Assessment of WINROP algorithm as screening tool for preterm infants in Manitoba to detect retinopathy of prematurity

Abstract Objective: Developing less invasive methods for early detection of retinopathy of prematurity (ROP) is vital to minimizing blindness in premature infants. Lofqvist and colleagues developed a computer-based ROP risk algorithm (WINROP) (https://winrop.com), which detects downtrends in postnatal weight gain that correlate with the development of sight-threatening ROP. The aim of this study is to investigate the sensitivity and specificity of the WINROP algorithm to detect vision-threatening ROP. Methods: This is a retrospective chart review study between January 2008 and December 2013. This study was conducted in the neonatal intensive care unit in Children’s Hospital at Health Sciences Centre, Winnipeg, Manitoba, Canada. The study included preterm infants, less than 32 weeks’ gestation, who were admitted to the hospital during the study period. The included 215 infants were eligible for ROP screening and had sufficient data to be entered into the WINROP algorithm. Infants were screened by a paediatric ophthalmologist for retinopathy of prematurity. The body weight of infants was measured weekly and entered into the WINROP algorithm; the sensitivity and the specificity of the WINROP algorithm were assessed. Results: The mean gestational age was 28.6 ± 1.8 weeks. The mean body weight was 1244 ± 294 g. The sensitivity of the WINROP algorithm to detect vision-threatening retinopathy of prematurity in our cohort was 90% (P=0.021) with a specificity of 60% (P=0.002). Conclusion: The WINROP algorithm lacks sufficient sensitivity to be used clinically in our population. The algorithm needs to be reassessed in contemporary populations.