Abstracts

Rationale: Caregiving for children with epileptic encephalopathies (EE) affects caregivers’ lives in many significant ways. To better understand the effects of providing care for a child with EE, a measure that assesses life domains most important to caregivers of children with EE is needed. The purpose of this study was to develop brief, flexible, and psychometrically sound tools that could be used to compare benefits and stress of caregiving across populations and studies, including assessment of EE treatment effects on caregivers.  Methods: An expert panel of pediatric neurologists and caregivers of children with EE provided feedback on the items to be included in the instrument. Candidate items assessed both stress and benefits, and were administered to caregivers of children (age < 18) with EE, Down syndrome (Down), Muscular Dystrophy (MD) and a community sample (CS) through an online survey. PROMIS global health mental (GMH) and physical (GPH) scores were used to assess caregiver health. Higher score indicates better health. Test-retest data were collected 40 to 80 hours apart and test-retest reliability was evaluated using the Intraclass correlation (ICC). The items were examined for unidimensionality, local dependence (LD) and differential item functioning (DIF). A 2-parameter graded response IRT model was used to develop the item bank.  Results: Caregivers of children with EE (n=126), Down (n=63), MD (n=129), and a CS (n=314) responded to candidate items (n=50) either once (n=632) or twice (n=112). Average age of caregivers was 41 years (SD=8), 80% were female, 81% were white, 71% were married, 19% had high school education or less, and 41% were employed full time. GMH ranged from 38.1 (EE) to 46.7 (CS) and GPH scores ranged from 44.1 (EE) to 48.0 (MD).Factor analyses suggested two distinct constructs, caregiver stress (n=35 candidate items) and caregiver benefit (n=15 candidate items). Here, we report on the Caregiver Stress item bank. Based on results of IRT analyses 17 items had to be removed due to problems with unidimensionality, LD, or model fit. The remaining 18 items were essentially unidimensional with no indication of LD or bias due to gender, age or education. Five response options functioned well for all items. Test-retest was excellent with an ICC for the total score from the first and second administration of 0.95. The t-score had reliability >0.80 between +/-3 standard deviations from the mean. Correlation with the Paediatric Renal Caregiver Burden Scale, r(630) = 0.7, p < 0.01 supported validity. Highest caregiver stress was reported by caregivers of children with EE (M=59.0, SD=6.7) compared to Down (M=51.2, SD=8.9), MD (M=50.2, SD=9.1) and CS (M=46.0, SD=9.2). Conclusions: The 18 item UW Caregiver Stress Scale (UW-CSS) item bank appropriate for caregivers of children with epilepsy was developed using person-centered methodology, has excellent psychometric properties, and is publicly available to researchers and clinicians to administer by computerized adaptive testing or by short forms. Initial findings suggest caregivers for children with EE have higher levels of stress than do those caring for children with Down and MD. Funding: This investigator initiated study was supported by Zogenix and in part by a grant from the National Institute on Disability, Independent Living, and Rehabilitation Research H133P120002.