Abstract
Purpose :
The VRFV-Nav (Fig. 1) was developed to assess functional vision in patients with inherited retinal diseases or other causes of low vision. The VRFV-Nav allows for limited setup, instant course configuration changes, and objective/automated scoring. We evaluated the construct validity (i.e. ability to distinguish between normal sighted and software-simulated visually impaired subjects) and the coefficient of repeatability (CR; i.e. maximum difference between repeated measurements) of the VRFV-Nav.
Methods :
The HAT Study was a single-masked, randomized, pilot study designed to evaluate the VRFV-Nav in normal sighted subjects. 30 subjects were randomized to 1 of 6 settings for varying degrees of software-simulated visual impairment (Fig. 2). Subjects were dark adapted before they were asked to navigate the VRFV-Nav course at increasing levels of luminance (measured in lux). Test-retest visits were conducted in Weeks 1 and 2, followed by a final visit in Week 3. The CR was calculated for each lux level and for each impairment condition by lux level.
Results :
Subjects with simulated visual impairment took longer to complete the course than subjects with no impairment. For the Week 1 test-retest assessment, the CR was 33.5, 60.5, and 54.8 seconds, whereas for Week 2, the CR was 24.2, 30.7, and 28.5 seconds for 368 lux, 29 lux and 4 lux, respectively. When grouping by each impairment condition, the Week 2 CR ranged from 11.0 (no impairment at 368 lux) to 47.9 seconds (20/800+12.5% LT+ no tunnel vision at 49 lux). The CR could not be computed for the higher levels of impairment at the lowest lux level, because no subjects were able to complete the course under those conditions.
Conclusions :
Increased time to complete the course with increasing levels of impairment supports the construct validity of the VRFV-Nav. The CR was generally smaller in Week 2 than in Week 1, indicating that there may have been an initial learning effect after the first test, therefore, data from Week 2 are more representative. The CR was smallest for subjects with no impairment, indicating that there was no/minimal learning effect in subjects that did not need to cope with simulated impairment. Under the conditions tested, a change greater than 30 seconds could be interpreted as a meaningful difference in time to complete the course.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.