RELIABILITY OF KINEMATICS DURING CLINICAL GAIT ANALYSIS: A COMPARISON BETWEEN NORMAL AND CHILDREN WITH CEREBRAL PALSY Freeman Miller, M.D.*, Patrick Castageno, M.S.*, James Richards, Ph.D.**, Nancy Lennon, P.T.*, Edward Quigley, M.S.*, Tim Niiler, M.S.* *Alfred *I. duPont Institute, Wilmington, DE 19899 **University of Delaware, Newark, DE, 19716 Introduction: Understanding the variations in gait parameters which are utilized to make clinical treatment decisions is important. These variations have not been quantified in children with cerebral palsy compared to their peers, although, clinical decisions are made with the presumption that variations are minimal. A review of literature relative to the repeatability of gait analysis measurements yields very little information. An initial report investigated between and within day variability of gait kinematics but used only 2 normal adults (Winter, 1984). additional research involving 40 normal adults has revealed extremely reproducible kinematic results over 3 visits (Kadaba et al., 1989). The only attempt to investigate variability and reproducibility in children with cerebral palsy over multiple visits concentrated on temporal-spatial characteristics and force plate recordings (Kirkpatrick, et. al., 1994). Subsequently, this study demonstrated highly reproducible results for within each subject and between multiple test days. The specific goal of this study is to compare variations in waling gait kinematics between children with cerebral palsy and age matched normal peers. Methodology: Ten children, five with CP and five non-disabled age matched peers volunteered for this study. Subjects ranged in age from six to sixteen, with a mean age of 9.6 years. Video data was obtained using a high-resolution 6 camera Motion Analysis® system. Trials were collected for a 5 second duration at 60 Hz. Patients were asked to walk at a self-selected speed for three trials per visit on 5 separate visits. Project-specific software was created to calculate bilateral 3-dimensional joint kinematics for the ankle, knee, and hip using a modified Helen Hayes marker set. Data from each subject consisted of 3 gait cycles bilaterally for each visit. Gait cycles were dilated to 100 samples. Intraclass correlation coefficient profiles (ICC) were created within and between visits to determine the reliability of the kinematic data for each joint. The ICC profiles as seen in figure 1, were generated for each sample during the gait cycle across all visits for each subject for the between visit comparison and for each sample during the gait cycle across each subject's 3 trials within each visit. This method was used to determine reliability of joint kinematic data.