Abstract
Purpose:
The 150 year-old Snellen chart is a common test of visual function. Although able to describe a subject’s visual resolution (visual acuity), it falls short in its ability to summarize a subject’s visual function in “real world” situations. Visual targets are very rarely of 100% contrast and available for an unlimited viewing time. The test and results described herein, as well as the analytics used to summarize it’s results, more comprehensively estimate a subject’s visual function relative to viewing real world targets.
Methods:
500 subjects with normal eyes (except for uncorrected refractive errors) were enrolled in an IRB approved single-site, prospective, non-interventional protocol. All subjects responded to 260 randomly presented Landolt rings, each of which was a unique visual target with respect to the combination of three variables: varying in size between (logMAR between -0.2 to 1.0); contrast (between 10% and 90%); and presentation time (between 400 ms and unlimited). Subjects were stratified into 5 groups by age, ranging from 21 to >60 years. The sample was 64.8% female. Test-retest data were obtained on the initial test day and at re-testing 7 days later. Rasch analysis was used to estimate person measures (visual function variable) and item measures (stimulus variable) from the observed responses.
Results:
Analyses of the visual thresholds showed that the estimated Rasch person measure is a monotonic multivariable function of log size, log contrast, and exposure duration thresholds. Estimated size thresholds agreed with baseline Snellen visual acuity (expected). There was no effect of age or gender on estimated person measures. A Pearson correlation of 0.98 was found between measures collected on day 0 and day 7. Additionally, a Bland Altman plot demonstrated no evidence of a learning effect or other bias between day 0 and day 7.
Conclusions:
This study demonstrates the validity of manipulating stimulus size, exposure time and contrast to obtain a single summary measure of visual function that is responsive to stimulus thresholds and their interactions. Additionally, the study demonstrated stable and repeatable measures for subjects when tested a week apart. Clinical testing of subjects with this technique may produce a more relevant measure for prediction of visual abilities in real world situations in which targets are of variable size, contrast and viewing time.
Keywords: 753 vision and action •
669 quality of life •
754 visual acuity