Related Experiment Video
Updated: Sep 10, 2025
How to Administer Near-Infrared Spectroscopy in Critically ill Neonates, Infants, and Children
Published on: August 19, 2020
Neonatal Jaundice Requiring Phototherapy Risk Factors in a Newborn Nursery: Machine Learning Approach
Yunjin Choi1, Sunyoung Park1, Hyungbok Lee1
1Nursing Department, Seoul National University Hospital, Seoul 03038, Republic of Korea.
Insights
Machine learning effectively identifies newborns needing phototherapy for jaundice. Key predictors include delivery mode, feeding patterns, and maternal factors, aiding early intervention for better infant outcomes.
Area of Science:
- Neonatal Medicine
- Data Science
- Predictive Analytics
Background:
- Neonatal jaundice is a prevalent condition that can lead to severe hyperbilirubinemia if not managed promptly.
- Early identification of neonates at high risk for jaundice remains a clinical challenge despite current guidelines.
- Existing diagnostic methods may not capture the full spectrum of risk factors for severe neonatal jaundice.
Purpose of the Study:
- To pinpoint critical maternal and neonatal risk factors associated with neonatal jaundice requiring phototherapy.
- To develop and validate a machine learning model for predicting the need for phototherapy in newborns.
- To enhance early clinical decision-making for managing neonatal hyperbilirubinemia.
Main Methods:
- Retrospective analysis of electronic medical records for 8242 neonates from 2017-2022.
- Application of machine learning algorithms, including XGBoost, to predict phototherapy requirements.
- Utilized SHAP values for interpreting the predictive model and identifying key risk factors.
Main Results:
- Mode of delivery, neonatal feeding indicators (formula intake, breastfeeding frequency), maternal BMI, and maternal white blood cell count were identified as significant predictors.
- Cesarean delivery and lower birth weight were associated with an increased likelihood of requiring phototherapy.
- The XGBoost model achieved a high predictive performance with an AUROC of 0.911.
Conclusions:
- Machine learning models trained on perinatal data can accurately predict the risk of neonatal jaundice requiring phototherapy.
- These predictive models offer a valuable tool for early clinical intervention, potentially improving infant health outcomes.
- Integrating machine learning into clinical practice can support timely decisions regarding phototherapy for neonatal jaundice.
Abstract:
Background: Neonatal jaundice is common and can cause severe hyperbilirubinemia if untreated. The early identification of at-risk newborns is challenging despite the existing guidelines. Objective: This study aimed to identify the key maternal and neonatal risk factors for jaundice requiring phototherapy using machine learning. Methods: In this study hospital, phototherapy was administered following the American Academy of Pediatrics (AAP) guidelines when a neonate's transcutaneous bilirubin level was in the high-risk zone. To identify the risk factors for phototherapy, we retrospectively analyzed the electronic medical records of 8242 neonates admitted between 2017 and 2022. Predictive models were trained using maternal and neonatal data. XGBoost showed the best performance (AUROC = 0.911). SHAP values interpreted the model. Results: Mode of delivery, neonatal feeding indicators (including daily formula intake and breastfeeding frequency), maternal BMI, and maternal white blood cell count were strong predictors. Cesarean delivery and lower birth weight were linked to treatment need. Conclusions: Machine learning models using perinatal data accurately predict the risk of neonatal jaundice requiring phototherapy, potentially aiding early clinical decisions and improving outcomes.
