The Lumbar Spine Outcomes Questionnaire: its development and psychometric properties

Abstract

There are currently a number of generic and disease-specific instruments for assessing complaints of low back pain (LBP). None provide the comprehensive coverage of the wide range of factors that are considered essential in evaluating treatment outcomes. To develop and evaluate the psychometric properties of a comprehensive, disease-specific questionnaire for characterizing complaints of LBP and evaluating the outcomes of treatments for these complaints. A clinical-cohort study of a large, convenience sample of LBP patients. We have developed a comprehensive, disease-specific questionnaire for characterizing complaints of LBP and evaluating the outcomes of treatments for these complaints. A large group of patients who sought treatment for LBP (n=2539) completed the Lumbar Spine Outcomes Questionnaire (LSOQ) before treatment, and at 12 and 24 months after treatment. For each subject and for each evaluation period, scores on six composite measures were derived from the subjects' responses to the questionnaire: a LBP severity score, a leg pain severity score, a functional disability score, a psychological distress score, a physical symptoms score, and a health-care utilization score. These scores were used to evaluate the reliability, validity, and responsiveness of the questionnaire. Test-retest reliability of the LSOQ was evaluated by correlating the subject's 12-month scores on each composite measure with the corresponding 24-month scores. Intraclass coefficients of correlation were used. The obtained coefficients of correlation [(a) LBP severity, 0.87; (b) leg pain severity, 0.85; (c) functional disability, 0.87; (d) psychological distress, 0.88; (e) physical symptoms other than pain, 0.82; and (f) health-care utilization, 0.76] indicate good test-retest reliability for the LSOQ. Construct validity was evaluated by correlating scores on the composite measures derived from the LSOQ with scores on measures of the same constructs derived from the Oswestry Low Back Pain Disability Questionnaire and the Short Form 36-Item Health Survey. The coefficients of correlation were relatively high (mostly between .7 and .9), indicating good construct validity. Construct validity was also evaluated by comparing the scores of groups of subjects who were known to differ or not to differ on the composite measures, using multivariate analyses of variance. Significant multivariate and univariate differences were obtained between groups who were expected to differ (ie, surgically and nonsurgically treated patients). No significant differences were found for groups who were not expected to differ (ie, patients with similar diagnosis, but different surgical treatments). Responsiveness was assessed by evaluating differences in the 24-month change scores between improved and unimproved subjects. Large and significant differences were obtained between improved and unimproved subjects for all composite measures derived from the LSOQ. The observed effect sizes ranged from .68 to 1.17 indicating that the LSOQ is highly responsive. The LSOQ appears to be acceptable to patients, easy to administer, highly reliable, valid, and responsive. It provides information on demographics, pain severity, functional disability, psychological distress, physical symptoms, health-care utilization, and satisfaction. It should be considered for use in both clinical and research applications as well as regulatory review involving patients with LBP complaints.