Background/Purpose: One goal of LBP assessment is to direct clinicians to specific sub-groups that benefit from particular treatment approaches.

The purpose of this study was to compare outcomes between patients assessed and treated based on a diagnostic system of LBP classification and those managed without a classification system.

Methods/Results: This was a prospective double-cohort study of mechanical LBP cases (n=2110) who started a rehabilitation program at 15 clinics across Canada between February 06 and August 07. A group of patients attending 7 clinics specifically trained in the use of a LBP classification system using patterns of pain, not anatomic site or pathological process (n=1356) were compared with patients from 8 clinics using a generic, non-classification approach (n=754). Patients are categorized into one of five classifications, each dictating a separate treatment approach.

The mean age of the cohort was 44.7 years (SD=13.3, range=18–89) with 55.1% males. The mean lagtime from injury to treatment was 110.7 days (median=32 days).

Treatment based on this classification system resulted in more pain reduction (p< 0.001), higher functional improvement (p< 0.001), less medication use (p< 0.001) and fewer treatment days (p< 0.001), compared to a generic approach.

Conclusion: Outcome differences between groups suggest that LBP is heterogeneous, and recognizable by clinical patterns of patient characteristics. The results support the effectiveness of a classification system that matches treatment to patient presentation and pattern of pain. Showing that one cohort had better outcomes than another as a result of classifying back pain is formidable test of validity.