Utilizing machine learning could help identify patients at risk of neonatal jaundice requiring phototherapy, according to a study recently published in Children.
Neonatal jaundice is a common complication of hemolytic disease of the fetus and newborn (HDFN). Machine learning is a field of artificial intelligence that teaches computers to learn from data, identify patterns and make decisions or predictions without being explicitly programmed for each task.
Risk factors and predictors of neonatal jaundice
The authors aimed to analyse the risk factors for neonatal jaundice through a retrospective study from a single hospital in South Korea. Researchers analyzed electronic medical records from 8242 neonates admitted to the well-baby nursery between 2017 and 2022.
The goal was to determine maternal and neonatal predictors for jaundice requiring phototherapy.
Data processing included outlier replacement and use of the SMOTE-Tomek method to address class imbalance. Multiple machine learning algorithms were tested, with XGBoost showing the best performance.
The study found that 20.6% of the neonates required phototherapy. Significant risk factors included cesarean section delivery, weight loss over 5% of birth weight and lower birth weight. Infants in the phototherapy group also had higher frequencies of urination and defecation, and consumed more formula.
Maternal factors such as blood type O, gestational hypertension, prior miscarriages, higher BMI and elevated white blood cell count were also associated with increased risk.
“This study highlights the potential for developing a neonatal jaundice risk prediction model using machine learning. It is expected to serve as a valuable foundation for early identification and prevention strategies for neonatal jaundice,” the authors wrote.
Sign up here to get the latest news, perspectives, and information about HDFN sent directly to your inbox. Registration is free and only takes a minute.
