RADAR: Wearable-Based Dysglycemia Detection and Warning in Diabetes
Diabetes mellitus affects approximately 537 million people worldwide, and a doubling of cases is expected within the next 25 years. The economic burden manifests in yearly expenditures of at least USD 966 billion. For most people with diabetes, self-monitoring of blood glucose (SMBG) is of utmost importance for the management of glycemic control. Conventional SMBG requires piercing the skin to retrieve a drop of blood. Such finger prick tests are cumbersome, painful, and provide only limited information on glucose dynamics.
Recently, continuous glucose monitoring (CGM) has been introduced, predominantly in wealthy countries. CGM sensors are attached to the skin and measure glucose permanently. Compared to SMBG, CGM has the advantage of displaying glucose dynamics in real-time and providing alarms in case of dysglycemia (too low/high glucose). However, CGMs are invasive and expensive (CHF 4’000–5’000 per patient-year, restricted coverage by insurances). Therefore, they are often rejected by patients and a considerable financial burden for individuals and the health care system.
Existing research provides evidence that the latest advancement of smartwatch-sensors and machine learning can be leveraged to infer blood glucose or at least dysglycemia. Hence, we aim at developing a smartwatch-based system (RADAR) for blood glucose and dysglycemia detection. To ensure reliable detection, we focus on interpretable and explainable machine learning and will test the system in a large-scale field study to maximize generalization capability.
RADAR is a collaboration of the Inselspital Bern (Department of Diabetes, Endocrinology, Clinical Nutrition and Metabolism) as well as the Bosch IoT Lab and the Center for Digital Health Interventions (CDHI) at ETH Zurich and the University of St. Gallen.
Maritsch, M., Föll, S., Lehmann, V., Bérubé, C., Kraus, M., Feuerriegel, S., Kowatsch, T., Züger, T., Stettler, C., Fleisch, E., Wortmann, F., Towards Wearable-based Hypoglycemia Detection and Warning in Diabetes, Conference on Human Factors in Computing Systems – Late Breaking Work (CHI 2020), 10.1145/3334480.3382808. [PDF]