Minimally clinically important relative change instead of difference following surgery for degenerative cervical radiculopathy: a nationwide study on 5300 patients using prospectively collected data.

Objective: The landscape of healthcare policies has undergone significant reforms, with a notable shift toward a patient-centered basis. This change has prioritized the use of metrics reflecting subjective patient-reported outcome measures (PROMs). However, these metrics are often limited, especially in terms of determining clinically relevant changes over time. The concept of the minimal clinically important difference (MCID) has been introduced to address this limitation. In this study, the authors aimed to assess the predictive capability of relative change and its corresponding threshold value, the minimal clinically important relative change (MCIRC), as an alternative to the absolute difference and MCID.

Methods: Prospectively collected data from 2006 to 2021 were obtained from the Swedish Spine Registry (Swespine). PROMs included the numeric rating scale (NRS) neck and arm pain scores, the EQ-5D index and EQ visual analog scale (VAS), and the Neck Disability Index (NDI). Anchor-based methods were used to determine MCID and MCIRC. Absolute differences and relative changes were compared in terms of predictive power using the area under the receiver operating characteristic curve (AUC).

Results: A total of 5323 patients were included. The chosen MCID values were -3 for NRS neck pain, -3 for NRS arm pain, 0.08 for EQ-5D index, 13 for EQ VAS, and -30 for NDI scores. The chosen MCIRC values were -40% for NRS neck pain, -50% for arm pain, 323% for EQ-5D index, 55% for EQ VAS, and -33% for NDI scores. Relative change was superior for NRS neck and arm pain scores and NDI, while absolute difference was superior for EQ-5D and EQ VAS scores.

Conclusions: Relative change and its corresponding MCIRC were found to be more appropriate as a surrogate of subjective satisfaction when studying the NRS and NDI scores, whereas absolute difference and MCID was more appropriate when addressing EQ-5D index and EQ VAS scores.