Comparison between visual and three-dimensional gait analysis in patients with spastic diplegic cerebral palsy.

Fifty patients with spastic diplegic cerebral palsy were included in this retrospective study which compared visual assessment of gait to three-dimensional (3D) gait analysis. Inter-observer variability was evaluated as well. Inclusion criteria comprehended independent ambulation (i.e. without assistive devices or orthoses). All subjects went through 3D gait analysis at the Gait Analysis Laboratory of the AACD Hospital. Four observers, viewing videotaped gait cycles, evaluated 10 specific points of interest of the cycle: hip flexion at terminal stance; knee flexion at initial contact; knee extension at terminal stance; knee flexion at initial swing; ankle dorsiflexion at initial contact; pelvic obliquity at mid stance; hip adduction at loading response; pelvic rotation; hip rotation at mid stance and foot progression angle, in relation to the lower limb, at mid stance. Their evaluation was then compared to the 3D kinematics data. A statistical analysis of the results was performed using kappa and McNemar's test in order to determine inter-observer and visual/3D analysis agreement. Results showed that inter-observer agreement was high but on the other hand, only two points of the gait cycle (knee flexion at initial contact and pelvic obliquity) were shown to have been similarly evaluated visually and with the 3D analysis. In conclusion, this study indicates that only knee flexion at initial contact and pelvic obliquity appear to be reliably evaluated on a visual basis alone. Visual observation is therefore inadequate for the evaluation of the other eight selected points of the gait cycle which require some form of quantitative assessment.