Evaluation of a smartphone photoscreening app to detect refractive amblyopia risk factors in children aged 1–6 years

Figures & data

Table 1 Patient demographics

	Age 1–3 (n = 132)		Age 4–6 (n = 155)		Total (N = 287)
	n	%	n	%	n	%
Race/Ethnicity
African American	7	5.3	6	3.9	13	4.5
Asian	19	14.4	19	12.3	38	13.2
Asian/Caucasian	0	0.00	3	1.9	3	1.1
Asian/Hispanic	0	0.00	1	0.7	1	0.4
Asian/other	1	0.8	1	0.7	2	0.7
Caucasian	74	56.1	77	49.7	151	52.6
Caucasian/Hispanic	1	0.8	1	0.7	2	0.7
Hispanic	17	12.9	29	18.7	46	16.0
Indian	0	0.00	1	0.7	1	0.4
Native American	7	5.3	4	2.6	11	3.8
Pacific Islander	3	2.3	3	1.9	6	2.1
Other	3	2.3	10	6.5	13	4.5
Gender
Male	68	51.5	74	47.7	142	49.5
Female	64	48.5	81	52.3	145	50.5

Table 2 Reasons for rejecting captured images

Image rejection reason	iPhone 7 portrait image errors (auto)	iPhone 7 portrait image errors (manual)
Blurry image	3	14
Bad gaze	21	11
Unable to identify pupil	6	5
Test distance error	7	2

Table 3 Combined sensitivity and specificity

	Manual grading, age 1–6	Auto grading, age 1–6
Sensitivity	76%	65%
Specificity	85%	83%
Positive Predictive Value	0.76	0.69
Negative Predictive Value	0.85	0.80