Assessing upper extremity skills in children with hemiplegic cerebral palsy: the agreement among examiners

Background: The evaluation of primary and secondary outcomes in clinical research in the field of rehabilitation may be hampered by the limitation of the measeurement tools and their administration by assessors. Evaluate the interobserver and intraobserver reliability of assessments of disabilities is fundamental. Aim: This study aims to evaluate the reliability of rating tests, to assess interobserver reproducibility (agreement and reliability) of upper limb function and to study the influence of clinical characteristics on reproducibility. Materials and Methods: Two tests were adminstered to 44 children, administration was videorecorded, 21 physical therapists were asked to rate videos. Both tests (Quality of Upper Extremity Skills Test (QUEST) and Besta Scale) aim at evaluating the upper extremity function . For each case information was collected on age and severty (mild/medium/severe). Agreement was calculated as the mean difference between examiners +/-1.96 x standard deviations. For QUEST the intraclass correlation coefficient (ICC) was calculated as a measure of reliability. For Besta Scale reliability was assessed by Cohen\ub4s kappa. Results: The highest level of agreement was for dissociated movement and for grasp and was increasing for older children and for more severe cases. The lowest level of agreement was for protective extension, and weight bearing. Conclusion: The measurements tools for the rating of upper limb impairment examined are only partially valid and reliable to assess the spontaneous and non-spontaneous use of affected limb in children with hemiplegic cerebral palsy. Further analysis is need to evaluate their usefulness in clinical research and clinical practice