Validity and test-retest reliability of the Ugandan version of the Pediatric Evaluation of Disability Inventory (PEDI-UG) in children and youth with cerebral palsy.

Abstract

Validity of the Ugandan version of the Pediatric Evaluation of Disability Inventory (PEDI-UG) was previously investigated on typically developing children. This study aimed to investigate the validity, test-retest reliability and minimal detectable change (MDC) of the PEDI-UG in children and youth (C&Y) with cerebral palsy (CP). A cross-sectional study design with 118 C&Y with CP (44.7% girls) aged 10months-22.5years were included in the study; 37 of them completed the PEDI-UG twice to investigate test-retest reliability, determined by calculating the intraclass correlation coefficient (ICC). Additionally, data from 249 typically developing children were used for differential item functioning (DIF) analysis. The validity of the PEDI-UG was investigated by Rasch analysis. The Kruskal-Wallis test and Spearman's correlation coefficient were calculated to investigate associations between PEDI-UG scores and external classification systems. The principal component analysis of residuals indicated unidimensionality in all domains. The ICC values were excellent (0.98-0.99), and the MDCs were less than 6 and 13 (on a 0-100 scale) for the functional skills and caregiver assistance parts, respectively. The four-category caregiver assistance rating scale fulfilled the criteria for the analysis of rating scale functioning. In total, 78 of 189 items in the functional skills domain and two items in the caregiver assistance domain demonstrated DIF between C&Y with CP and TD children. The Kruskal-Wallis test (p < 0.05) and Spearman's correlation (coefficients of -0.93 to -0.78) supported the validity of PEDI-UG. The current diagnose-specific version of PEDI-UG demonstrates evidence for validity as a measure of ability in C&Y with CP in Uganda and other similar settings, being a promising tool for use in clinical practice and research. Conversion tables and MDC values are provided to facilitate clinical adoption of the measure.