Classification of manual abilities in children with cerebral palsy under 5 years of age: how reliable is the Manual Ability Classification System?

Objective: To assess the interobserver reliability of the Manual Ability Classification System (MACS) in young children (age 1—5 years) with cerebral palsy. Design: Interobserver reliability study. Setting: A cross-sectional study of a hospital-based population of children with cerebral palsy. Subjects: Thirty children, 18 boys and 12 girls between 1 and 5 years of age (mean age 2.5 years ± 14.2 SD, Gross Motor Function Classification System level I—IV). Measures: the children were classified by means of the MACS by two independent observers. Interobserver reliability was analysed using Cohen's kappa. Results: Overall interobserver reliability of the MACS for children aged 1—5 years was moderate, with a linear weighted kappa (κ) of 0.62 (95% confidence interval (CI) 0.49—0.76). According to the generally accepted categories of agreement, reliability was moderate for children under 2 years of age (κ = 0.55), and good for children between 2 and 5 years of age (κ = 0.67). Conclusion: Classification of manual ability of young children with cerebral palsy is possible between 2 and 5 years of age. For children younger than 2 years old, it should be done with caution. Further development of the MACS for children under 5 years of age is recommended with an emphasis on age-appropriate descriptions of manual abilities.