Qualitative evaluation of meaningful change in Dravet syndrome as measured by the Vineland-3: Caregiver and clinician perspectives.

The Vineland Adaptive Behavior Scales, Third Edition (Vineland-3) is used to assess non-seizure outcomes in developmental and epileptic encephalopathies, including Dravet syndrome (DS); meaningful change thresholds (MCTs) have not been established. This qualitative study examined MCTs for the Vineland-3 and associated life impacts. Semi-structured interviews with 18 caregivers of children/adolescents with DS, including concept elicitation then completion and qualitative exploration of meaningful change for eight Vineland-3 subdomains considered most relevant to DS were conducted. Findings were used in clinical expert discussions. Caregivers spontaneously reported 35 signs/symptoms; most frequent were seizures (n = 16). Disruptive behavior, delays in adaptive behavior, and impaired communication were considered important to change with treatment. Raw score improvements ranging from 1 to 3 points (2 to 3 growth scale values) across subdomains were considered meaningful to ≥50 % of caregivers. These meaningful changes were associated with greater patient independence and safety. Caregivers ranked expressive and receptive communication as most important for improvement. Clinical experts' feedback regarding clinically meaningful changes aligned with caregivers' feedback. Clinical experts highlighted multiple factors for determining Vineland-3 MCTs in DS, including age, underlying intellectual and developmental ability, and presence of autism spectrum disorder. Overall, small changes on the Vineland-3 were considered meaningful to caregivers and clinical experts and were associated with important impacts for patients and families. This study provides novel insights supporting Vineland-3 MCT selection in DS, supporting the Vineland-3's use to explore non-seizure treatment-related benefit in clinical trial settings. Further evaluation of MCTs via in-trial interviews and quantitative anchor-based methods may be considered.