Comparison of Different Risk Assessment Models for Predicting Postdischarge Phototherapy Requirement in Term and Late Preterm Neonates.

Abstract

This study was conducted to compare the Bhutani nomogram and the difference between the total serum bilirubin measurement at discharge and the AAP phototherapy threshold at the time of measurement (Δ-TSB) for predicting the postdischarge bilirubin level that will exceed the phototherapy threshold.Healthy neonates born at ≥35 weeks of gestation, followed in the newborn nursery, who did not receive phototherapy during hospital stay after birth, and who followed up in the outpatient clinic at least for 1 month from 2019 to 2024 were included in the study. Four logistic models were compared (D-TSB-plus,which encompassed additional variables such as gestational age, delivery type, and blood incompatibility, and DAT positivity, the Bhutani risk zones, and Bhutani-plus which encompassed additional variables such as gestational age, delivery type, and blood incompatibility) to predict postdischarge phototherapy requirement.Of the 2,040 neonates included in the study, 208 were readmitted for phototherapy treatment after discharge. ABO and/or Rh incompatibility, Bhutani risk zone, lower gestational age, and being born vaginally increased the need for phototherapy. Δ-TSB was the strongest predictor of postdischarge phototherapy requirement (p < 10^-50). Among the logistic models, the Δ-TSB-plus model had the highest predictive power (AUC: 0.83), followed by Δ-TSB alone (AUC: 0.82), Bhutani-plus (AUC: 0.80) and Bhutani alone (AUC: 0.74) models, respectively.Δ-TSB models had higher predictive ability regarding postdischarge phototherapy requirement compared with Bhutani models. Combining Bhutani risk zones with risk factors, especially with gestational age, increased the discrimination but did not reach the success of Δ-TSB models. · It is important to predict neonates likely to develop jaundice to start treatment in a timely manner.. · Combining bilirubin measurement with hemolysis findings and clinical parameters improves the prediction of postdischarge phototherapy.. · Δ-TSB model appears to be the strongest model for the prediction of postdischarge phototherapy requirement..