Ongoing Debate About Models for Artificial Pancreas Systems and In Silico Studies

Technological advancements in glucose monitoring and artificial pancreas systems for shaping diabetes care

The management of diabetes mellitus has undergone remarkable progress with the introduction of cutting-edge technologies in glucose monitoring and artificial pancreas systems. These innovations have revolutionized diabetes care, offering patients more precise, convenient, and personalized management solutions that significantly improve their quality of life. This review aims to provide a comprehensive overview of recent technological advancements in glucose monitoring devices and artificial pancreas systems, focusing on their transformative impact on diabetes care. A detailed review of the literature was conducted to examine the evolution of glucose monitoring technologies, from traditional invasive methods to more advanced systems. The review explores minimally invasive techniques such as continuous glucose monitoring (CGM) systems and flash glucose monitoring (FGM) systems, which have already been proven to enhance glycemic control and reduce the risk of hypoglycemia. In addition, emerging non-invasive glucose monitoring technologies, including optical, electrochemical, and electro-mechanical methods, were evaluated. These techniques are paving the way for more patient-friendly options that eliminate the need for frequent finger-prick tests, thereby improving adherence and ease of use. Advancements in closed-loop artificial pancreas systems, which integrate CGM with automated insulin delivery, were also examined. These systems, often referred to as "hybrid closed-loop" or "automated insulin delivery" systems, represent a significant leap forward in diabetes care by automating the process of insulin dosing. Such advancements aim to mimic the natural function of the pancreas, allowing for better glucose regulation without the constant need for manual interventions by the patient. Technological breakthroughs in glucose monitoring and artificial pancreas systems have had a profound impact on diabetes management, providing patients with more accurate, reliable, and individualized treatment options. These innovations hold the potential to significantly improve glycemic control, reduce the incidence of diabetes-related complications, and ultimately enhance the quality of life for individuals living with diabetes. Researchers are continually exploring novel methods to measure glucose more effectively and with greater convenience, further refining the future of diabetes care. Researchers are also investigating the integration of artificial intelligence and machine learning algorithms to further enhance the precision and predictive capabilities of glucose monitoring and insulin delivery systems. With ongoing advancements in sensor technology, connectivity, and data analytics, the future of diabetes care promises to deliver even more seamless, real-time management, empowering patients with greater autonomy and improved health outcomes.

Offline reinforcement learning for safer blood glucose control in people with type 1 diabetes

The widespread adoption of effective hybrid closed loop systems would represent an important milestone of care for people living with type 1 diabetes (T1D). These devices typically utilise simple control algorithms to select the optimal insulin dose for maintaining blood glucose levels within a healthy range. Online reinforcement learning (RL) has been utilised as a method for further enhancing glucose control in these devices. Previous approaches have been shown to reduce patient risk and improve time spent in the target range when compared to classical control algorithms, but are prone to instability in the learning process, often resulting in the selection of unsafe actions. This work presents an evaluation of offline RL for developing effective dosing policies without the need for potentially dangerous patient interaction during training. This paper examines the utility of BCQ, CQL and TD3-BC in managing the blood glucose of the 30 virtual patients available within the FDA-approved UVA/Padova glucose dynamics simulator. When trained on less than a tenth of the total training samples required by online RL to achieve stable performance, this work shows that offline RL can significantly increase time in the healthy blood glucose range from 61.6±0.3% to 65.3±0.5% when compared to the strongest state-of-art baseline (p<0.001). This is achieved without any associated increase in low blood glucose events. Offline RL is also shown to be able to correct for common and challenging control scenarios such as incorrect bolus dosing, irregular meal timings and compression errors. The code for this work is available at: https://github.com/hemerson1/offline-glucose.