Table 2.

Comparison between groups with small and large prediction errors, respectively, for mean deviation.

	Training/validation set			Testing set
	Small prediction error*	Large prediction error	P value	Small prediction error*	Large prediction error	P value
No. of images	944	56	N/A	190	10	N/A
Age (years)	57.8 ± 13.6	57.4 ± 14.1	0.823	54.5 ± 8.7	59.0 ± 9.3	0.097
Sex (Female, %)	144 (58.3%)	19 (52.8%)	0.656	14 (41.2%)	2 (18.2%)	0.279
IOP (mmHg)	14.3 ± 3.0	15.8 ± 4.3	0.036	14.0 ± 3.3	14.9 ± 2.6	0.218
SE (D)	− 2.6 ± 3.5	− 3.5 ± 3.4	0.130	− 2.8 ± 3.0	− 1.5 ± 3.1	0.067
CCT (μm)	540.1 ± 34.4	535.5 ± 33.9	0.454	534 ± 36.1	529 ± 39.9	0.788
SAP MD (dB)	− 2.1 ± 3.7	− 8.0 ± 6.3	< 0.001	− 1.3 ± 2.9	− 2.1 ± 2.6	0.404
SAP PSD (dB)	4.1 ± 3.8	9.8 ± 5.5	< 0.001	3.7 ± 2.8	4.0 ± 3.1	0.691
Moderate-to-severe stage†	127 (13.5%)	39 (69.6%)	< 0.001	18 (9.5%)	2 (18.1%)	0.315

dB, Decibels; D, Diopters; CCT, central corneal thickness; SAP, standard automated perimetry; MD, mean deviation; PSD, pattern standard deviation.

*The group of which prediction error is equal to or smaller than 95% limits of agreement.

^†MD lower than − 6 dB.