Abstract
Purpose :
There is a pressing need for an accessible color blindness screening test that quantifies the individualistic points along the color vision spectrum. To this end, we developed a digital color blind assessment to discriminate between normative color vision (NCV) and color deficient vision (CDV) using error margins. This pilot study examines the potential of the tool to quantitatively identify color deficiencies based on proportional error margins.
Methods :
Male individuals aged 18-22 with normative color vision (control, n=20) and clinically diagnosed acute color blindness (n=20) were asked to match the color of two rings on a mobile device, separated by negative black space, by using a color wheel ten consecutive times while proctored. The outermost ring (Ring 1) depicted a given prompt color and the innermost ring (Ring 2) adjusted to show the user’s response color as they attempted to match them using the color wheel. The sizes of which were constrained to the following parameters: wheel 17.23-25.81 mm, space 19.86-29.25 mm, Ring 1 20.67-30.97 mm, and Ring 2 19.04-28.52 mm. Utilizing the prompt color value and the user’s response values of red, green, and blue on the RGB spectrum, test results were then weighted by the proportion of the color value present. We employed logistic regression to predict red-green color blindness (deuteranopia) using the user’s corrected red error. This survey aligned with the Declaration of Helsinki and posed no risk to participants.
Results :
The logistic regression showed the statistical significance (p<0.05) of four principal components used by the screening model to distinguish between NCV and CDV, specifically red-green color blindness [(95% CI=0.013-0.157; p=0.02) (95% CI=0.03-0.219; p=0.043) (95% CI=0.016-0.311; p=0.03) (95% CI=0.012-0.217; p=0.028)]. The in-sample performance, denoting the accuracy of the trueHue model to identify CDV and NCV, proved to be 95%. This analysis demonstrated the effectiveness of error margin analysis for the future distinction and identification of the aforementioned groups.
Conclusions :
This pilot study demonstrates the potential utility of a digital color blindness assessment to quantitatively distinguish between CDV and NCV. Limitations of this study include the sample size and lack of genetically verified study participants. Future studies will be required to compare the testing accuracy of this method with previously validated methods.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.