A Self-Learning Knowledge Based System to Mitigate the Impact of Drug Interactions in Type Two Diabetes Mellitus

Abstract

The topic of drug interactions has received great attention worldwide recently. If a drug is not administered in appropriate quantity with an appropriate combination of the drug with drug or other substances, the result will be a high risk of dangerous interactions which lead to potentially harmful side effects which are ranging from treatment failure, economic degradation, and death due to lack of drug information and maladministration of drugs among health professionals. Thus, this study was initiated with the main aim to develop a self-learning knowledge-based system to mitigate the impact of drug interactions. To develop this system, design science methodology with the integration of knowledge engineering method was used; and semi-structured interview, document analysis to acquire knowledge; questionnaire to grasp users‟ feedback, and purposive sampling technique to select domain experts were used. The acquired knowledge was then represented using a production rule and modelled using a decision tree. The system was implemented using PROLOG on SWI-PROLOG editor and evaluated using system performance and user acceptance testing. The developed system was evaluated and 96% of the users were satisfied. As well as the performance of the system was evaluated and recorded 80% accuracy, thus it can be concluded that the system achieves good performance. However, in order to make the system fully applicable in the domain area, further research work to incorporate adequate knowledge, develop online knowledge-based systems and mobile applications were recommended to enhance the accessibility of the developed system.