OBJECTIVES: The NEI VFQ-28R, an alternative Rasch-based scoring of the National Eye Institute Visual Functioning Questionnaire (NEI VFQ) items, previously demonstrated improved psychometric performance and interpretability due to its cleaner conceptualization. The objective of this analysis was to provide confirmatory data on its psychometric performance in neovascular age-related macular degeneration (nAMD).

METHODS: The NEI VFQ-25 original and six supplemental items were collected in 1817 patients with nAMD from two clinical trials, HAWK (NCT02307682) and HARRIER (NCT02434328), leading to total of 8428 NEI VFQ observations. Rasch Measurement Theory (RMT) analysis was conducted on two domains of the NEI VFQ-28R, with similar item rescoring as in the original analysis: Activity limitations and Socio-emotional functioning. The analysis evaluated scale-to-sample targeting, dimensionality, reliability, item and person fit, local dependency, and invariance in subgroups (age,sex).

RESULTS: NEI VFQ items covered a wide range of visual functioning, but, in both domains, no items allowed to discriminate patients with high level of visual functioning (ceiling effect). Item rescoring defined in the NEI VFQ-28R scoring was working appropriately for all items of both domains. Reliability was good for activity limitations and modest for socio-emotional functioning (reliability coefficients of 0.89 and 0.78 respectively). Both domains showed only minor deviation to Rasch model in terms of item fit, local dependency, and invariance. The individual estimates of visual functioning from our RMT analysis were highly correlated to NEI VFQ-28R scores obtained with scoring algorithm from the original analysis (r>0.99), demonstrating stability of the results in this new sample.

CONCLUSIONS: The RMT results in a sample of nAMD patients were similar to those obtained originally in the development of NEI VFQ-28R score in a more heterogeneous sample: two domains of NEI VFQ-28R have adequate psychometric performance for measuring visual functioning in nAMD overall, but may not allow showing improvement in already highly functioning patients.