The effect of methodological quality, sample size and statistical rigour on outcomes of randomised controlled trials in low back pain

Purpose and Background: Reliable and valid RCTs are essential in guiding clinical practice, but shortcomings in methodological quality have been reported in RCTs on LBP. The aim of this paper was to use the results of a systematic review on exercise or manual therapies for persistent LBP to evaluate the effect of methodological quality, sample size and statistical rigour on the outcomes of these trials.Methods and Results: The systematic review included 41 RCTs on exercise or manual therapies (i.e. manipulation, mobilisation and/or massage) for persistent (>6 weeks) non-specific LBP. Quality of the RCTs was assessed using an adapted 10-point Van Tulder scale. Sample size was defined as the number of subjects in the intervention group. Adequate statistical testing was defined as analyses that compared the change in pain or function achieved by the intervention group with the change in the same parameter achieved by the control or alternative group. The results showed that the RCTs with smaller sample sizes or RCTs of lower methodological quality more often reported larger differences in effectiveness than RCTs of higher methodological quality or larger sample sizes. Furthermore, small differences in effectiveness reported by smaller RCTs were often not statistically significant, while larger trials showed that such differences actually were statistically significant.Conclusion: Low methodological quality and small sample size has resulted in misinterpretation of RCTs. Small or low quality RCTs overestimated differences in effectiveness or failed to detect smaller but statistically significant differences. Future RCTs and systematic reviews should address these shortcomings in order to provide reliable guidance for clinical practice