Background This study validates the newly developed WINROP algorithm aimed at detecting retinopathy of prematurity (ROP) requiring treatment at an early stage. The study was conducted at two middle-sized hospitals in Sweden, prospectively and retrospectively.

Methods A total of 104 children participated in this study. Their mean gestational age at birth was 28.7 weeks (range, 23.6–32.1 weeks), and their mean birth weight was 1208 g (range, 477–2340 g). Weekly weight measurements were used in WINROP to calculate the risk of developing ROP.

Results 80% of infants (83/104) had no ROP, 15% (16/104) had mild ROP (stage 1 or 2), 5% (5/104) had severe ROP, and 2% (2/104) were treated for ROP. The alarm was registered at an average of 2 weeks postnatal age (range 1–6 weeks).

Conclusions WINROP identified all the infants at risk for developing stage 3 ROP (100% sensitivity) and had a 59% specificity. The alarm was registered several weeks before screening for ROP began. WINROP can be used to complement conventional ROP screening.