has been included occasionally where new high quality randomised trials may have changed the conclusions from the most recent systematic review. Risk of bias in randomised controlled trials has mainly been assessed by the included systematic reviews using the Cochrane GRADE system of low, unclear or high risk of bias across a number of domains (8). The quality of evidence has also been assessed principally using the GRADE system with randomised controlled trials being assessed as high, moderate, low or very low quality evidence. See the Appendix, Table 10 for an outline of GRADE levels. Where systematic reviews have used other systems of quality assessment (such as PRISMA or Jadad) an attempt has been made to convert these scores to an equivalent within the GRADE system. Some systematic reviews have not reported an assessment of quality of evidence of included trials, and due to time constraints, this review has not attempted to make such an assessment. In the US Veterans Affairs evidence map, evidence has been assessed at four levels: evidence of positive effect, evidence of potential positive effect, unclear evidence and evidence of no effect. In this review this terminology has largely been adopted with the exception that ‘evidence of no effect’ has been replaced with ‘no evidence of effect’ which seems to be more strictly accurate. ‘Unclear evidence’ was described as conflicting evidence between reviews or between authors within a review, with reviewers summarising the evidence as inconclusive. ‘Evidence of potential positive effect’ refers to reviews reporting all individual RCTs or pooled effects across RCTs as positive, however the reviewers deeming the evidence insufficient to draw firm conclusions. ‘Evidence of positive effect’ refers to reviews with consistent statistically significant positive effects and where authors have recommended the intervention (6). The Acupuncture Evidence Project © AACMA 3 Table 1. Summary of levels of evidence used in this review Level Description GRADE level (8) Evidence of positive effect Reviews with consistent statistically significant positive effects and where authors have recommended the intervention. Strong positive evidence. Moderate or high quality Evidence of potential positive effect Reviews reporting all individual RCTs or pooled effects across RCTs as positive, but the reviewers deeming the evidence insufficient to draw firm conclusions. Moderate positive evidence. Moderate or high quality Unclear/insufficient evidence Reviews consisted mostly of weak positive evidence or conflicting evidence between reviews or between authors within a review, with reviewers summarising the evidence as inconclusive. Low or very low quality; or conflicting levels of evidence within or between reviews No evidence of effect Reviews have consistently found little support for acupuncture. Consistently low or very low quality There is not necessarily an exact equivalence between the ‘insufficient evidence’ category used in the Australian DVA review and ‘unclear’ in the USVA evidence map, as some conditions rated as ‘insufficient evidence’ in the former may appear in the ‘evidence of potential positive effect’ category in the latter. Generally speaking, conditions rated as positive or potential positive in this review are conditions for which evidence levels were either moderate or high quality evidence according to the GRADE definitions, and conditions with low or very low quality evidence were rated as unclear even when all included RCTs reported positive outcomes. Conditions previously rated as ‘potential positive’ have been down-graded to ‘no evidence of effect’ where subsequent Cochrane reviews have found little support for acupuncture. 1.3 Efficacy versus effectiveness Efficacy refers to the ability of an intervention to achieve its intended effect in ideal conditions, i.e. a clinical trial, usually with a placebo control. Effectiveness refers to the result of an intervention in real world clinical practice. The most common method used to assess efficacy is the randomised sham/placebo controlled blinded clinical trial. This methodology is derived from pharmaceutical research where an inert tablet (‘a sugar pill’) is compared to a medication. ‘Efficacy’ in this example is the measure of the effects of the medication on one group of test subjects minus the effects from the sugar pill on the other group of test subjects. The validity of this method of measuring efficacy rests heavily on the assumption that the sham protocol comparator intervention is inert. If the comparator intervention is not inert, this creates difficulties in measuring efficacy accurately, as discussed in the next section. 1.4 Problems with placebo controls in acupuncture trials It has been observed from the over 8,000 randomised controlled trials which appear in the Cochrane Database of Controlled Clinical Trials that, in