Smartphone-Based Screening for Neonatal Hyperbilirubinemia in a U.S. Tertiary Center

Abstract Noninvasive screening for neonatal hyperbilirubinemia typically requires dedicated transcutaneous bilirubin devices, which may not be universally available. Smartphone-based screening tools offer a potential alternative but remain incompletely validated in U.S. clinical settings. We evaluated the accuracy of the Picterus Jaundice Pro smartphone application (app) compared with total serum bilirubin (TSB) in a diverse U.S. newborn population. A prospective, cross-sectional study of 74 newborns (gestational age ≥35 weeks, birth weight ≥1,800 g, age >21 hours to 14 days) admitted to the nursery or neonatal intensive care unit (NICU) was performed. Within 60 minutes of clinically indicated TSB collection, trained staff placed a calibration card on the infant's chest and used the smartphone Picterus Jaundice Pro app to obtain two scans (six images per scan). Images were analyzed by an algorithm to estimate bilirubin levels. Staff were blinded to application results. App estimates correlated with TSB (r = 0.72, p < 0.001). Using 2022 American Academy of Pediatrics (AAP) phototherapy criteria, the app correctly identified all infants requiring phototherapy (sensitivity 100%; 95% confidence interval [CI] 54.1–100%) with 78% specificity (95% CI 67.8–87.0%) and 100% negative predictive value (95% CI 94.2–100%). Bland–Altman analysis yielded a mean bias of −2.1 mg/dL (95% limits of agreement: −7.8 to +3.6 mg/dL). These results provide an indirect comparison with performance for currently reported transcutaneous devices. The Picterus Jaundice Pro app correlated with TSB and identified all infants requiring phototherapy based on current AAP thresholds, suggesting potential utility as a noninvasive smartphone-based screening tool for neonatal hyperbilirubinemia. Key Points