Abstract
Current artificial pancreas (AP) systems are hybrid closed-loop systems that require manual meal announcements to manage postprandial glucose control effectively. This poses a cognitive burden and challenge to users with T1D since this relies on frequent user engagement to maintain tight glucose control. In order to move towards fully automated closed-loop glucose control, we propose an algorithm based on a deep learning framework that performs multitask quantile regression, for both meal detection and carbohydrate estimation. Our proposed method is evaluated in silico on 10 adult subjects from the UVa/Padova simulator with a Bio-inspired Artificial Pancreas (BiAP) control algorithm over a 2 month period. Three different configurations of the AP are evaluated -BiAP without meal announcement (BiAP-NMA), BiAP with meal announcement (BiAP-MA), and BiAP with meal detection (BiAP-MD). We present results showing an improvement of BiAP-MD over BiAP-NMA, demonstrating 144.5 ± 6.8 mg/dL mean blood glucose level (−4.4 mg/dL, 0.01) and 77.8 ± 6.3% mean time between 70 and 180 mg/dL (+3.9%, 0.001). This improvement in control is realised without a significant increase in mean in hypoglycaemia (+0.1%, 0.4). In terms of detection of meals and snacks, the proposed method on average achieves 93% precision and 76% recall with a detection delay time of 38 ± 15 min (92% precision, 92% recall, and 37 min detection time for meals only). Furthermore, BiAP-MD handles hypoglycaemia better than BiAP-MA based on CVGA assessment with fewer control errors (10% vs. 20%). This study suggests that multitask quantile regression can improve the capability of AP systems for postprandial glucose control without increasing hypoglycaemia.
Highlights
The self-management of diabetes is a burdensome, delicate, and yet critical undertaking that individuals with diabetes have to engage in daily in order to avert adverse glycaemic events
We primarily report the following glycaemic metrics: percentage time spent in euglycaemia (70 mg/dL < BG < 180 mg/dL), percentage time spent in hyperglycaemia (BG ≥ 180 mg/dL), percentage time spent in hypoglycaemia (BG ≤ 70 mg/dL), mean glucose concentration level
We evaluate the level of control with a number of indices, the high blood glucose index (HBGI), low blood glucose index (LBGI), and risk index (RI)
Summary
The self-management of diabetes is a burdensome, delicate, and yet critical undertaking that individuals with diabetes have to engage in daily in order to avert adverse glycaemic events. Current diabetes management systems such as the artificial pancreas (AP) and decision support systems have been developed in recent years in order to improve the management of diabetes. The envisaged endpoint in the development of the insulin-based artificial pancreas is a fully automated system that does is not depend on user input throughout the day [1]. One major challenge in the realising this objective with the artificial pancreas centres on postprandial glucose control. Studies have shown that when small meals (e.g., 30 g) are missed AP systems are capable of handling the resulting postprandial increase in glucose [2,3]. The current artificial pancreas systems is classified as a hybrid closed-loop system since it requires meals to be announced prior to mealtime in order to ensure good control
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
