Skip to main content
. Author manuscript; available in PMC: 2024 May 1.
Published in final edited form as: Ophthalmol Glaucoma. 2022 Nov 18;6(3):228–238. doi: 10.1016/j.ogla.2022.11.004

Table 2.

Model performance as a function of different demographics and clinical characteristics at baseline.

Variable AUC (95% confidence interval)
Disease status
 Suspect 0.914 (0.862, 0.953)
 Glaucoma 0.953 (0.933, 0.969)
Age, yearsa
 Younger than 51.9 0.982 (0.950, 0.996)
 Between 51.9 and 64.5 0.898 (0.847, 0.933)
 Between 64.5 and 77.1 0.943 (0.918, 0.962)
 Older than 77.1 0.955 (0.902, 0.99)
Sex
 Male 0.929 (0.901, 0.955)
 Female 0.942 (0.909, 0.961)
Race
 Black 0.895 (0.842, 0.939)
 White 0.945 (0.926, 0.961)
 Others 0.978 (0.940, 0.999)
Time from baseline, yearsb
 Less than 1.2 0.970 (0.942, 0.986)
 Between 1.2 and 2.3 0.911 (0.880, 0.938)
 Between 2.3 and 3.5 0.910 (0.878, 0.943)
 More than 3.5 0.912 (0.869, 0.940)
RNFL thickness at baseline, μmb
 Less than 74 0.974 (0.953, 0.989)
 Between 74 and 86 0.940 (0.904, 0.965)
 Between 86 and 97 0.927 (0.892, 0.958)
 More than 97 0.888 (0.842, 0.947)
SAP MD at baseline, dB
 Higher than −6 0.933 (0.911, 0.949)
 Between −6 and −12 0.940 (0.628, 1.000)
 Lower than −12 0.994 (0.985, 0.999)
a

Categories defined using mean − 1 SD, mean, mean + 1 SD as cut-offs.

b

Categories defined using p25, p50, p75 as cut-offs.