Reliability and validity of Hebrew Pediatric Evaluation of Disability Inventory (PEDI) in children with cerebral palsy -- health care professionals vs. mothers.

Aim: To evaluate the reliability and validity of the PEDI in Hebrew (PEDI-H) in children with cerebral palsy (CP) using health care professionals' (HCP) and mothers' evaluations.
Methods: The sample comprised 73 participants (40 males, 33 females) with CP. Two modes of PEDI-H administration were used: interview of the mothers by a social worker and HCP evaluation. PEDI-H reliability was examined by two modes: 1) internal consistency via Cronbach's alpha and 2) overall absolute agreement within subject reliability via intraclass correlation coefficient (ICC). Discriminative validity using collapsed strata of the Gross Motor Functional Classification System (GMFCS) (area under the curve=AUC) were examined for each of the PEDI-H sub-domains.
Results: Participants' mean age was 8 years 8 months (standard deviation (SD) 2 years 10 months). The reliability of mothers' PEDI-H was good-to-excellent (Cronbach's alpha=0.889-0.964, ICC=0.845-0.938). The HCPs' reliability was excellent (Cronbach's alpha and ICCs > 0.90). The PEDI-H was also reliable in children with mild, moderate, and severe CP (GMFCS=I+II, III and IV+V, respectively), in younger (6-7 years) and older children (8-12 years), and in children with various CP distribution. Mothers and HCPs had low accuracy in Social-Function domains (AUC=0.538-0.686) and moderate-to-high accuracy in Mobility and Self-Care domains (AUC=0.887-0.967). PEDI-H was able to distinguish between children with various CP severities.
Conclusion: The PEDI-H has good psychometric properties when administered by mothers and HCPs and can be used in older children with CP.