We would like to thank the BC Medical Journal for the opportunity to respond to some of the issues raised by Dr Bishop in Counterpoint September 2004 [BCMJ 2004;46:349, 353-354].
Dr Bishop raises the important point of reliability, an essential element of every assessment system. We were remiss in not including this in our paper. Dr Bishop questions the reliability of the McKenzie assessment based on one article by Riddle et al. Why he selected this one study and omitted the four subsequent studies that demonstrate strong reliability is unclear.[1-4] It is noteworthy that these other four all used clinicians fully educated in McKenzie methods while the Riddle study used novice clinicians.
Dr Bishop describes the importance of establishing guidelines; we agree. The Danish Institute for Health Technology Assessment published in 1999[5] differs from other guideline groups in that they are the first to review the literature investigating the McKenzie assessment procedures independent from the literature on treatment. They assigned the McKenzie assessment their highest grade for scientific support for reliability and validity.
Borkan and colleagues 1998,[6] Bouter and colleagues 1998,[7] and Bouter and colleagues 2003 of the Cochrane Back Review Group[8] all described the vital necessity of trying to identify distinct subgroups in the broad category of non-specific back pain. Randomized controlled trials are not the appropriate study design for identifying subgroups. Guidelines panels will need to look to other study designs, as Denmark did, to fulfill our fundamental need to identify and validate subgroups.
Our original article [BCMJ 2004;46:348, 350-352] on the centralization phenomenon cited numerous studies that, without exception, demonstrated that patients whose symptoms centralize form a distinct subgroup in the low back pain population based on their consistently superior outcomes compared with non-centralizers, another distinct subgroup, that typically have slow or non-recoveries. To be able to make such a distinction at the outset of care is of great value, especially if the assessment also identifies patient-specific exercise and posture strategies that centralize, abolish, and then prevent the return of the pain.
These international experts we have referenced correctly state that future randomized controlled trials should focus on treatment interventions for these subgroups, illustrated by the Long and colleagues study we cited, due to be published in Spine in December. This study dramatically demonstrates the superiority of subgroup-specific treatments with every outcome measured. It is rather like conducting a trial of nitroglycerin for patients with angina vs only studying those with non-specific chest pain. The medication is clearly demonstrated as useful and efficacious in the former study but barely better than doing nothing in the latter. Results will always be confusing, even misleading, if we continue to study non-specific low back pain, just as in the 1998 Cherkin study Dr Bishop cited, which reported a lack of efficacy of either chiropractic or McKenzie in a non-specific back pain sample.
We consider Dr Bishop’s dismissal of the value of the McKenzie assessment to be based on an incomplete knowledge of the relevant literature. In actuality, the evidence supporting the reliability and validity of a McKenzie-based examination is, as we previously stated, already strong and growing steadily.
—C.L. Davies, Vancouver
—C.M. Blackwood, MD, Mission
