Table 3

Comparison of DIGIROP-Screen versus other existing ROP prediction models

Comparison model and time pointN/N
Specificity (95% CI) (%)
N/N
Sensitivity (95% CI) (%)
DIGIROP-ScreenComparison ROP modelDIGIROP-ScreenComparison ROP model
DIGIROP-Screen (up to PNA 8 w) vs CHOP-ROP17 (up to PNA 8 w)278/571
48.7
(44.5 to 52.9)
157/571
27.5
(23.9 to 31.4)
26/26
100.0
(86.8 to 100.0)
26/26
100.0
(86.8 to 100.0)
DIGIROP-Screen (up to PNA 12 w) vs CHOP-ROP17 (up to PNA 12 w)362/569
63.6
(59.5 to 67.6)
159/569
27.9
(24.3 to 31.8)
26/26
100.0
(86.8 to 100.0)
26/26
100.0
(86.8 to 100.0)
DIGIROP-Screen (up to PMA 36 w) vs OMA-ROP18 (up to PMA 36 w)250/541
46.2
(41.9 to 50.5)
206/541
38.1
(34.0 to 42.3)
24/25
96.0
(79.6 to 99.9)
24/25
96.0
(79.6 to 99.9)
DIGIROP-Screen (up to WINROP risk flag or last measurement) vs
WINROP12
256/568
45.1
(40.9 to 49.3)
257/568
45.2
(41.1 to 49.4)
25/26
96.2
(80.4 to 99.9)
23/26
88.5
(69.8 to 97.6)
DIGIROP-Screen (at birth) vs
CO-ROP16 (at PNA 4 w)
231/564
41.0
(36.9 to 45.1)
54/564
9.6
(7.3 to 12.3)
25/26
96.2
(80.4 to 99.9)
25/26
96.2
(80.4 to 99.9)
  • CHOP-ROP, Children’s Hospital of Philadelphia-ROP; CO-ROP, Colorado-ROP; OMA-ROP, Omaha ROP; PMA, postmenstrual age; PNA, postnatal age; ROP, retinopathy of prematurity; w, weeks; WINROP, weight, insulin-like growth factor 1, neonatal, ROP.