BACKGROUND: Minimal clinically important difference (MCID) is the change on a measure that may be interpreted as treatment benefit by caregivers and clinicians, and support coverage by payers. Currently there exists no definition of MCID or responder for patients with later-onset spinal muscular atrophy (SMA).
OBJECTIVE: This study identified patients with later-onset SMA who responded to treatment and calculated a threshold of meaningful change in Expanded Hammersmith Functional Motor Scale (HFMSE) scores. Targeted analyses were conducted to estimate MCIDs and responder definitions for the HFMSE for later-onset SMA.
METHODS: Data from CHERISH, a phase 3 randomized, double-blind, multicenter, sham procedure-controlled clinical trial of later-onset SMA were analyzed. The HFMSE, the primary endpoint in CHERISH, was administered at screening and days 92, 169, 274, 365, and 456, in addition to 7-point global change ratings (CGICs) by the clinician and patient’s caregiver at all follow-up days. Anchor-based methods using the CGICs were applied to determine MCIDs and responder definitions. Supportive, distribution-based methods, the half-standard deviation (SD) and standard error of measurement (SEM), were computed, as were receiver-operating characteristic (ROC) curve analyses and cumulative distribution functions.
RESULTS: Applying a conventional anchor-based definition for MCID yielded 3.82 based on the caregiver CGIC and 3.46 using the clinician CGIC at day 456; half-SD=4.1 and SEM=2.6. ROC curve analyses supported a smaller MCID of 2 HFMSE points of improvement as meaningful to caregivers and clinicians. A provisional MCID and responder definition for the HFMSE in later-onset SMA was approximately 3-4 points of improvement on the 0 to 66 HFMSE scale. Applying a 3-point change in HFMSE as the responder threshold to the CHERISH data, 60% (n=21) of patients in the nusinersen group versus 21% (n=4) patients in the sham group improved by ≥3 HFMSE points at day 456.
CONCLUSIONS: An MCID to help interpret HFMSE scores in later-onset, treated SMA patients was calculated using multiple methods. Provisional MCID and responder values converged on 3-4 points of improvement and were deemed meaningful and important to stakeholders such as caregivers and clinicians, to be confirmed in future studies. These values may also inform healthcare coverage decisions.
