The eye of the beholder: understanding discrepancy between visual acuity and self-reported vision in age-related-macular degeneration.

Abstract

Purpose : Although visual acuity (VA) is used during clinical care of patients with age-related macular degeneration (AMD), AMD patients often report vision problems out of proportion with their VA. Our objective was to estimate the prevalence of discrepancy between VA and self-reported vision (SRV) and to evaluate factors that may contribute to the discrepancy.

Methods : We analyzed data from an observational study. Distance VA was measured with Snellen chart and converted to logMAR scale. SRV was assessed with a single question rating eyesight with both eyes on a scale of 0 to 5 (0=completely blind, 5=excellent). VA/SRV discrepancy was defined consistent with previous literature (VA≤0.3 SRV≥4 or VA>0.3 SRV<4). Depressive symptoms were assessed with the geriatric depression screen. Contrast sensitivity (CS) was assessed by Pelli-Robinson chart. Cognition was assessed with the Brief Test of Adult Cognition by Telephone. Type of AMD (wet vs dry) was abstracted from medical charts. Correlation analyses between VA and SRV and separately between CS and SRV were done using Spearman rank correlation, and Fisher r-to-z transformation was used to compare the correlation coefficients. Group differences were assessed with Chi-square tests (for proportions) and t tests (for continuous variables).

Results : Of 56 participants with AMD, 21 (38%) displayed VA/SRV discrepancy and of those, 19 (90.5%) reported SRV worse than VA. Discrepant and non-discrepant participants did not significantly differ based on age, gender, race, years of education, depression score, cognition or AMD type (wet vs. dry). The correlation coefficient between VA and SRV was -0.58, and between SRV and CS was 0.53. There was no significant difference between the absolute value of the two correlation coefficients (z=-0.36, p=0.36).

Conclusions : More than 1 in 3 AMD participants exhibited VA/SRV discrepancy, and the majority of discrepant individuals perceived worse vision than suggested by VA measurement. The discrepancy was not explained by any of the factors investigated. Both VA and CS were moderately correlated to SRV, but CS was not a superior indicator of SRV. Future research is needed to investigate potential contributors to VA/SRV discrepancy not examined here, such as near visual acuity, timing and course of visual changes, and patient expectations.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.