Auto adaptation of closed-loop insulin delivery system using continuous reward functions and incremental discretization.

Several closed or hybrid loop controllers for Blood Glucose (BG) regulation, which are also known as Artificial Pancreas (AP) Systems or Automated Insulin Delivery systems (AIDs), are in development worldwide. Most AIDs are designed and evaluated for short-term performance, with a particular emphasis on the post-meal period. However, if controllers are not adapted properly to account for variations in physiology that affect Insulin Sensitivity (IS), the AIDs may perform inadequately. In this work, the performance of two Reinforcement Learning (RL) agents trained under both piecewise and continuous reward functions is evaluated in-silico for long-term adaptation of a Fully Automated Insulin Delivery (fAID) system. An automatic adaptive discretization scheme that expands the state space as needed is also implemented to avoid disproportionate state space exploration. The proposed agents are evaluated for long-term adaptation of the Automatic Regulation of Glucose (ARG) algorithm, considering variations in IS. Results show that both RL agents have improved performance compared to a rule-based decision-making approach and the baseline controller for the majority of the adult population. Moreover, the use of a continuous shaped reward function proves to enhance the performance of the agents further than a piecewise one.

First Outpatient Clinical Trial of a Full Closed-Loop Artificial Pancreas System in South America.

BACKGROUND: The first two studies of an artificial pancreas (AP) system carried out in Latin America took place in 2016 (phase 1) and 2017 (phase 2). They evaluated a hybrid algorithm from the University of Virginia (UVA) and the automatic regulation of glucose (ARG) algorithm in an inpatient setting using an AP platform developed by the UVA. The ARG algorithm does not require carbohydrate (CHO) counting and does not deliver meal priming insulin boluses. Here, the first outpatient trial of the ARG algorithm using an own AP platform and doubling the duration of previous phases is presented. METHOD: Phase 3 involved the evaluation of the ARG algorithm in five adult participants (n = 5) during 72 hours of closed-loop (CL) and 72 hours of open-loop (OL) control in an outpatient setting. This trial was performed with an own AP and remote monitoring platform developed from open-source resources, called InsuMate. The meals tested ranged its CHO content from 38 to 120 g and included challenging meals like pasta. Also, the participants performed mild exercise (3-5 km walks) daily. The clinical trial is registered in ClinicalTrials.gov with identifier: NCT04793165. RESULTS: The ARG algorithm showed an improvement in the time in hyperglycemia (52.2% [16.3%] OL vs 48.0% [15.4%] CL), time in range (46.9% [15.6%] OL vs 50.9% [14.4%] CL), and mean glucose (188.9 [25.5] mg/dl OL vs 186.2 [24.7] mg/dl CL) compared with the OL therapy. No severe hyperglycemia or hypoglycemia episodes occurred during the trial. The InsuMate platform achieved an average of more than 95% of the time in CL. CONCLUSION: The results obtained demonstrated the feasibility of outpatient full CL regulation of glucose levels involving the ARG algorithm and the InsuMate platform.

Automatic glycemic regulation for the pediatric population based on switched control and time-varying IOB constraints: an in silico study.

Artificial pancreas (AP) systems have shown to improve glucose regulation in type 1 diabetes (T1D) patients. However, full closed-loop performance remains a challenge particularly in children and adolescents, since these age groups often present the worst glycemic control. In this work, an algorithm based on switched control and time-varying IOB constraints is presented. The proposed control strategy is evaluated in silico using the FDA-approved UVA/ Padova simulator and its performance contrasted with the previously introduced Automatic Regulation of Glucose (ARG) algorithm in the pediatric population. The effect of unannounced meals is also explored. Results indicate that the proposed strategy achieves lower hypo- and hyperglycemia than the ARG for both announced and unannounced meals. Graphical Abstract Block diagram and illustrative example of insulin and glucose evolution over time for the proposed algorithm (ARGAE).