|Integration of subclassification strategies in randomised controlled clinical trials evaluating manual therapy treatment and exercise therapy for non-specific chronic low back pain: a systematic review
|Fersum KV, Dankaerts W, O'Sullivan PB, Maes J, Skouen JS, Bjordal JM, Kvale A
The review concluded that patient classification-based manual and exercise therapy interventions for non-specific chronic low back pain significantly reduced post-intervention pain and disability, but were only significant for disability in the long term. The authors' conclusions should be treated with caution due to limitations to the review process, study numbers and quality. Improved subclassification strategies were advocated.
To determine the effect of integrating subclassification strategies for patients in the evaluation of manual therapy treatment and exercise therapy for non-specific chronic low back pain.
MEDLINE (from 1966), EMBASE (from 1988), CINAHL (from 1982), PEDro and Cochrane Central Register of Controlled Trials (CENTRAL) were searched to December 2008 for studies published in English; search terms were reported. Bibliographies of included studies were handsearched.
Randomised controlled trials (RCTs) of manual therapy or exercise interventions for adults with non-specific chronic low back pain with symptom duration of more than 12 weeks were eligible for inclusion. Studies included patients with cancer, inflammatory disorders, infection, fractures, specific pathology (spondylolisthesis, disc herniation with radicular pain, degenerative disc with Modic changes, central or foraminal stenosis), pregnancy or surgery were excluded. Studies were eligible for inclusion only if they considered time frame (acute, subacute, chronic), age of participants, area of pain, level of pain, level of disability, psychosocial status, work status and whether there was compensation involved or litigation pending. Eligible studies had to attempt patient classification strategies and match treatment to the classified subgroups; any classification subgroups, physical and psychological, were accepted. Studies were also included if the classifications were made after study completion. Studies had to report on the primary outcomes of pain intensity (visual analogue scale) and disability (Roland Morris or Oswestry).
Interventions were described for three studies and included: combined manipulation, exercise and physician consultation versus consultation alone; McKenzie treatment with combined education and active patient involvement; and early morning lumbar flexion versus sham treatment. Two studies divided patients into groups according to the Multidimensional Pain Inventory, which aids in predicting rehabilitation outcome: one study had separate interpersonally distressed and dysfunctional patients; the other study had four subgroups based on both Multidimensional Pain Inventory and lumbar dynamometry results (expected performance lower than normal but consistent; grey zone with inconsistent behaviour; dysfunctional patients; and interpersonally distressed patients). Patients were classified in the other studies by: the presence and severity of radiculopathy and type of pain; the McKenzie classification of spinal pain; and age, sex, pain location and psychological overlay. More than half of the subclassification strategies (60%) reflected the biophysical model, were known to be reliable and matched the intervention; no studies were validated for non-specific chronic low back pain.
Two independent reviewers performed the study selection. Disagreements were resolved at a consensus reviewer meeting.
Assessment of study quality
The PEDro 10-point scale for RCTs was used to assess internal and external validity. Criteria were of randomisation, allocation concealment, similarity of patient characteristics at baseline, blinding of participants, therapists and assessors, measure of at least one key outcome in 85% of initial participants, results available for all patients for at least one key outcome or an intention-to-treat (ITT) analysis performed, between-group statistical result available for at least one key outcome and point measures and measure of variability available for at least one key outcome.
The authors did not report how many reviewers performed the quality assessment.
Group mean disability and pain scores were extracted with standard deviations for long- and short-term data. Where standard deviations were not reported, they were calculated from other variance data or a reasonable estimate made. Whether patients with a known psychiatric disorder had been excluded from a study was recorded.
The authors did not report how many reviewers performed the data extraction.
Methods of synthesis
Results were pooled using weighted mean differences (WMDs) with 95% confidence intervals (CIs) for pain scores and standardised mean differences (SMDs) with 95% CIs for disability scores. A fixed-effect model was used where there was low heterogeneity (Ι²<25%). A random-effects model was used when there was high heterogeneity. Between-group heterogeneity was measured using the Cochrane Q statistic and Ι² statistic. Meta-analyses were performed at the end of treatment and at 36 to 52 weeks follow-up.
Results of the review
Five RCTs were included. Data were provided for 359 participants on pain and 432 participants on disability. Two studies scored 7 for quality and three studies scored 4 or 5 out of 10.
There was a statistically significant difference in favour of classification-based interventions for reduction in pain score (100mm visual analogue scale) at end of treatment (WMD 8.59, 95% CI 2.67 to 14.50, Ι²=58%; five comparisons), but the long-term effect was not significant (WMD 0.14, 95% CI -0.01 to 0.29, Ι²=24%; nine comparisons). There was a statistically significant difference in favour of classification-based interventions for reduction in disability score at end of treatment (SMD 6.89, 95% CI 2.67 to 11.10, Ι²=0%; five comparisons) and long term (SMD 0.43, 95% CI 0.19 to 0.67, Ι²=59%; nine comparisons). Effect sizes ranged from moderate (0.43) for short term to minimal (0.14) for long term.
RCTs that evaluated manual therapy treatment and exercise therapy in patients with chronic low back pain using a classification system approach and matched treatments were very limited or non existent. A better integration of subclassification strategies in non-specific chronic low back pain outcome research was needed.
The review addressed a well-defined question in terms of study design, participants, interventions and relevant outcomes. The search for unpublished studies was not extensive and only studies published in English were included, so some relevant studies may have been missed. Study quality was assessed using relevant criteria and study quality was moderate. Efforts were made to reduce error and bias in study selection, but the authors did not report whether this applied to other aspects of the review process.
Some relevant study details were reported, but insufficient details of interventions and patient groups and numbers (including those used in the meta-analyses) were provided. The synthesis may not have been appropriate in view of the heterogeneity in the patient classification systems and interventions. The authors' conclusions in the text for pain and disability in the long term did not agree with those in the figures. The authors did not explain how they had derived the effect sizes in the abstract.
In view of the limited number and heterogeneity of the studies and review process limitations, the authors acknowledged that their meta-analysis conclusions should be treated with caution.
Implications of the review for practice and research
Practice:The authors did not state any implications for practice.
Research: The authors identified a need for further high-quality RCTs with similar comparisons. Classification systems should be based on identifying the underlying mechanisms driving the pain disorder and thereby guiding the targeted interventions, be based on a biopsychosocial construct and go through a validation process. Interventions should also target the underlying mechanisms. Well-defined inclusion and exclusion criteria should be used and minimum standards for evaluation of outcome and long-term follow-up should be established.
Norwegian Fund for Postgraduate Training in Physiotherapy.
Fersum KV, Dankaerts W, O'Sullivan PB, Maes J, Skouen JS, Bjordal JM, Kvale A. Integration of subclassification strategies in randomised controlled clinical trials evaluating manual therapy treatment and exercise therapy for non-specific chronic low back pain: a systematic review. British Journal of Sports Medicine 2011; 44(14): 1054-1062
Subject indexing assigned by NLM
Adult; Chronic Disease; Disabled Persons /rehabilitation; Evaluation Studies as Topic; Exercise Therapy; Humans; Low Back Pain /therapy; Musculoskeletal Manipulations; Randomized Controlled Trials as Topic; Treatment Outcome
Database entry date
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.