Outcome assessments in children with cerebral palsy, Part II: Discriminatory ability of outcome tools

Discriminatory ability of several pediatric outcome tools was assessed relative to Gross Motor Function Classification System (GMFCS) level in patients with cerebral palsy. Five hundred and sixty-two patients (400 with diplegia, 162 with hemiplegia; 339 males, 223 females; age range 4-18y, mean 11y 1mo [SD 3y 7mo]), classified as GMFCS Levels I to III, participated in this prospective multicenter, cross-sectional study. All tools were completed by parents and participants when appropriate. Effect size indices (ESIs) for parametric variables and odds ratios for non-parametric data quantified the magnitude of differences across GMFCS levels. Binary logistic regression models determined discrimination, and receiver operating characteristic curves addressed sensitivity and specificity. Between Levels I and II, the most discriminatory tools were Gross Motor Function Measure (GMFM-66), velocity, and WeeFIM Mobility. Between Levels II and III, the most discriminatory tools were GMFM Dimension E, Pediatric Functional Independence Measure (WeeFIM) Self-Care and Mobility, cadence, and Gillette Functional Assessment Questionnaire Question 1. Large ESIs were noted for Parent and Child reports of Pediatric Outcomes Data Collection Instrument (PODCI) Sports & Physical Function, Parent report of PODCI Global Function, GMFM Dimension E, and GMFM-66 across all GMFCS level comparisons. The least discriminatory tools were the Quality of Life and cognition measures; however, these are important in comprehensive assessments of treatment effects.