How reliable are the Core Outcome Measures Index and Global Treatment Outcome as measures of treatment success after surgery for central lumbar spinal canal stenosis?

INTRODUCTION

The reliability of patient-reported outcome measures is typically evaluated in patients with chronic or stable symptoms, pre-treatment. After treatment, symptoms may be less extreme and less variable within a group, which may influence indices of reliability. Further, few studies have examined the reliability of retrospective, single item global assessment scores after treatment, despite the latter often being used as the external criterion when determining an instrument's minimal clinically important change (MCIC) score. This study examined the stability of Core Outcome Measures Index (COMI) and Global Treatment Outcome (GTO) scores after spine surgery.

METHODS

Data were extracted for patients with lumbar spinal stenosis who had completed the COMI and GTO twice at follow-up, for both an in-house outcomes database (ROUTINE) and as part of a separate prospective study (LSOS). To be included, the questionnaires had to have been completed within 3 months of each other for 1-year follow-up (FU), within 4 months for 2-year FU and within 5 months for 5-year FU. Repeated measures ANOVA, intraclass correlation coefficients (ICC; 2-way mixed, absolute agreement), and weighted Kappa values were calculated.

RESULTS

64 patients (72.9 ± 6.9 y; 48% female) had a COMI and GTO available from both ROUTINE and LSOS, completed on average 9.8 ± 6.2 weeks apart. There were no significant differences between test and retest scores for any of the COMI domains or for the COMI summary score (all p > 0.05), and ICCs/weighted Kappas were moderate to good (0.73-0.87). In the ROUTINE and LSOS datasets, 83% and 81% patients, respectively, reported a "good global outcome" on the GTO (i.e. treatment helped/helped a lot); the corresponding Kappa for agreement between the ratings on an individual basis was 0.74 ("good agreement").

CONCLUSION

Despite the more stringent nature of the evaluation given by this "real-life" analysis of the stability of outcome scores, the FU COMI scores showed moderate to good reliability comparable to that reported in previous studies with reliability assessed 1-2 weeks apart. The GTO was also confirmed as a reliable variable yielding stable values at mid to longer term follow-up.