Visualization of Heart Rate Abnormality and Its Application for Detection of Possible False-positive Neonatal Sepsis

Neonatal sepsis is a type of bloodstream infection that occurs within 28 days of birth, mostly in low birth weight or premature newborns. According to a 2020 report by the World Health organization, an estimated number of 1.3 million to 3.9 million newborns suffer from sepsis each year, and 4.0 million to 7.0 million die from sepsis-related complications. A newborn with an infection will show signs or changes in their appearance. Symptoms like temporary acceleration or slowing of the heart rate, and low HRV, are commonly associated with neonates who have sepsis Accurate detection of heart rate anomalies is essential for diagnosing sepsis. In this study, heart rate abnormality encompasses low HRV, transient acceleration or slowing of the heart rate, and other unusual events. The current methods of clinical observation of abnormalities are not necessarily intuitive or informative. For example, differentiation between proven sepsis and culture-negative sepsis is difficult by solely observing the average heart rate on a bedside monitor. To this end, we proposed a new visualization tool to help clinicians in the detection of suspected false-positive neonatal sepsis. Our visualization tool provides information such as the variability and distribution of the heart rates, which can be used to detect the abnormal events. These functions will help clinicians to detect possible false-positive sepsis and reduce unnecessary antibiotic treatment.