OBJECTIVES:

Minimal clinically important differences (MCIDs) represent the smallest change following a clinical intervention considered meaningful for patients. The objective was to estimate the correlation of MCIDs for patient reported outcome measures (PROMs) related to spinal surgery.

METHODS:

Using data from the Swedish spine register Swespine, we estimated changes in PROMs Oswestry Disability Index (ODI), Visual Analog Scale (VAS) for pain and EQ-5D in patients who had lumbar spine surgery between 2000 – 2012 from surgery to two-year follow up. VAS leg pain was used for patients diagnosed with spinal stenosis and disc herniation whereas VAS back pain was used for spondylolisthesis and degenerative disc disease. Two binary variables were constructed for each of the PROMs indicating whether the change could be defined as an improvement according to the highest and lowest MCID cut-offs identified in the literature signifying improvements in back pain following surgical interventions. Correlations between MCID-defined improvements in the PROMs were estimated with the phi coefficient.

RESULTS:

MCID cut-offs identified in the literature ranged from -5 to -20 for ODI, from -18 to -35 for VAS, and from 0.05 to 0.52 for EQ-5D. Considering the lower (upper) thresholds, the correlation between improvement in ODI and VAS was 0.40 (0.39), between ODI and EQ-5D 0.48 (0.38), and between EQ-5D and VAS 0.35 (0.25). All correlations were significant at the 1% level. This indicate moderate relationships between having improved according to MCID in one PROM and having improved by another PROM, except for VAS and EQ-5D considering the upper MCID threshold, in which case the correlation was low but not negligible.

CONCLUSIONS:

Low to moderate relationships of having improved outcomes using different PROMs were found. Relying on only one PROM to infer whether a patient has had a clinically meaningful improvement is associated with uncertainty.