A newly developed mobile application is set to revolutionize how parents monitor their newborns for jaundice, a common condition that affects a significant number of infants. Jaundice causes a yellowish tint in the skin and eyes and is present in approximately 60% of full-term babies and up to 80% of premature infants.
The app, named BiliSG, allows parents to easily screen their babies for neonatal jaundice from the comfort of their own homes. Unlike traditional methods, which involve using a light-emitting handheld device or a blood sample from a heel prick, BiliSG uses a machine learning model to accurately predict bilirubin levels by analyzing skin color.
The algorithm works by assessing the yellowness of a baby’s skin through photos taken of the baby’s forehead, chest, and abdomen. Special color-calibration stickers are used to ensure accurate measurements. The app is designed to accommodate babies of varying ethnicities and skin tones, including both lighter and darker shades, primarily focusing on Asian populations.
Bilirubin, a yellow pigment produced during the breakdown of red blood cells, is typically processed by the liver. However, in newborns, the liver may not yet be fully developed, leading to elevated bilirubin levels and the characteristic yellow hue. While jaundice is usually mild and resolves within the first two to three weeks of life, severe cases can pose serious risks, including permanent brain damage if left untreated.
The BiliSG app, developed by researchers from Singapore General Hospital (SGH) in collaboration with SingHealth Polyclinics (SHP) and national health tech agency Synapxe, was tested in a study involving nearly 550 babies from November 2023 to June 2024. The results, published in December 2024 in the peer-reviewed JAMA Network Open, show that the app was able to accurately detect significant cases of jaundice, comparable to traditional testing methods.
Clinical Assistant Professor Alvin Ngeow, senior consultant at SGH’s neonatal and developmental medicine department, explained that while other jaundice-screening apps exist, many rely on a single reference point—typically the forehead—to determine the condition. He noted that jaundice can vary in intensity across different parts of the body, making multi-point analysis more reliable.
Professor Ngeow, who also led the study, pointed out that some existing apps may be less effective in Singapore due to differences in skin tones between populations, which could impact the accuracy of their results.
SHP research director Tan Ngiap Chuan, the study’s senior author, described the app as a “game changer” in infant healthcare. “It will transform the way we protect the brain health of babies by enabling parents to closely monitor jaundice levels at home, with the added oversight of their family doctors via telemedicine,” said Tan, who is also the vice-chair of research at SingHealth-Duke NUS Family Medicine Academic Clinical Programme.
While the app has been tested on a single smartphone model, further comprehensive testing is planned over the next 18 months. During this period, BiliSG will be trialed on a wider range of smartphones, with more diverse camera specifications and on a broader group of babies. The team also aims to expand the app’s ability to analyze an even wider variety of skin tones.
Synapxe’s chief data officer, Andy Ta, emphasized the app’s potential to assist parents in monitoring jaundice with the help of AI technologies. “While still in the testing phase, the development of the BiliSG app showcases how artificial intelligence can be harnessed to support parents in caring for their infants,” he said.
BiliSG is expected to be commercially available within the next two years, bringing the convenience of home-based jaundice monitoring to parents everywhere.
Related topics
