Samenvatting
The aim of this thesis was to develop and evaluate the performance of machine learning algorithms applied across a breadth of clinical challenges in patients undergoing cardiac surgery and in critically ill patients. Different studies focused on different technical and clinical challenges: from the complexity of modelling increasingly richer iterations of a single-centre perioperative dataset, to the task of producing and modelling synthetic data. One important technical aspect studied in this thesis was how to optimally employ ML algorithms to predict mortality in patients undergoing different types of cardiac surgery. Through a journey of learning from the early days of ML applications in healthcare, particularly in cardiac surgery and critical care, this thesis can be seen as a roadmap of the first steps required for the development of predictive algorithms for prognostication in medicine. Any potential implementation of the algorithms developed would require thorough external validation, as well as extensive research on real-world implementation, effectiveness, and adoption.
Originele taal-2 | English |
---|---|
Kwalificatie | Doctor of Philosophy |
Toekennende instantie |
|
Begeleider(s)/adviseur |
|
Datum van toekenning | 14-feb.-2024 |
Plaats van publicatie | [Groningen] |
Uitgever | |
Gedrukte ISBN's | 978-94-93330-52-8 |
Elektronische ISBN's | 978-94-93330-53-5 |
DOI's | |
Status | Published - 2024 |