Pre-eclampsia Risk Prediction Using Machine Learning and Classical Regression Approaches: Towards Personalised Prenatal Care

Management of pre-eclampsia in pregnant women is challenging, but early screening and low-dose aspirin before 16 weeks can reduce risk in high-risk women by two-thirds. This thesis addresses gaps by evaluating existing risk prediction models, developing and validating prediction models using regression and machine learning, and proposing a simple model with readily available predictors. The findings support personalised prenatal care by leveraging the value of routinely collected data. My thesis presents a parsimonious early pre-eclampsia screening model, which could be optimised and integrated in the future as a baseline first-step screening to identify high-risk women for subsequent, more specific testing.