Presentation (CTI)

OBJECTIVES: The MSAS scale is a new patient-reported outcome (PRO) instrument assessing autonomy in multiple sclerosis. To interpret score changes, establishing a minimal clinically important difference (MCID) is essential. This analysis aims to determine the MCID for the global impact score (range: 0-100, higher indicating worsened autonomy).
METHODS: Data from the Focal MS 2 study were analyzed (N=199 patients included; n=158 completers at D360). MCIDs were calculated across five anchor-based methods (within-patient change, between-patient change, linear regression, ROC optimization using Youden and Top-Left indices) using two strategies:

- Global unidirectional: Absolute change in MSAS score from baseline to D360.

- Stratified bidirectional: Improvement (score decrease) or deterioration (score increase).

Responders were defined as patients with a change ≥1 point in their anchor score between baseline and D360 (|ΔAnchor|≥1), while non-responders were those whose anchor score remained unchanged (|ΔAnchor|=0). A sensitivity analysis was planned to exclude discordant patients, defined as those whose global score evolution contradicted their anchor response (e.g., reported as deteriorated by the anchor but showing an improved global score). To account for noise measurement, observed thresholds were compared against the minimal detectable change (MDC), defined as the upper limit of the 95% CI for the average change score in non-responders.
RESULTS: Analyses included 158 patients. The average change score in non-responders (n=82) was 7.86, 95%CI:[6.73-9.0]. Consequently, the MDC was established at 9.0 points. The unidirectional approach yielded thresholds exceeding MDC, ranging from 9.55 (ROC Top-Left) to 11.41 points (ROC Youden). Bidirectional analysis yielded a threshold of -4.44 points for improvement and +4.21 points for deterioration, while the bidirectional sensitivity analysis (excluding 24 discordant patients), converged at -10.20 for improvement and +10.78 for deterioration.
CONCLUSIONS: This analysis using different anchor-based methods shows a convergent MCID threshold of ~10 points as a meaningful clinical change in patient autonomy.