A context sensitive, advisory decision support approach for mobile, knowledge based, time critical environments

2017-03-22T01:46:58Z (GMT) by Grigsby, Shane Anthony
This research aimed to explore decision support in mobile, knowledge based, time critical environments involving ill-defined decision situations. The research was concerned with how decision makers in such environments made their decisions and whether/how the decision making process could be supported by knowledge based decision support technology. The aim presented within the thesis is to ascertain the necessary components of a decision support system capable of incorporating contextual knowledge, real time context modelling and providing useful decision support for ill-defined decision situations. A design science research methodology was adopted for the research. This methodology incorporated researching the domain, knowledge acquisition, artefact development in the form of a prototype, and a qualitative evaluation of the system's feasibility and potential usefulness in a particular context. The research utilised decision support, mobile technology and knowledge acquisition concepts, and of particular importance, the areas of context sensitive decision support and advisory systems to develop the prototype. This research developed a new approach towards decision support within the target problem domain. This approach involved utilising context sensitive decision support techniques to represent expert knowledge in a generalised form. This knowledge was then incorporated within a mobile advisory system prototype to address the time critical nature of the target domain. Hospital triage was selected as an example of the time critical problem domain. Expert knowledge was obtained from triage nurses, which was not otherwise detailed in official triage documentation. The prototype system was developed to explore the implementation of the approach and acquired knowledge. Triage nurses were also involved in a qualitative evaluation of the prototype mobile decision support system. The evaluation was aimed at establishing the importance of the contextual components of the system and its suitability within the triage environment. Feedback from the evaluation suggested domain experts' approval of this form of decision support. Comments issued indicate decision support in this form could be of benefit for triage nurses to support their decisions with specific experiential knowledge. The domain experts also indicated that this form of support would be most beneficial for trainee triage nurses whose lack of experience necessitates additional guidance. Another potential application was suggested for the purpose of education of staff in specific knowledge of medical triage. This research contributes to the theory and practice of development of knowledge based mobile decision support systems within such domains where domain knowledge is critical but not well documented. The implementation of generalised guiding contextual rules was found to be useful within a mobile and time critical environment involving ill-defined decision situations. The generalised contextual rules provided safe decision support and in a form that would have minimal impact on the time taken to make decisions. The key concepts within the system which addressed the research questions are generalised contextual rules, context management objects, and an interface that presents guiding information to address time constraints in mobile decision support environment.