Degree Name

Master of Computer Science


School of Computer Science and Software Engineering


A major challenge for healthcare systems is to manage the continuously increasing amount of disparate data. Such large quantities of unstructured data, spread across multiple data sources, makes it difficult and time consuming for clinicians to locate the information they need to perform their tasks efficiently and accurately. This makes it difficult for knowledge workers to identify the relevant information that they need to perform their daily tasks. The ability to retrieve this information is constrained by not only the sheer volume of data but also the logical disconnection between the physical storage of the data and the processes that created the data as, in most situations, clinical processes and process related data are managed separately. It is of prime importance not only in clinical decision making but also for the safety of the patients for clinicians to be able to retrieve the appropriate information, in the appropriate level of granularity, in a timely manner. Clinical artefact networks (CANets) are introduced with the aim of addressing this issue by representing the existing data in a contextual format. With CA-Nets, we aim to correlate the semantic aspect of the data irrespective of how, where and in what format the data are stored. The resulting model will then provide an ability to navigate through the collection of data items as one would navigate through a graph or network. Such a model will provide the appropriate information at the required level of granularity. CA-Nets are developed based on semantic networks and data provenance techniques. Thus, this thesis addresses an important problem in supporting clinical decisions-that of generating and representing contextual knowledge.