Purchase this article with an account.
R. C. Becker, G. Teichler, M. Graef; Influence of Interpolated Log Mar Scoring on Test-Retest Variability of ETDRS Visual Acuity. Invest. Ophthalmol. Vis. Sci. 2008;49(13):2582. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Test-retest-variability (TRV) of a visual acuity test determines the ability of the test to detect true clinical change. The purpose of this study was to investigate whether a relationship exists between the scoring method (threshold vs. interpolation, i.e. line-by-line vs. letter-by-letter) and TRV of ETDRS visual acuity.
95 patients (age 8-90 years, median 60.5) with various eye disorders (39 strabismus amblyopia, 24 cataract, 32 retinal disease) and 13 healthy volunteers (age 18-33 years, median 24) were tested. Visual acuity assessment was performed using ETDRS 1 & 2 charts. Both charts use the 10 letters described by Sloan in different combinations of five letters. 3 out of 5 optotypes per line had to be correctly identified for line-by-line-scoring, while a 0.02-logMAR score per letter was assigned for interpolation. In the patients group, the eyes with the lower visual acuity were evaluated, and for the healthy subjects, the right eyes. All units of acuity are given in logMAR notation. TRV was quantified in terms of its 95% range using the approach of Bland and Altman. P-values < 0.05 were considered statistically significant.
While TRV was lower for letter-by-letter counting, there was no statistically significant difference in all groups.entire group (p=0.7):line-by-line: 0.06 ± 0.07 (95% TRV = 0.14)letter-by-letter: 0.06 ± 0.05 (95% TRV = 0.10)cataract (p=0.96):line-by-line: 0.07 ± 0.08 (95% TRV = 0.16)letter-by-letter: 0.07 ± 0.06 (95% TRV = 0.12)retinal disease (p=0.28):line-by-line: 0.05 ± 0.07 (95% TRV = 0.13)letter-by-letter: 0.06 ± 0.05 (95% TRV = 0.10)strabismus amblyopia (p=0.98):line-by-line: 0.06 ± 0.06 (95% TRV = 0.12)letter-by-letter: 0.06 ± 0.04 (95% TRV = 0.08)
There was no significant difference between threshold scoring and interpolation. This is in accordance with previous reports of both scoring methods suggesting that differences between the two are small (Carkeet 2001, Optom Vis Sci; Chen 2006, IOVS). Existing differences can be explained by the smaller scale increment of interpolated scoring. Thus both methods can be used.
This PDF is available to Subscribers Only