Feasibility and reliability of classifying gross motor function among children with cerebral palsy using population-based record surveillance

Summary Benedict RE, Patz J, Maenner MJ, Arneson CL, Yeargin‐Allsopp M, Doernberg NS, Van Naarden Braun K, Kirby RS, Durkin MS. Feasibility and reliability of classifying gross motor function among children with cerebral palsy using population‐based record surveillance. Paediatric and Perinatal Epidemiology 2010. For conditions with wide‐ranging consequences, such as cerebral palsy (CP), population‐based surveillance provides an estimate of the prevalence of case status but only the broadest understanding of the impact of the condition on children, families or society. Beyond case status, information regarding health, functional skills and participation is necessary to fully appreciate the consequences of the condition. The purpose of this study was to assess the feasibility and reliability of enhancing population‐based surveillance by classifying gross motor function (GMF) from information available in medical records of children with CP. We assessed inter‐rater reliability of two GMF classification methods, one the Gross Motor Function Classification System (GMFCS) and the other a 3‐category classification of walking ability: (1) independently, (2) with handheld mobility device, or (3) limited or none. Two qualified clinicians independently reviewed ed evaluations from medical records of 8‐year‐old children residing in southeast Wisconsin, USA who were identified as having CP (n = 154) through the Centers for Disease Control and Prevention's Autism and Developmental Disabilities Monitoring Network. Ninety per cent (n = 138) of the children with CP had information in the record after age 4 years and 108 (70%) had adequate descriptions of gross motor skills to classify using the GMFCS. Agreement was achieved on 75.0% of the GMFCS ratings (simple kappa = 0.67, 95% confidence interval [95% CI 0.57, 0.78], weighted kappa = 0.83, [95% CI 0.77, 0.89]). Among case children for whom walking ability could be classified (n = 117), approximately half walked independently without devices and one‐third had limited or no walking ability. Across walking ability categories, agreement was reached for 94% (simple kappa = 0.90, [95% CI 0.82, 0.96], weighted kappa = 0.94, [95% CI 0.89, 0.98]). Classifying GMF in the context of active records‐based surveillance is feasible and reliable. Future surveillance efforts that include functional level among children with cerebral palsy may provide important information for monitoring the impact of the condition for programmatic and p