Responsiveness of pain, disability, and physical impairment outcomes in patients with low back pain

• Published in: Spine (Philadelphia, Pa. 1976) Pa. 1976), 2004-04, Vol.29 (8), p.879-883
• Main Authors: PENGEL, Liset H. M, REFSHAUGE, Kathryn M, MAHER, Chris G
• Format: Article
• Language: English
• Subjects: Activities of Daily Living; Adolescent; Adult; Aged; Aged, 80 and over; Australia; Biological and medical sciences; Cerebrospinal fluid. Meninges. Spinal cord; Cohort Studies; Disability Evaluation; Female; Humans; Low Back Pain - diagnosis; Low Back Pain - physiopathology; Male; Medical sciences; Middle Aged; Nervous system (semeiology, syndromes); Neurology; Outcome Assessment (Health Care) - methods; Pain Measurement - statistics & numerical data; Predictive Value of Tests; Severity of Illness Index; Surveys and Questionnaires
• ISSN: 0362-2436; 1528-1159
• DOI: 10.1097/00007632-200404150-00011

Cohort study. To conduct a head-to-head comparison of the responsiveness of pain, disability, and physical impairment measures in patients with low back pain. Pain, disability, and physical impairment measures are routinely measured in clinical practice and clinical research. However, to date, a head-to-head comparison has not been performed. A numerical pain scale (0-10), the 24-item and 2 modified 18-item versions of the Roland Morris questionnaire, the patient specific functional scale, and physical impairment measures were completed by 155 patients with low back pain at baseline and then again after 6 weeks together with an 11-point global perceived effect scale. Responsiveness was evaluated by using effect sizes and t tests, correlating the change scores for each outcome with the global perceived effect score and by calculating the Guyatt responsiveness index. The most responsive outcome proved to be the patient specific functional scale (effect size = 1.6), followed by the numerical pain scale (effect size = 1.3) and 24-item Roland Morris questionnaire (effect size = 0.8). The responsiveness of the two 18-item Roland Morris versions was equal to the 24-item version. However, the physical impairment measures were not very responsive (effect size 0.1-0.6). The ranking of the responsiveness indices was consistent across all statistical analyses. Physical impairments are routinely measured in clinical practice and clinical research, but the lower responsiveness indicates that this approach is not optimal. Our findings suggest that more emphasis should be placed on change in pain and disability scores than on change in physical impairments.