The Core Outcome Measures Index (COMI) is a responsive instrument for assessing the outcome of treatment for adult spinal deformity

Introduction: The Core Outcome Measures Index for the back (COMI-back) is a very brief instrument for assessing the main outcomes of importance to patients with back problems (pain, function, symptom-specific well-being, quality of life, disability). However, it might be expected to be less responsive than a disease-specific instrument when evaluating specific pathologies. In patients with adult spinal deformity, we compared the performance of COMI-back with the widely accepted SRS-22 questionnaire.

Methods: At baseline and 12 months after non-operative (N = 121) and surgical (N = 83) treatment, patients (175 F, 29 M) completed the following: COMI-back, SRS-22, Oswestry Disability Index (ODI) and SF-36 PCS. At 12 months' follow-up, patients also indicated on a 15-point Global Rating of Change Scale (GRCS) how their back problem had changed relative to 1 year ago. Construct validity for the COMI-back was assessed by the correlation between its scores and those of the comparator instruments; responsiveness was assessed with receiver operating characteristics (ROC) analysis of COMI-back change scores versus the criterion 'treatment success' (dichotomized GRCS).

Results: Baseline values for the COMI-back showed significant (p < 0.0001) correlations with SRS-22 (r = -0.85), ODI (r = 0.83), and SF-36 PCS (r = -0.82) scores; significantly worse scores for all measures were recorded in the surgical group. The correlation between the change scores (baseline to 12 months) for COMI and SRS-22 was 0.74, and between each of these change scores and the external criterion of treatment success were: COMI-back, r = 0.58; SRS-22, r = -0.58 (each p < 0.0001). The ROC areas under the curve for the COMI-back and SRS-22 change scores were 0.79 and 0.82, respectively.

Conclusion: Both baseline and change scores for the COMI-back correlated strongly with those of the SRS-22, and differed significantly in surgical and non-operative patients, suggesting good construct validity. With the "change in the back problem" serving as external criterion, COMI-back showed similar external responsiveness to SRS-22. The COMI-back was well able to detect important change. Coupled with its brevity, which minimizes patient burden, these favourable psychometric properties suggest the COMI-back is a suitable instrument for use in registries and can serve as a valid instrument in clinical studies emerging from such data pools.