The Ugandan version of the Pediatric Evaluation of Disability Inventory (PEDI-UG). Part II: Psychometric properties.

Abstract

The Pediatric Evaluation of Disability Inventory (PEDI) has been recommended as a gold standard in paediatric rehabilitation. A Ugandan version of PEDI (PEDI-UG) has been developed by culturally adapting and translating the original PEDI. The aim of this study was to investigate the psychometric properties of the PEDI-UG in Ugandan children by testing the instrument's rating scale functioning, internal structure, and test-retest reliability. Two hundred forty-nine Ugandan children (125 girls) aged 6months to 7.5years (Mean=3.4, SD=1.9) with typical development were tested using the PEDI-UG. Forty-nine children were tested twice to assess test-retest reliability. Validity was investigated by Rasch analysis and reliability by intraclass correlation coefficient. The PEDI-UG domains showed good unidimensionality based on principal component analysis of residuals. Most activities (95%) showed acceptable fit to the Rasch model. Six misfit items were deleted from the Functional Skills scales and one from the Caregiver Assistance scales. The category steps on the Caregiver Assistance scales' rating scale were reversed but functioned well when changed from a 6-point to 4-point rating scale. The reliability was excellent; intraclass correlation coefficient was 0.87-0.92 for the domains of the Functional Skills scales and 0.86-0.88 for the domains of the Caregiver Assistance scales. The PEDI-UG has good to excellent psychometric properties and provides a valid measure of the functional performance of typically developing children from the age of 6months to 7.5years in Uganda. Further analysis of all items, including misfit and deleted items, in children with functional disability is recommended.