Outcome assessment in low back pain: how low can you go?

The present study examined the psychometric characteristics of a "core-set" of six individual questions (on pain, function, symptom-specific well-being, work disability, social disability and satisfaction) for use in low back pain (LBP) outcome assessment. A questionnaire booklet was administered to 277 German-speaking LBP patients with a range of common diagnoses, before and 6 months after surgical (N=187) or conservative (N=90) treatment. The core-set items were embedded in the booklet alongside validated 'reference' questionnaires: Likert scales for back/leg pain; Roland and Morris disability scale; WHO Quality of Life scale; Psychological General Well-Being Index. A further 45 patients with chronic LBP completed the booklet twice in 1-2 weeks. The minimal reliability (similar to Cronbach's alpha) for each core item was 0.42-0.78, increasing to 0.84 for a composite index score comprising all items plus an additional question on general well-being ('quality of life'). Floor or ceiling effects of 20-50% were observed for some items before surgery (function, symptom-specific well-being) and some items after it (disability, function). The intraclass correlation coefficient (ICC) ("test-retest reliability") was moderate to excellent (ICC, 0.67-0.95) for the individual core items and excellent (ICC, 0.91) for the composite index score. With the exception of "symptom-specific well-being", the correlations between each core item and its corresponding reference questionnaire ("validity") were between 0.61 and 0.79. Both the composite index and the individual items differentiated (P<0.001) between the severity of the back problem in surgical and conservative patients (validity). The composite index score had an effect size (sensitivity to change) of 0.95, which was larger than most of the reference questionnaires (0.47-1.01); for individual core items, the effect sizes were 0.52-0.87. The core items provide a simple, practical, reliable, valid and sensitive assessment of outcome in LBP patients. We recommend the widespread and consistent use of the core-set items and their composite score index to promote standardisation of outcome measurements in clinical trials, multicentre studies, routine quality management and surgical registry systems.