NEO-SEPSIS: Early Detection of Neonatal Sepsis Using Clinical and Physiological Data

Neonatal early-onset sepsis (EOS) is a life-threatening medical condition that affects newborns. EOS involves an extreme inflammatory response to infection and remains a major cause of morbidity and mortality in newborns. Current clinical assessments rely on subtle and often nonspecific signs, such as lethargy, hypothermia and poor feeding, leading to delayed diagnoses and unnecessary antibiotic treatments. To overcome these limitations, this project aims to create a machine learning-based method capable of detecting the conditions leading to sepsis onset. By identifying newborns at risk of developing sepsis and triggering a warning alert, the main goal of this project is to decrease the time to antibiotic therapy and reduce the mortality rate. By integrating structured clinical data with physiological signals and combining model interpretability with robust evaluation in clinical workflows, this computational tool will support neonatologists in making safer, faster, and more precise diagnostic decisions.