Usefulness of minimum clinically important difference for assessing patients with subaxial degenerative cervical spine disease: statistical versus substantial clinical benefit

Abstract

Background: The measurement of the therapeutic outcome of cervical spine surgeries commonly relies on four main patient reported outcomes (PROs): Neck Disability Index (NDI), Visual Analog Scale (VAS) for pain, and Short Form-36 (SF-36) Physical (PCS) and Mental (MCS) Component Summary. However, the clinical impact of such scores and how they could effectively measure therapeutic efficacy remains unclear. In this context, the concept of minimum clinically important difference (MCID) is developing into the standard by which to evaluate treatments, patient satisfaction and cost-effectiveness.

Methods: Eighty-eight consecutive patients undergoing surgery for subaxial degenerative cervical spine disease were selected from a prospective blinded database. PROs (NDI, PCS, MCS and VAS) were collected preoperatively, and together with blinded Surgeon Ratings (SR) at 3 months and 6 months post-surgery. Four anchor-based approaches were used to calculate different MCIDs. Three anchors (VAS, HTI (Health Transition Item of the SF-36) and SR) were used to evaluate surgery outcome. The best clinically and statistically relevant MCID was chosen.

Results: On average, all patients presented with a statistically significant improvement (p < 0.001) postoperatively for NDI (27.42 to 19.42), PCS (33.02 to 42.03), MCS (44 to 50.74) and VAS (2.85 to 1.93). The four MCID anchor-based approaches yielded a range of values for each PRO: 2.23-16.59 for PCS, 0.11-16.27 for MCS and 2.72-12.08 for NDI. When compared to the VAS and HTI anchors, the area under the ROC curve was greater for SR. This finding suggests that SR may be a more reliable anchor for MCID calculation.

Conclusion: The MDC (minimum detectable change) approach together with the SR anchor appears to be the most appropriate MCID method. It offers the greatest area under the ROC curve (threshold above the 95 % CI), and the choice of the anchor did not significantly affect this result. MCID values for this dataset were 5.6 for PCS, 5.12 for MCS and 2.41 for NDI.