Abstract
Purpose :
Ultra-low vision (ULV) refers to a level of visual acuity worse than 20/1600, where one's vision is limited to, at best, seeing moving shadows. Currently, there are limited tools available to assess visual performance in people with ULV. The ULV-VFQ is a questionnaire developed to assess self-reported ability in people with ULV. However, it is not known how well their self-report relates to task performance. In this study, we compare visual ability assessed with ULV-VFQ with performance on a range of tasks involving visual information gathering in people with ULV.
Methods :
Forty-three participants with ULV answered the 150-item ULV-VFQ and performed 19 m-alternative (2≤m≤4) forced choice activities assessing visual information gathering, presented in virtual reality; each activity was presented at 3 difficulty levels, each repeated 3 times, and scored as 1 (correct) or 0 (incorrect). Item (difficulty) and person (ability) measures for this “Wilmer VRI” were estimated using signal detection theory. ULV-VFQ scores ranged from and 1 (not difficult) to 5 (extremely difficult), with the option to score 0 (not applicable). Item and person measures were estimated using the method of successive dichotomizations. Person measures from both tests were compared to assess agreement.
Results :
Estimated item measures for the 150 scenarios ranged from -2.24 to 2.89 logits; the scenario with the least reported difficulty was locating a window in a dark room on a sunny day, while the scenario with the most reported difficulty was spotting a deer a distance away on a sunny day. Person measures ranged from -3.69 to 5.53 logits. In comparison, estimated person measures for the Wilmer VRI ranged from -1.80 to 5.22 logits. There was a moderate (R2 = 0.35) but significant (p < 0.001) association between the person measures assessed from both the tests.
Conclusions :
Both instruments have varied everyday scenarios with a a wide range of difficulty that can assess visual ability in individuals with ULV, ranging from bare light perception to 20/1600. The significant association between self-reported ability and task performance can be useful for rehabilitation in vision restoration trials, as improving task performance through training should correspond to an improvement in one's perceived ability, leading to improved quality of life.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.