Intra-rater reliability of the cinematic assessment of the paretic upper-limb using a rehabilitation robot

Objective Rehabilitation robots are commonly used to help deliver high dosage of physical treatment in hemiparesis after stroke. Kinematic measurements are provided by some robotic devices. This study aims to measure the intra-rater reliability of robot-based kinematic assessments of the paretic upper limb in subacute stroke. Rehabilitation robots are commonly used to help deliver high dosage of physical treatment in hemiparesis after stroke. Kinematic measurements are provided by some robotic devices. This study aims to measure the intra-rater reliability of robot-based kinematic assessments of the paretic upper limb in subacute stroke. Material/patients and methods Twenty subjects with subacute stroke-induced hemiparesis (age 57 ± 18, time from stroke, 60 ± 41 days, mean ± SD) participated in this study. All subjects perform 2 pointing tasks, 24 hours apart, using the InMotion™ device, which involves a planar end-effector emphasizing shoulder and elbow movements from supported hand displacement in the horizontal plane. During each pointing task, the same investigator asks the subject to reach peripheral targets equally distant on a 14-cm radius circumference around the center target. Four robot-based kinematic measures were computed: distance covered, velocity, accuracy (RMS error from straight line) and smoothness (number of velocity peaks). Intra-class correlation coefficients (ICC), coefficient of variation (CV) and mean raw difference were used to analyze intra-rater reliability of kinematic parameters. Results Intra-rater reliability was good for velocity (ICC, 69 ± 19%), accuracy (ICC, 60 ± 24%) and smoothness (ICC, 55 ± 35%) except for distance covered (ICC, 7 ± 68%). Mean CVs were 7.6 ± 4.5% for distance covered, 19.8 ± 3.9% for velocity, 33.8 ± 5.9% for accuracy and 32.7 ± 5.3% for smoothness. Mean raw differences were 0.9 ± 0.3 cm for distance covered, 2.0 ± 0.5 cm/s for velocity, 0.5 ± 0.1 cm for accuracy and 2.2 ± 0.5 peaks for smoothness. Discussion – conclusion Kinematic assessment from the InMotion™ device has good intra-rater reliability in subjects with subacute hemiparesis. Relative homogeneity and ceiling effects of distances covered across subjects (many of the subjects reached the maximal 14 cm requested by the task) may be responsible for artificially low ICCs in this study. Kinematic analysis based on robot evaluation could be useful to monitor recovery after stroke.