"uuid","repository link","title","author","contributor","publication year","abstract","subject topic","language","publication type","publisher","isbn","issn","patent","patent status","bibliographic note","access restriction","embargo date","faculty","department","research group","programme","project","coordinates" "uuid:ab7c5ff8-16d4-4a13-a013-61f69dafc11f","http://resolver.tudelft.nl/uuid:ab7c5ff8-16d4-4a13-a013-61f69dafc11f","Predicting sepsis on the Neonatal Intensive Care Unit","Mank, A.B.","Dankelman, J. (mentor); Goos, T.G. (mentor)","2017","To reduce the mortality rate of sepsis, treatment needs to be started as soon as possible. One should however first be sure about the occurrence of sepsis. Currently, the suspicion of sepsis in premature infants is based on non-specific physiological changes. The detection of alterations in the inter-heartbeat interval preceding sepsis can offer a solution. The aim of this study was to design a predictive model which captures the relation between physiological changes and variability in the inter heart beat interval, Heart Rate Variability (HRV) and the development of sepsis. A retrospective study was performed in the Neonatal Intensive Care Unit (NICU) at the Sophia Children’s Hospital from January 2016 to June 2016. The study focussed on very preterm infants, of Gestational Age (GA) <32 weeks, with a birth weight below 1500 gram. Sepsis was defined as a blood culture proven Late-ONset Sepsis (LONS) with elevated levels of C-reactive Protein (CRP). Logistic mixed effect modelling was used to estimate the probability of developing sepsis based on GA, gender, birth weight, CRP, percentage weight, bradycardia, median RR and HRV. During the inclusion period, 18 of the 60 infants developed at least one event of sepsis. Infants with sepsis had lower birth weight and GA, with higher mortality rates and length of stay. Logistic mixed effect models showed a significant relation between CRP, bradycardia, median RR, HRV and the probability of developing sepsis. Bradycardia, CRP, median RR and HRV are found to be useful predictors of sepsis. Since the authorisation of a blood culture to prove sepsis takes about four days and the model predicts sepsis at the moment the blood culture is taken, the model is faster in providing results.","premature infant; sepsis; heart rate variability","en","master thesis","","","","","","","","","Mechanical, Maritime and Materials Engineering","Biomedical Engineering","","Biomechatronics","",""