Responsiveness and minimal important change of specific and generic patient-reported outcome measures for back patients: the Norwegian Neck and Back Register

Abstract

Purpose: To evaluate responsiveness and minimal important change (MIC) of Oswestry Disability Index (ODI), pain during activity on a numeric rating scale (NRSa) and health related quality of life (EQ-5D) based on data from the Norwegian neck and back registry (NNRR). Methods: A total of 1617 patients who responded to NNRR follow-up after both 6 and 12 months were included in this study. Responsiveness was calculated using standardized response mean and area under the receiver operating characteristic (ROC) curve. We calculated MIC with both an anchor-based and distribution-based method. Results: The condition specific ODI had best responsiveness, the more generic NRSa and EQ-5D had lower responsiveness. We found that the MIC for ODI varied from 3.0 to 9.5, from 0.4 to 2.5 for NRSa while the EQ5D varied from 0.05 to 0.12 depending on the method for calculation. Conclusion: In a register based back pain population, the condition specific ODI was more responsive to change than the more generic tools NRSa and EQ5D. The variations in responsiveness and MIC estimates also indicate that they should be regarded as indicative, rather than fixed estimates.