Abstract
PURPOSE: Develop a short form of the National Eye Institute Visual Function Questionnaire-25 (NEI VFQ-25) using Rasch methodology in patients with neovascular age-related macular degeneration. METHODS: Data on 32 items of the NEI VFQ-25 were analyzed from 1,294 patients in 3 studies of ranibizumab in neovascular age-related macular degeneration. Items were first grouped by visual function type; iterative multiple-group rating scale Rasch analyses were used to derive the Visual Function Questionnaire-Short Form (VFQ-SF). Item and scale-level VFQ-SF psychometric properties were assessed with estimation of meaningful within-patient change. RESULTS: Fourteen of 32 items were eliminated initially; the final optimal subset included 7 items measuring 5 domains. The seven-item VFQ-SF had strong internal consistency (average item-total correlation 0.67, Cronbach alpha 0.88), test-retest reliability (intraclass correlation coefficient 0.79-0.91), and high NEI VFQ-25 composite score correlation (r = 0.93). Convergent (mean r = 0.76) and divergent (mean r = 0.24) validity were demonstrated. Known-groups validity was shown by mean VFQ-SF scores monotonically increasing with best-corrected visual acuity severity categories with large between group effect sizes (>2.8). Responsiveness was confirmed with VFQ-SF change linearly related to best-corrected visual acuity change. A meaningful within-patient change estimate of seven points was derived. CONCLUSION: The seven-item VFQ-SF measures composite vision-related functioning, met all psychometric benchmarks, and is fit-for-purpose in neovascular age-related macular degeneration patients.