IoT-Based Diabetic Students' Decision Support System

Abstract

In decision support systems, IoT plays a crucial role, with Continuous Glucose Monitoring (CGM) devices exemplifying IoT sensors. Diabetic patients can maintain continuous control of diabetes using CGM devices and with minimal fingerstick tests. These devices alert the users in case of a decrease or increase in glucose levels. Although these devices are widely used in the healthcare sector, CGM is seldom utilised for educational purposes. However, students with diabetes could benefit from the Internet of Things. A three-phase design science research (DSR) approach was adopted for this research. An investigative study and a review of the literature were carried out during the first phase. The review revealed that there is concern regarding current conflicts in the literature. Previous research suggested that diabetic students do not perform worse academically and that the condition has no effect on cognitive abilities. However, more recent research revealed that in addition to attendance issues, cognitive functioning problems and difficulty focusing cause these students to perform less effectively in their academic tasks. The main aim of this phase was to establish the issues and develop a system to solve them. Then, an investigative study design based on mixed methods was conducted with 231 teachers and administrative staff, 133 parents of diabetic students, and ‎154 diabetic students in Saudi Arabia high schools. In addition, semi-structured interviews were conducted with 3 diabetic specialists at the University of Warwick and 5 parents of diabetic students in Saudi Arabia. The investigative study concluded that more than half of all participants agreed that poor glycaemic control affects academic achievement of students in addition to the fact that they suffer with focus, attendance, and attention as implications of repeated hypoglycaemia and hyperglycaemia. Therefore, the development phase proposed an Internet of Things-based support system to solve the extracted issues. In addition, this phase included the development and implementation of the deterministic classification model which classifies glucose readings. Then, the development of mobile and web applications and the integration of these applications with the model was built. Finally, the evaluation phase consisted of a main experiment implemented in a Saudi Arabian high school with 5 teachers and 4 students utilising Dexcom CGM sensors. Semi-structured interviews with teachers were conducted to assess their perspectives on utilising the mobile application during the assessment period. Furthermore, the conclusions drawn demonstrated that using the developed mobile and web applications enhanced teachers' experiences. Recommendations for further research as well as the study's limitations are provided in the conclusion of the study.