BACKGROUND: Mechanical Diagnosis and Therapy (MDT) is a treatment-based classification system founded on 3 core principles: Classification into diagnostic syndromes, classification-based intervention, and appropriate application of force. Many randomized controlled trials have investigated the efficacy of MDT for low back pain; however, results have varied. The inconsistent delivery of MDT across trials may explain the different findings. UUOBJECTIVES: To compare treatment effect sizes for pain or disability between trials that delivered MDT consistent with the core principles of the approach and trials that met some or none of these principles. METHODS: In this systematic review, databases were searched from inception to June 2018 for studies that delivered MDT compared to nonpharmacological, conservative control interventions in patients with low back pain and reported outcomes of pain or disability. Studies were classified as "adherent" (meeting the core principles of MDT) or "nonadherent" (using some or none of the principles of MDT). Data were extracted by 2 independent reviewers. Meta-regression procedures were used to analyze the effect of delivery mode on clinical outcomes, adjusting for covariates of symptom duration (less than or greater than 3 months) and control intervention (minimal or active). RESULTS: Studies classified as adherent to the MDT approach showed greater reductions in pain and disability of 15.0 (95% confidence interval: 7.3, 22.7) and 11.7 (95% confidence interval: 5.4, 18.0) points, respectively, on a 100-point scale compared to nonadherent trials. CONCLUSION: This review provides preliminary evidence that treatment effects of MDT are greater when the core principles are followed.