Clinical intelligence framework for decision-support

Nammour, Fadi Louis (2018) Clinical intelligence framework for decision-support. (PhD thesis), Kingston University, .


The healthcare setting has been evolving in many ways to keep up with the changing technology and information challenges. The modern healthcare setting hosts patient and clinical data over many co-existing information systems and medical devices. It is rarely the case that these data sources correctly exchange clinical data in a standard way, and for that reason answering questions that come up during the management of the healthcare processes requires looking up the pieces of information in this distributed and loosely connected ecosystem. As the healthcare organization grows in both size and operational activities, decision-making stakeholders become more dependent on status summaries, which become harder to manually prepare with large volumes of scattered data. As the decision-makers follow a performance-driven evaluation approach, their enquiries span multiple data sources and require adaptation of data models in many information systems and data silos. Decision-support officers and clinicians find themselves in need to work with technical personnel to help them connect to the many data systems and combine their data in a way the delivers the correct expected results to the stakeholders. As the demands are not the same every time, this process depends on continuous and expensive technical involvement. The aim of this thesis is to enable non-technical clinicians and decision-support officers to work with data from different information systems and medical devices using a framework software system that consolidates this data and allows them to create their own analytical enquiries without the continuous assistance of technical personnel. To achieve this aim, the thesis critically examined current approaches and existing solutions, then identified the need for a new approach to eliminate the limitations identified, analysed these approaches and proposed a new approach. The proposed solution was designed and developed following well-established and tested methodologies and was then evaluated and its impacts were identified in the healthcare environment. The research in this thesis questioned the possibility of having a consistently usable framework system where non-technical users can execute their technical analytics against clinical data from different data sources in the healthcare setting. A case study was performed in multiple hospitals to collect the needs of different users and to identify the gap in the current situation. The results showed that the adoption process was possible and was tested through an application scenario where the users contributed collectively to the creation of a monthly status report portal. The portal was used by the stakeholders to follow-up on the healthcare setting's current and historical performances and they added new requirements in the form of aggregation requests that were executed through the analytical framework system. The framework system was developed using a Model-Viewer-Controller software design approach and followed the best-practices of modern software engineering. The service-oriented architecture was adopted to govern the delivery of data from an application server to thin clients such as a web-browser or a mobile device. An algorithm was developed to enable the dynamic execution of user analytical enquiries against a dynamic data model. In addition to having a dynamic aggregation algorithm at the core of the framework system, this thesis provided a flexible user-experience where the users were able to contribute throughout the phases of the data loading and transformation, analytic execution against the loaded data, and the possibility of expressing the outcome in different visualization representations. A questionnaire was conducted to further comprehend the significance of the framework system as part of the healthcare setting, and to get feedback from the users after using it for some months. The results of the questionnaire were analysed for statistical correlations, and the outcome was an evaluation of users' experience that targeting different aspects of their engagement with the framework. The outcome was summarized and further work was suggested in order to improve it in future versions.

Actions (Repository Editors)

Item Control Page Item Control Page