Background: To develop a predictive model for the diagnosis of ROP in Indonesia. In the author's experience, no Indonesian ROP prediction model has been published. Early diagnosis of Retinopathy of Prematurity (ROP) is essential to prevent visual impairment and blindness in the preterm infant. The incidence of ROP reported approximately 65% of the infant with birth weight (BW) <1250 grams and represents a significant clinical problem accounting for up to 40% of childhood blindness worldwide. Many predictive models had been developed for ROP screening tools in the setting of high-income countries. However, the application somewhat irrelevant to the NICU service condition in the developing countries, including Indonesia. Methods: This retrospective cohort study was performed at Dr. Soetomo General Hospital Surabaya, Indonesia, from May 2009 to April 2011. The ophthalmologist performed binocular indirect ophthalmoscope (BIO) examination to the 248 preterm infants who admitted to the Neonatal Intensive care unit between May 2009 to April 2011. The potential predictors consisted of gestational age (GA), BW, oxygen therapy, sepsis, small for gestational age (SGA), blood transfusion, asphyxia, and respiratory distress syndrome (RDS) were analyzed. Results: Of the 248 preterm infants, 32(12.9%) were diagnosed with various stages of ROP. The model had three significant variables: oxygen therapy (OR= 36.93; 95% CI= 4.73 to 288.02), sepsis (OR= 4.86; 95% CI= 1.85 to 12.82), and SGA (OR= 3.99, 95% CI= 1.47 to 10.82). The area below the Receiver Operating Characteristics (ROC) curve was 90.4%. The score had a sensitivity of 89.0% and a specificity of 78.0%. Conclusions: This is the first Indonesian predictive model and a scoring system developed for the ROP screening tool of the preterm infant. This simpler tool was expected to reduce the number of ophthalmologic examinations required to detect ROP, especially in resource-limited settings.

BIO

premature infant

predictive model

ROP

scoring system