Acupuncture is effective for chronic knee pain: A reanalysis of the Australian acupuncture trial


Health Promotion

Document Type


Publication Source

Alternative Therapies in Health and Medicine

Publication Date






First Page


Last Page



© 2016, InnoVision Communications. All rights reserved. Context • In the October 2014 issue of the Journal of the American Medical Association (JAMA), Hinman et al published the results of an Australian clinical trial on acupuncture in a paper entitled “Acupuncture for Chronic Knee Pain: A Randomized Clinical Trial” (JAMA report), in which they concluded that neither acupuncture nor laser acupuncture had any greater effects than sham laser acupuncture for pain or function for patients aged 50 y and older with moderate-to-severe knee pain. That study has been criticized extensively by international scholars for its validity because serious methodological flaws existed throughout the study’s design, implementation, and conclusions. Objective • The current study intended to re-examine the prior study’s conclusions about the efficacy of acupuncture for chronic knee pain. Design • The current research team performed a reanalysis of relevant data from the JAMA report. Intervention • The original study included 4 groups: (1) an acupuncture group, which received needle acupuncture, inferred by the current authors to have been set up to be a positive control in the original study; (2) a laser acupuncture group, which received laser acupuncture; (3) a sham laser acupuncture group, which received sham laser acupuncture and acted as the negative controls for the laser acupuncture intervention; and (4) a control group, which received conventional care but no acupuncture or laser treatments. The study lasted 12 wk. Outcome Measures • The measures included evaluations in the following areas: (1) poststudy modifications—an evaluation of the consistency of the JAMA report with the study’s intentions as identified for a grant that was originally approved and funded by the Australian National Health and Medical Research Council (NHMRC) in 2009, as indicated in the study’s trial registration, and as compared with the published protocols and to the study’s originally stated objectives; (2) high heterogeneity—an assessment of the heterogeneity among the 4 groups for the overall outcome related to pain; (3) ineffectiveness of laser acupuncture—an analysis of laser acupuncture’s efficacy for chronic knee pain as stated in the JAMA report, using effect size (ES); (4) effectiveness of acupuncture—a reanalysis of acupuncture’s efficacy for chronic knee pain in comparison with the original analysis in the JAMA report, using ES; and (5) acupuncture after data adjustment—a new analysis of acupuncture’s efficacy for chronic knee pain using data from the original study that was discussed in the JAMA report, using ES, with an estimation after data adjustment and elimination of the dilution effect of the Zelen design. Results • Contrary to a general impression that acupuncture was the focus, laser acupuncture was the primary intervention tested in the actual study, “Laser Acupuncture in Patients With Chronic Knee Pain: A Randomized, Placebo Controlled Trial.” The study discussed in the JAMA report was neither a truly randomized, controlled trial (RCT) for acupuncture nor was it an appropriately designed, randomized study in general. High heterogeneity was found among its groups in the evaluation of overall pain in patients. Both the ES of 0.60 that had been set by Hinman et al for the minimal clinically important difference (MCID) and the resulting interpretation of results in the JAMA report were not appropriate. Using the original study’s criteria of efficacy, the reanalysis has confirmed that the laser acupuncture was not effective, whereas the acupuncture was found to be moderately effective for chronic knee pain (P <.05) for both overall pain and function at 12 wk, with an ES of 0.58, or after the adjustment of the data, with an ES of 0.67. Conclusions • The JAMA study was neither a conventional RCT nor an appropriately randomized trial, and its results are probably invalid. The ES of 0.60 for the MCID that was used in the JAMA study and the resulting explanation were not appropriate. Even with an ES of 0.60 for the MCID, acupuncture remained effective after data adjustment. Consequently, compared with conventional care, acupuncture treatment was found to be moderately effective for chronic knee pain in patients aged 50 y and older.

This document is currently not available here.