Researchers from KTH, in collaboration with Neonatal care at Karolinska University Hospital, have developed an advanced AI system that can detect infections in patients, particularly premature babies, as early as 24 hours before conventional symptoms appear. The project “Explainable Machine Learning for Early Warning Systems” is funded by Digital Futures and led by Saikat Chatterjee, Associate Professor, Division of Information Science and Engineering at KTH and member of Digital Futures Faculty. Since infection symptoms can be subtle and delayed, the technology, which uses sensors connected to hospital beds, allows healthcare providers to initiate treatment much earlier, potentially saving lives. The AI system monitors key physiological parameters, providing valuable insights into the body’s response to infection.
The goal is to alert doctors three days before symptoms manifest, enabling more effective and timely intervention. The technology not only prevents life-threatening conditions like sepsis but also predicts the type of infection likely to occur. While the aim is not to replace doctors, these AI tools are seen as crucial additions to healthcare, offering valuable support to medical professionals in providing optimal patient care.
Radio interview with Saikat Chatterjee on Sveriges Radio P4
More in this interview with Saikat Chatterjee on KTH website
Photo: Praisaeng/Mostphotos
