Interrater and Retest Reliability of Mulit-Joint Upper Limb Position Sense in Children

Thumbnail Image
Henderson, Carla Yvonne
Proprioception , Robot-Based Assessment , Paediatrics , Reliability , Sensory-Motor Control , Neurodevelopmental Disorders , Development , KINARM
The contribution of deficits in limb position sense to the motor impairments of children with cerebral palsy, as well as other neurodevelopmental disorders, is increasingly being recognized. A more complete understanding of the development of multi-joint upper limb position sense is needed and has been limited, to date, by the absence of a reliable measurement technique to produce clinically meaningful information. The KINARM Exoskeleton’s bilateral position matching task, which involves passive movement of one of the subject’s arms to one of eight positions requiring different combinations of elbow and shoulder positions and active matching by the participant’s other arm, was evaluated for interrater and retest reliability. Intraclass correlation coefficients, absolute difference, minimum detectable difference that would be considered a significant change in performance, standard error of the measure, coefficient of variation, index of reliability, limit of agreement and confidence intervals were used to determine reliability on three measures of multi-joint position sense: (1) inter-trial variability in end-point position, (2) the ratio between actual and matched position, or spatial contraction/expansion, which provides a measure of the absolute accuracy of position matching, and (3) systematic errors in matching. Interrater index of reliability was very good to excellent with values of 72% for systematic errors in matching to 93% for contraction/expansion. Interrater intraclass correlation values were fair to excellent at 0.46 for systematic errors in matching to 0.81 for contraction/expansion. Standard errors in measurement were low and ranged from 0.002 to 0.06, for inter-trial variability and contraction/expansion respectively. Similarly, minimal detectable difference values for retest reliability ranged from 0.005 for inter-trial variability to 0.161 for contraction/expansion. Retest intraclass correlation values were fair to excellent at 0.38 for systematic errors in matching to 0.82 for contraction/expansion. Moderate to strong interrater and retest reliability and high measurement precision support the use of robot-based assessment of multi-joint position sense for developmental studies and promises to be a reliable clinical and research tool in the advancement of knowledge on sensory-motor coordination difficulties in children with neurodevelopmental disorders.
External DOI