Abstract
PURPOSE: The purpose of this study was to determine the psychometric validity of the Visual Function Index (VF-14) for use by patients with Leber hereditary optic neuropathy (LHON). METHODS: Rasch analysis was conducted in two stages using data for 196 individuals (74.5% male) carrying one of the three primary LHON mutations and affected by vision loss. In stage 1, scale unidimensionality, scale-to-sample targeting, response category threshold ordering, item fit statistics, local dependency, and reliability were assessed. In stage 2, iterative post-hoc revisions of the VF-14 structure (VF-14R) were applied and psychometrically re-evaluated. RESULTS: Issues identified with the VF-14 included disordered response thresholds (12/14 items), local dependency (10/91 pairwise dependencies), and evidence of multidimensionality. However, the distribution of person estimates and item thresholds were fairly well matched, only one item showed misfit to the Rasch model, and there was good reliability (Person Separation Index 0.84). Rasch-informed VF-14 revisions included removing both driving items and the misfitting sports item, rescoring response options across all items by merging two response categories, and accounting for the dependency between two reading items. The VF-14R demonstrated improved psychometric validity. CONCLUSIONS: Clinicians and researchers using the VF-14 with LHON patients should be aware of its limitations. Compared to the original version, the proposed Rasch-based structure of the VF-14R appears to offer improved psychometric performance and interpretation of vision-related activity limitation. TRANSLATIONAL RELEVANCE: The original version of the VF-14 exhibits several limitations that undermines its psychometric validity as a patient-reported outcome measure for patients with LHON.