The Research Of Predicting Neonatal Sepsis In Preterm Infants By Using The Clinical Monitoring Data In NICU

Premature birth and its associated clinical problems are importantcauses of neonatal deaths and long-term adverse consequences, and thushave become more and more concerned in the field of perinatology andneonatology. About25%of the very low birth weight infants occurredneonatal sepsis. This disease has a mortality that approaches20%, and willalso cause permanent damage to the infants’ never system. Therefore, theresearch of early diagnosis of neonatal sepsis has great clinicalsignificance.Predictive monitoring system is a kind of clinical monitoring systemusing in the intensive care unit, which can intelligently and automaticallypredict the disease may occur in the patient by using clinical monitoringdata. Predictive monitoring system is based on mathematical analysistechnologies and computed signal processing technologies, and enablesclinicians to take effective measures in advance to prevent the risk ofoccurring serious diseases or even death. Predictive monitoring system is anon-invasive, continuous, and real-time system, and so it is important inthe prediction of neonatal sepsis.This thesis is dedicated to the research of predictive monitoringsystem for neonatal sepsis, and is mainly focused on the predictive effectof clinical ECG data in predicting neonatal sepsis. A data acquisition platform is established in the NICU of International Peace Maternity andChild Health Hospital for24-hours clinical ECG acquisition, and then anECG processing system is designed for ECG characteristic recognition,which is based on the slope threshold method and adaptive thresholdmethod. At last, this thesis analysis the relationship between ECG dataacquired in NICU and their clinical diagnosis, which can provide atheoretical basis for the establishment of predictive monitoring system forneonatal sepsis in the future.