Visual Function Classification System for children with cerebral palsy: development and validation

Aim To develop and validate the Visual Function Classification System (VFCS), which was created to classify how children with cerebral palsy (CP) use visual abilities in daily life. Method The process of development and validation of the VFCS involved four phases: (1) drafting of the five levels from the analysis of literature and clinical experience; (2) validation of constructs and revision of the levels for concept meaningfulness, using nominal group process; (3) refinement by international Delphi survey; and(4) assessment of interrater reliability among professionals and with caregivers, and of test-retest reliability. Results Five nominal groups involved 29 participants; 65 people completed the first round and 51 the second round of the Delphi survey. Construct validity was demonstrated within an expert group and external validation through several stakeholders, with the involvement of patients and families to ensure meaningfulness of the concept. Discussions continued until consensus was reached about the construct and content of the five levels. Participants in the reliability study included 29 professionals, 39 parents, and a total sample of 160 children with CP (mean age [SD] 6y 6mo [3y 4mo]; median 5y 7mo, range 1-19y). Absolute interrater agreement among professionals was 86% (weighted kappa=0.88; 95% confidence interval [CI] 0.83-0.93). Test-retest reliability was high (weighted kappa=0.97; 95% CI 0.95-0.99). Parent-professional interrater reliability on 39 children was moderate (weighted kappa=0.51; 95% CI 0.39-0.63). Interpretation The VFCS has been appropriately constructed and provides a reliable system to classify visual abilities of children with CP both in clinical and in research settings. What this paper adds The Visual Function Classification System is a valid and reliable system. It classifies visual abilities of children with cerebral palsy in clinical and research settings. At a clinical level, it can be used to harmonize communication among professionals and identify patients' strengths and weaknesses. In research settings, it can be used to stratify patients, define natural history evolution, and interpret intervention studies.