Using Machine Learning to Model Early-onset Neonatal Sepsis Risk in Late Preterm and Term Neonates in Uganda and Zimbabwe
The goal of this observational study is to develop a risk prediction model for early-onset neonatal sepsis in term and late preterm neonates in Uganda and Zimbabwe. The main questions it aims to answer are: * What are the risk factors for early-onset neonatal sepsis in low-resource settings? * How can these be combined into a risk prediction model? Mother-baby pairs will be recruited in Uganda. They will have extensive data taken on their medical and obstetric histories and lifestyles, and their newborns will have a blood sample taken just after birth for culture. Machine learning techniques will be used to create the risk prediction model, which will then be validated in a second population in Zimbabwe.
• Neonates born at one of the study sites
• Neonates born at ≥34 weeks gestational age, defined by:
• Ultrasound scan, where participant has had an ultrasound scan during routine antenatal care
• If no ultrasound scan has been done, according to last menstrual period date, according to mother
• If no ultrasound scan has been done and the mother does not know her last menstrual period date, the neonate will be eligible for inclusion if their birth weight is ≥1400g which will include the majority of births ≥34 weeks according to the results of the INTERGROWTH-21st Project(44)