Derivation of an abbreviated instrument for use in emergency department low back pain research: The five-item Roland Morris Questionnaire

Objectives Low back pain (LBP) is a common reason for emergency department (ED) visits in the United States. Pain and functional outcomes after ED visits for LBP tend to be poor. ED-based clinical LBP research is hampered by complexity of available outcome instruments, which can be time-consuming to administer. The purpose of this investigation was to determine if a shorter version of the well-validated and commonly used Roland Morris Disability Questionnaire (RMDQ) would retain the original 24-item instrument's ability to assess functional outcomes accurately in ED patients with LBP. Methods The authors used deidentified data obtained from a prospective LBP cohort study, which enrolled 674 patients during index ED visits for LBP, and followed them by telephone 1 week and 3 months later. Five items were selected from the original 24 items of the RMDQ using regression techniques. Internal consistency of the abbreviated scale was measured using Cronbach's alpha. The strength of association between the five-item scale (RM5) and the parent scale was determined at baseline, 1 week, and 3 months. The association between change in the parent scale and change in the RM5 and the change in RM5 that was most closely associated with a minimum clinically significant difference on the RMDQ were also determined. Results The Cronbach's alpha for the RM5 was 0.88 (95% confidence interval [CI] = 0.87 to 0.89) at baseline, 0.96 (95% CI = 0.96 to 0.96) at 7 days, and 0.97 (95% CI = 0.97 to 0.97) at 3 months. The Pearson correlation coefficient for the RM5 versus the parent scale was 0.93 (R ² = 0.86) at baseline, 0.98 (R² = 0.96) at 1 week, and 0.98 (R² = 0.96) at 3 months. The correlations between change from baseline in the abbreviated scale and the parent scale were 0.95 (R² = 0.90) and 0.96 (R² = 0.92) at 7 days and 3 months, respectively. A one-point change in the RM5 has a sensitivity of 96% (95% CI = 93% to 98%) and a specificity of 92% (95% CI = 89% to 94%) for the minimum clinically significant change on the RMDQ. Conclusions An abbreviated five-item version of the RMDQ was developed. Pending independent validation, this shortened instrument should streamline ED-based low back pain research.