Modelling glucose dynamics during moderate exercise in individuals with type 1 diabetes.

The artificial pancreas is a closed-loop insulin delivery system that automatically regulates glucose levels in individuals with type 1 diabetes. In-silico testing using simulation environments accelerates the development of better artificial pancreas systems. Simulation environments need an accurate model that captures glucose dynamics during exercise to simulate real-life scenarios. We proposed six variations of the Bergman Minimal Model to capture the physiological effects of moderate exercise on glucose dynamics in individuals with type 1 diabetes. We estimated the parameters of each model with clinical data using a Bayesian approach and Markov chain Monte Carlo methods. The data consisted of measurements of plasma glucose, plasma insulin, and oxygen consumption collected from a study of 17 adults with type 1 diabetes undergoing aerobic exercise sessions. We compared the models based on the physiological plausibility of their parameters estimates and the deviance information criterion. The best model features (i) an increase in glucose effectiveness proportional to exercise intensity, and (ii) an increase in insulin action proportional to exercise intensity and duration. We validated the selected model by reproducing results from two previous clinical studies. The selected model accurately simulates the physiological effects of moderate exercise on glucose dynamics in individuals with type 1 diabetes. This work offers an important tool to develop strategies for exercise management with the artificial pancreas.

Effect of frequency of sensor use on glycaemic control in individuals on sensor-augmented pump therapy with and without Predictive Low Glucose Management System.

Improved frequency of sensor use improves glycaemic control. Furthermore, there is no deterioration of glycaemic control with increased sensor use in individuals on Predictive Low Glucose Management (PLGM) system. Younger children are more likely to have better sensor uptake than older children.

Assessing patient-reported outcomes for automated insulin delivery systems: the psychometric properties of the INSPIRE measures.

AIM: Participants in clinical trials assessing automated insulin delivery systems report perceived benefits and burdens that reflect their experiences and may predict their likelihood of uptake and continued use of this novel technology. Despite the importance of understanding their perspectives, there are no available validated and reliable measures assessing the psychosocial aspects of automated insulin delivery systems. The present study assesses the initial psychometric properties of the INSPIRE measures, which were developed for youth and adults with Type 1 diabetes, as well as parents and partners. METHODS: Data from 292 youth, 159 adults, 150 parents of youth and 149 partners of individuals recruited from the Type 1 Diabetes Exchange Registry were analysed. Participants completed INSPIRE questionnaires and measures of quality of life, fear of hypoglycaemia, diabetes distress, glucose monitoring satisfaction. Exploratory factor analysis assessed factor structures. Associations between INSPIRE scores and other measures, HbA1c , and technology use assessed concurrent and discriminant validity. RESULTS: Youth, adult, parent and partner measures assess positive expectancies of automated insulin delivery systems. Measures range from 17 to 22 items and are reliable (α = 0.95-0.97). Youth, adult and parent measures are unidimensional; the partner measure has a two-factor structure (perceptions of impact on partners versus the person with diabetes). Measures showed concurrent and discriminant validity. CONCLUSIONS: INSPIRE measures assessing the positive expectancies of automated insulin delivery systems for youth, adults, parents and partners have meaningful factor structures and are internally consistent. The developmentally sensitive INSPIRE measures offer added value as clinical trials test newer systems, systems become commercially available and clinicians initiate using these systems.

Is an artificial pancreas (closed-loop system) for Type 1 diabetes effective?

The artificial pancreas is now a viable treatment option for people with Type 1 diabetes and has demonstrated improved glycaemic outcomes while also reducing the onus of self-management of Type 1 diabetes. Closed-loop glucose-responsive insulin delivery guided by real-time sensor glucose readings can accommodate highly variable day-to-day insulin requirements and reduce the hypoglycaemia risk observed with tight glycaemic control in Type 1 diabetes. In 2011, the James Lind Alliance research priorities for Type 1 diabetes were produced and priority 3 was to establish whether an artificial pancreas (closed-loop system) for Type 1 diabetes is effective. This review focuses on the progress that has been made in the evolution of closed-loop systems as an effective treatment option for Type 1 diabetes. Development of closed-loop systems has advanced from feasibility evaluations in highly supervised settings over short periods, to clinical studies in free-living, unsupervised conditions lasting several months. The approval in the USA of the first hybrid closed-loop system (MiniMed® 670G pump, Medtronic, Northridge, CA, USA) in 2016 for use in Type 1 diabetes reflects these advancements. We discuss the evidence from clinical studies that closed-loop systems are effective with improved glycaemic outcomes, reduced hypoglycaemia and had positive end-user acceptance in children, adolescents, adults and pregnant women with Type 1 diabetes. We also present the outlook for future closed-loop systems in the treatment of Type 1 diabetes and identify the challenges facing the wide-spread clinical adoption of this technology.

Model individualization for artificial pancreas.

BACKGROUND AND OBJECTIVE: The inter-subject variability characterizing the patients affected by type 1 diabetes mellitus makes automatic blood glucose control very challenging. Different patients have different insulin responses, and a control law based on a non-individualized model could be ineffective. The definition of an individualized control law in the context of artificial pancreas is currently an open research topic. In this work we consider two novel identification approaches that can be used for individualizing linear glucose-insulin models to a specific patient. METHODS: The first approach belongs to the class of black-box identification and is based on a novel kernel-based nonparametric approach, whereas the second is a gray-box identification technique which relies on a constrained optimization and requires to postulate a model structure as prior knowledge. The latter is derived from the linearization of the average nonlinear adult virtual patient of the UVA/Padova simulator. Model identification and validation are based on in silico data collected during simulations of clinical protocols designed to produce a sufficient signal excitation without compromising patient safety. The identified models are evaluated in terms of prediction performance by means of the coefficient of determination, fit, positive and negative max errors, and root mean square error. RESULTS: Both identification approaches were used to identify a linear individualized glucose-insulin model for each adult virtual patient of the UVA/Padova simulator. The resulting model simulation performance is significantly improved with respect to the performance achieved by a linear average model. CONCLUSIONS: The approaches proposed in this work have shown a good potential to identify glucose-insulin models for designing individualized control laws for artificial pancreas.

Optimizing Hybrid Closed-Loop Therapy in Adolescents and Emerging Adults Using the MiniMed 670G System.

OBJECTIVE: The MiniMed 670G System is the first commercial hybrid closed-loop (HCL) system for management of type 1 diabetes. Using data from adolescent and young adult participants, we compared insulin delivery patterns and time-in-range metrics in HCL (Auto Mode) and open loop (OL). System alerts, usage profiles, and operational parameters were examined to provide suggestions for optimal clinical use of the system. RESEARCH DESIGN AND METHODS: Data from 31 adolescent and young adult participants (14-26 years old) at three clinical sites in the 670G pivotal trial were analyzed. Participants had a 2-week run-in period in OL, followed by a 3-month in-home study phase with HCL functionality enabled. Data were compared between baseline OL and HCL use after 1 week, 1 month, 2 months, and 3 months. RESULTS: Carbohydrate-to-insulin (C-to-I) ratios were more aggressive for all meals with HCL compared with baseline OL. Total daily insulin dose and basal-to-bolus ratio did not change during the trial. Time in range increased 14% with use of Auto Mode after 3 months (P < 0.001), and HbA1c decreased 0.75%. Auto Mode exits were primarily due to sensor/insulin delivery alerts and hyperglycemia. The percentage of time in Auto Mode gradually declined from 87%, with a final use rate of 72% (-15%). CONCLUSIONS: In transitioning young patients to the 670G system, providers should anticipate immediate C-to-I ratio adjustments while also assessing active insulin time. Users should anticipate occasional Auto Mode exits, which can be reduced by following system instructions and reliably bolusing for meals. Unique 670G system functionality requires ongoing clinical guidance and education from providers.

Evaluation of a novel continuous glucose monitoring guided system for adjustment of insulin dosing - PumpTune: a randomized controlled trial.

OBJECTIVE: Retrospective continuous glucose monitoring (CGM) can guide insulin pump adjustments, however, interpretation of data and recommending new pump settings is complex and subjective. We aimed to compare the safety and glycaemic profiles of children after their diabetologist or a novel algorithm (PumpTune) adjusted their insulin pump settings. RESEARCH DESIGN AND METHODS: In a randomized cross-over trial of 22 patients aged 6-14 yr with type 1 diabetes with mean Hba1c 7.4% (57 mmol/mol) using CSII, CGM was used over two periods each of 6.5 d to assess percentage time glucose remained within, above and below 3.9-10.0 mmol/L. Before the start of one period pump settings were adjusted by the patient's diabetologist, and before the other insulin pump settings were adjusted by PumpTune. RESULTS: A total of 63.4% of the sensor glucose levels were within target range with PumpTune settings and 57.4% were within range with the clinician settings (p = 0.016). The time spent above target range with PumpTune was 26.9% and with clinician settings was 33.5% (p = 0.021). The time spent below target range with PumpTune was 9.7% and with clinician settings was 9.2% (p = 0.77). The mean number of times when a sensor glucose level <2.75 mmol/L was recorded with PumpTune settings was 2.9 compared with 3.7 with clinician settings (p = 0.39). There were no serious adverse outcomes and no difference in parent-assessed satisfaction. CONCLUSIONS: Automated insulin pump adjustment with PumpTune is feasible and warrants testing in a larger more varied population over a longer time. In this well-controlled group of children, PumpTune achieved a more favorable glucose profile.

Diabetes Device Interoperability for Improved Diabetes Management.

Scientific and technological advancements have led to the increasing availability and use of sophisticated devices for diabetes management, with corresponding improvements in public health. These devices are often capable of sharing data with a few other specific devices but are generally not broadly interoperable; they cannot work together with a wide variety of other devices. As a result of limited interoperability, benefits of modern diabetes devices and potential for development of innovative new diabetes technologies are not being fully realized. Here we discuss diabetes device interoperability in general, then focus on 4 examples that show how diabetes management could benefit from enhanced interoperability: remote monitoring and data sharing, integrating data from multiple devices to better inform diabetes management strategies, device consolidation, and artificial pancreas development.

Cybersecurity in Artificial Pancreas Experiments.

Medical devices have transformed modern health care, and ongoing experimental medical technology trials (such as the artificial pancreas) have the potential to significantly improve the treatment of several chronic conditions, including diabetes mellitus. However, we suggest that, to date, the essential concept of cybersecurity has not been adequately addressed in this field. This article discusses several key issues of cybersecurity in medical devices and proposes some solutions. In addition, it outlines the current requirements and efforts of regulatory agencies to increase awareness of this topic and to improve cybersecurity.

Advancement of the artificial pancreas through the development of interoperability standards.

Despite advancements in the development of the artificial pancreas, barriers in the form of proprietary data and communication protocols of diabetes devices have made the integration of these components challenging. The Artificial Pancreas Standards and Technical Platform Project is an initiative funded by the JDRF Canadian Clinical Trial Network with the goal of developing device communication standards for the interoperability of diabetes devices. Stakeholders from academia, industry, regulatory agencies, and medical and patient communities have been engaged in advancing this effort. In this article, we describe this initiative along with the process involved in working with the standards organizations and stakeholders that are key to ensuring effective standards are developed and adopted. Discussion from a special session of the 12th Annual Diabetes Technology Meeting is also provided.

Fitting dynamic models with forcing functions: application to continuous glucose monitoring in insulin therapy.

The artificial pancreas is an emerging technology to treat type 1 diabetes (T1D). It has the potential to revolutionize diabetes care and improve quality of life. The system requires extensive testing, however, to ensure that it is both effective and safe. Clinical studies are resource demanding and so a principle aim is to develop an in silico population of subjects with T1D on which to conduct pre-clinical testing. This paper aims to reliably characterize the relationship between blood glucose and glucose measured by subcutaneous sensor as a major step towards this goal. Blood-and sensor-glucose are related through a dynamic model, specified in terms of differential equations. Such models can present special challenges for statistical inference, however. In this paper we make use of the BUGS software, which can accommodate a limited class of dynamic models, and it is in this context that we discuss such challenges. For example, we show how dynamic models involving forcing functions can be accommodated. To account for fluctuations away from the dynamic model that are apparent in the observed data, we assume an autoregressive structure for the residual error model. This leads to some identifiability issues but gives very good predictions of virtual data. Our approach is pragmatic and we propose a method to mitigate the consequences of such identifiability issues.

Generic safety requirements for developing safe insulin pump software.

BACKGROUND: The authors previously introduced a highly abstract generic insulin infusion pump (GIIP) model that identified common features and hazards shared by most insulin pumps on the market. The aim of this article is to extend our previous work on the GIIP model by articulating safety requirements that address the identified GIIP hazards. These safety requirements can be validated by manufacturers, and may ultimately serve as a safety reference for insulin pump software. Together, these two publications can serve as a basis for discussing insulin pump safety in the diabetes community. METHODS: In our previous work, we established a generic insulin pump architecture that abstracts functions common to many insulin pumps currently on the market and near-future pump designs. We then carried out a preliminary hazard analysis based on this architecture that included consultations with many domain experts. Further consultation with domain experts resulted in the safety requirements used in the modeling work presented in this article. RESULTS: Generic safety requirements for the GIIP model are presented, as appropriate, in parameterized format to accommodate clinical practices or specific insulin pump criteria important to safe device performance. CONCLUSIONS: We believe that there is considerable value in having the diabetes, academic, and manufacturing communities consider and discuss these generic safety requirements. We hope that the communities will extend and revise them, make them more representative and comprehensive, experiment with them, and use them as a means for assessing the safety of insulin pump software designs. One potential use of these requirements is to integrate them into model-based engineering (MBE) software development methods. We believe, based on our experiences, that implementing safety requirements using MBE methods holds promise in reducing design/implementation flaws in insulin pump development and evolutionary processes, therefore improving overall safety of insulin pump software.

Analysis of novel methods to determine the accuracy of the OmniPod insulin pump: a key component of the artificial pancreas system.

In this issue of Journal of Diabetes Science and Technology, Zisser and collegues describe two inexpensive methods for accurate measurement of dosage delivered by OmniPod insulin pump. The first method is based on imaging a meniscus movement in a micro-pipette and using simple image analysis; the second relies on using a digital microscope to measure the volume of a dispensed droplet while it is still attached to the cannula tip. Both methods produce accurate measurements for doses >0.2 U, and the latter method is especially appropriate for doses <0.2 U, with accuracies down to 0.9%.

Novel methodology to determine the accuracy of the OmniPod insulin pump: a key component of the artificial pancreas system.

BACKGROUND: This article describes two novel and easy approaches for assessing the accuracy of insulin pumps as implemented within the artificial pancreas system. The approaches are illustrated by data testing the OmniPod Insulin Management System at its lowest delivery volume (0.05 U) and at doses of 0.1, 0.2, 1, and 6U. METHOD: In method 1, a pipette, digital microscope, and imaging software were used to measure average bolus delivery on a linear scale for multiple volumes. In method 2, a digital microscope and imaging software were used to measure the volume of a spherical bolus of 0.05 U of insulin. RESULTS: Bench testing results using the two novel methods demonstrated that the OmniPod is extremely accurate, with a relative error ranging from -0.90% to +0.96% for all measured doses (0.05, 0.1, 0.2, 1, and 6 U). In method 1, at target bolus dose of 0.05 U, the mean delivered dose (± standard deviation) was 0.0497 ± 0.003 U, 0.099 ± 0.005 U at 0.1 U, 0.2 ± <1e-5 U at 0.2 U, 1.001 ± 0.018 U at 1 U, and 6.03 ± 0.04 U at 6 U. In method 2, at target bolus dose of 0.5 ml, the mean delivered dose for both OmniPods was 0.505 ± 0.014. CONCLUSIONS: Both methods confirmed a high degree of accuracy for the OmniPod insulin pump. These techniques can be used to estimate delivery volume in other infusion pumps as well.