Psychometric evaluation of the functional walking test for children with cerebral palsy.

Purpose. This study examined the psychometric properties of the functional walking test (FWT). Method. Fifty-six subjects with cerebral palsy (CP) (21 females and 35 males, mean age 9 years 6 months, SD 3 years 9months, range 4-17 years) were assessed on two occasions, 6 months apart, using both the FWT and the gross motor function measure (GMFM). Results. Generalisability correlation coefficients (GCC) for all 11 items were high (0.91-0.99). Inter-rater reliability was also high with excellent consensus in the scores given by the eight raters (intra-class correlation coefficient and GCC 0.99). Intra-rater reliability was equally high (GCC 0.99). The internal consistency of the FWT was estimated using Cronbach's α as 0.95 and 0.94 at Time 1 and 2, respectively. The FWT had a high degree of correlation with the GMFM, when total scores were compared at Time 1 and 2 (Pearson's r = 0.86 and 0.87, n = 56, p < 0.01). The FWT also found statistically significant differences in total scores between the three Gross Motor Function Classification System (GMFCS) levels. The correlation between the FWT scores and GMFCS was -0.70 at Time 1 and -0.76 Time 2 (p < 0.01) indicating the construct validity of the FWT. Conclusions. This study has demonstrated that the FWT has sound psychometric properties and is valid and reliable in a sample population of ambulant children with CP.