Show simple item record

dc.descriptionPhD thesisen_US
dc.description.abstractBayesian networks have been widely proposed to assist clinical decision making. Their popularity is due to their ability to combine different sources of information and reason under uncertainty, using sound probabilistic laws. Despite their benefit, there is still a gap between developing a Bayesian network that has a good predictive accuracy and having a model that makes a significant difference to clinical decision making. This thesis tries to bridge that gap and proposes three novel contributions. The first contribution is a modelling approach that captures the progress of an acute condition and the dynamic way that clinicians gather information and take decisions in irregular stages of care. The proposed method shows how to design a model to generate predictions with the potential to support decision making in successive stages of care. The second contribution is to show how counterfactual reasoning with a Bayesian network can be used as a healthcare governance tool to estimate the effect of treatment decisions other than those occurred. In addition, we extend counterfactual reasoning in situations where the targeted decision and its effect belong to different stages of the patient’s care. The third contribution is an explanation of the Bayesian network’s reasoning. No model is going to be used if it is unclear how it reasons. Presenting an explanation, alongside a prediction, has the potential to increase the acceptability of the network. The proposed technique indicates which important evidence supports or contradicts the prediction and through which intermediate variables the information flows. The above contributions are explored using two clinical case studies. A clinical case study on combat trauma care is used to investigate the first two contributions. The third contribution is explored using a Bayesian network developed by others to provide decision support in treating acute traumatic coagulopathy in the emergency department. Both case studies are done in collaboration with the Royal London Hospital and the Royal Centre for Defence Medicine.en_US
dc.publisherQueen Mary University of London
dc.subjectgraphene nanoplateletsen_US
dc.titleBayesian Networks for Clinical Decision Making : Support, Assurance, Trusten_US
dc.rights.holderThe copyright of this thesis rests with the author and no quotation from it or information derived from it may be published without the prior written consent of the author

Files in this item


This item appears in the following Collection(s)

  • Theses [3584]
    Theses Awarded by Queen Mary University of London

Show simple item record