loading  Checking for direct PDF access through Ovid


We appreciate Dr Puljak's interest2 in our recent article published in PAIN.1 Dr Puljak points out that we did not provide specific details on several methodologic points, such as the definition of high-quality trials and the assessment of included trials. This is a secondary analysis of an individual patient data meta-analysis which has been published,3 along with a protocol.4 We did not provide full details of all methodologic procedures because it would be duplicative to do so.
We would like to assure Dr Puljak that we support the Cochrane approach. Indeed, the senior author of our article, Dr Vickers, helped start the Cochrane field for Complementary Medicine in 1994 and has coauthored several Cochrane reviews on acupuncture. A careful comparison of our methodology with that recommended by the Cochrane collaboration would demonstrate that the 2 are highly congruent. For example, our assessment of randomization corresponds to the Cochrane tool's selection bias domains. We required unambiguously adequate allocation concealment, with additional details requested from authors if sufficient details on the full process of generating and concealing randomization were not provided. In addition, we addressed performance bias in our assessment of blinding for each trial and reported on the risk of detection bias in our results. We did not address the issue of incomplete outcome data when selecting eligible trials, given that we had individual patient data that could be used to investigate bias empirically. We also investigated additional sources of bias, such as simulations to investigate the possible effects of publication bias.
In sum, our methodology follows contemporary approaches for determining the risk of bias. Naturally, any reader who believes that we have erroneously included a specific trial at high risk of bias is welcome to bring that to our attention. We note that Dr Puljak does not provide such details.
    loading  Loading Related Articles