Validation of the East London Retinopathy of Prematurity algorithm to detect treatment-warranted retinopathy of prematurity: a cohort study.

Abstract

To validate the East London Retinopathy of Prematurity algorithm (EL-ROP) in a cohort of infants at risk of developing retinopathy of prematurity (ROP). The EL-ROP algorithm was applied retrospectively to routinely collected data from two tertiary neonatal units in England on infants eligible for ROP screening. The EL-ROP recommendation, to screen or not, was compared with the development of treatment-warranted ROP (TW-ROP) for each infant. The main outcome measures were (1) EL-ROP's sensitivity for predicting the future development of TW-ROP and (2) potential to reduce ROP screening examinations. Data from 568 infants were included in the trial. The median (IQR) birth weight (g) was 875 (704 - 1103) and gestational age (weeks) was 27.0 (25.4 - 29.0). Maternal ethnicity was black (33%) and non-black (67%). 58(10%) developed TW-ROP and in every case this was predicted by the EL-ROP algorithm. It's sensitivity was 100% (95% CI 94-100%) specificity: 44% (95% CI 39-48%) positive predictive value: 17% (95%CI 16-18%), negative predictive value: 100%. EL-ROP has been validated in a cohort of infants from two tertiary neonatal units in England. Further validation is required before its clinical usefulness can be assessed.