To the Editor: I read with interest the systematic review of Posadzki and Ernst regarding randomized controlled trials of osteopathy for musculoskeletal pain (MSP) patients [1]. The authors are correct in asserting that there is no compelling evidence for the effectiveness of osteopathic manipulative treatment (OMT) in managing musculoskeletal pain generally. However, a valid assessment of OMT in the management of patients with low back pain (LBP) was precluded by several methodological shortcomings of their review. Consequently, their results were discrepant from those of our previous systematic review and meta-analysis (SRMA) [2]. Posadzki and Ernst claim that our SRMA is burdened with a high risk of bias for two reasons: (1) four of the six included studies showed no difference between OMT and control treatments, and (2) it lacked a critical assessment of the methodology and validity of the included studies. With regard to the first criticism, the North Texas trial found significant differences between OMT and the nointervention control treatment at all three reported followup intervals [(Fig. 3) 3]. Other studies in our SRMA also found significant differences between OMT and the control treatments at one or more time intervals [4, 5]. Thus, the complexity of such clinical trials is too great to be adequately captured with a simple binary measure. For this reason, the vote-counting method invoked in this criticism has been largely abandoned since the earliest days of metaanalysis because it yields an unacceptably large risk of drawing erroneous conclusions [6]. Alternatively, metaanalysis is most useful in summarizing research studies that are too small to yield valid conclusions individually [7]. The second criticism, regarding the methodology and validity of included studies in our SRMA, can be dismissed by simply accepting the Jadad scores reported by Posadzki and Ernst in Table 1. Four of our six SRMA studies (references 9, 13, 22, and 24 in Table 1) were assigned Jadad scores of 3, 4, 4, and 2, respectively (mean Jadad score, 3.25). By contrast, the 16 included studies in Table 1 of the Posadzki and Ernst review achieved a mean Jadad score of only 2.63. Another shortcoming of their review was that they did not include the two remaining studies of OMT for LBP, which were previously identified in our SRMA [4, 8]. Even under adverse assumptions regarding the quality of these two additional studies [4, 8], the resulting mean Jadad score for our six SRMA studies would be comparable or superior to the mean Jadad score for the 16 included studies in Table 1 of the Posadzki and Ernst review. Finally, their review states, “to date no definitive conclusions about the effectiveness of osteopathy in the management ofMSP has [sic] been drawn, and the usefulness of OMT in treating MSP is unclear.” However, our SRMA [2] is now the basis for an evidence-based clinical practice guideline that promotes the use of OMT in patients with LBP [9]. Further, this guideline has been accepted by the National Guideline Clearinghouse (NCG:007504), which is sponsored by the Agency for Healthcare Research and Quality. J. C. Licciardone The Osteopathic Research Center, University of North Texas Health Science Center, Fort Worth, TX, USA
Read full abstract