Functional visual fields: A cross-sectional UK study to determine which visual field paradigms best reflect difficulty with mobility function

Objectives: To develop an appropriate method of assessing visual field (VF) loss which reflects its functional consequences, this study aims to determine which method(s) of assessing visual fields best reflect mobility difficulty. Setting: This cross-sectional observational study took place within a single primary care setting. Participants attended a single session at a University Eye Clinic, Cambridge, UK, with data collected by a single researcher (HS), a qualified optometrist. Participants: 50 adult participants with peripheral field impairment were recruited for this study. Individuals with conditions not primarily affecting peripheral visual function, such as macular degeneration, were excluded from the study. Primary and secondary outcome measures: Participants undertook three custom and one standard binocular VF tests assessing visual field to 60 degrees, and also integrated monocular threshold 24-2 visual fields (IVF). Primary VF outcomes were average mean threshold, percentage of stimuli seen, and VF area. VF outcomes were compared to self-reported mobility function assessed with the Independent Mobility Questionnaire, and time taken and patient acceptability were also considered. Receiver Operating Characteristic (ROC) curves determined which tests best predicted difficulty with mobility tasks. Results: Greater VF loss was associated with greater self-reported mobility difficulty with all field paradigms (R2 0.38-0.48, all p<0.001). All four binocular tests were better than the IVF at predicting difficulty with at least three mobility tasks in ROC analysis. Mean duration of the tests ranged from 1min 26sec (±9sec) for kinetic assessment to 9min 23sec (±24 sec) for IVF. Conclusions: The binocular VF tests extending to 60 deg eccentricity all relate similarly to self-reported mobility function, and slightly better than integrated monocular VFs. A kinetic assessment of VF area is quicker than and as effective at predicting mobility function as static threshold assessment.