Reliability of Observational Kinematic Gait Analysis

Gait analysis, like all clinical assessments, is subject to measurement error. Specification of the extent of measurement error is imperative before drawing conclusions from any test. The purpose of this study was to determine the within-rater and between-rater reliability of observational gait analysis in a pediatric sample wearing knee-ankle-foot orthoses. Three expert observers, using a 3-point scale, rated videotaped gait kinematics of 15 children who had lower limb disability and who wore braces. The rating sessions were then repeated, with one month between sessions. Total agreement (identical ratings), both between-raters and within-raters, occurred in two-thirds of the observations, and an additional 29% of the observations differed by one point. Between-rater intraclass correlation coefficient type 2, 1 was .73; within-rater Pearson product-moment correlation averaged .60. Observational kinematic gait analysis appears to be a convenient, but only moderately reliable, technique.