Reliability and sources of variation of the ABILHAND-Kids questionnaire in children with cerebral palsy

Abstract

© 2016 Informa UK Limited, trading as Taylor & Francis Group Purpose: To determine reliability of the ABILHAND-Kids, explore sources of variation associated with these measurement results, and generate repeatability coefficients. Method: A reliability study with a repeated measures design was performed in an ambulatory rehabilitation care department from a rehabilitation center, and a center for special education. A physician, an occupational therapist, and parents of 27 children with spastic cerebral palsy independently rated the children’s manual capacity when performing 21 standardized tasks of the ABILHAND-Kids from video recordings twice with a three week time interval (27 first-, and 25?second video recordings available). Parents additionally rated their children’s performance based on their own perception of their child's ability to perform manual activities in everyday life, resulting in eight ratings per child. Results: ABILHAND-Kids ratings were systematically different between observers, sessions, and rating method. Participant?×?observer interaction (66%) and residual variance (20%) contributed the most to error variance (9%). Test–retest reliability was 0.92. Repeatability coefficients (between 0.81 and 1.82 logit points) were largest for the parents’ performance-based ratings. Conclusion: ABILHAND-Kids scores can be reliably used as a performance- and capacity-based rating method across different raters. Parents’ performance-based ratings are less reliable than their capacity-based ratings. Resulting repeatability coefficients can be used to interpret ABILHAND-Kids ratings with more confidence.Implications for RehabilitationThe ABILHAND-Kids is a valuable tool to assess a child?s unimanual and bimanual upper limb activities.The reliability of the ABILHANDS-Kids is good across different observers as a performance- and capacity-based rating method.Parents' performance-based ratings are less reliable than their capacity-based ones.This study has generated repeatability coefficients for clinical decision making.