Reliability of kinematic measures of functional reaching in children with cerebral palsy

Abstract

The determination of rehabilitation effectiveness in children with cerebral palsy (CP) depends on the metric properties of the outcome measure. We evaluated the reliability of kinematic measures of functional upper limb reaching movements in children with CP. Thirteen children (ten females, three males) with spastic hemiplegic, diplegic, or quadriplegic CP affecting at least one arm (mean age 9y, SD 1.6y; range 6-11y; Manual Ability Classification System [MACS] levels II-IV) were evaluated three times over 5 weeks. The kinematics of the more affected arm reaching to grasp a 2cm(3) block placed at three distances from the body midline were analysed. The reliability (test-retest) of six kinematic variables (endpoint trajectory straightness and smoothness, trunk displacement, elbow extension, shoulder horizontal adduction, and shoulder flexion] was tested and expressed as intraclass correlation coefficients (ICC, model 2,K) and 95% confidence intervals. Trajectory smoothness, trunk displacement, elbow extension, and shoulder flexion (far target) had the highest ICCs (0.82-0.95). Other kinematic variables had moderate (0.50< or =ICC< or =0.81) or low (0.17-0.38) reliability. Test-retest reliability was task dependent, as reaches required different degrees of trunk displacement and joint excursion. Kinematic variables can be used as outcomes in clinical trials to test upper limb intervention effectiveness on motor performance and movement quality. As kinematic variables are task specific, reliability should be interpreted in the context of task requirements.