Supervised and Unsupervised Approaches for the Real-Time Detection of Undesired Insulin Suspension Caused by Malfunctions

BACKGROUND

Automated insulin delivery (AID) systems, permit improved treatment of type 1 diabetes (T1D). Unfortunately, malfunctioning in the insulin pump or in the infusion set can prevent insulin from being administered, reducing the AID efficacy and posing the patient at risk. Different data-driven methods available in the literature can be used to deal with the problem of automatically detecting complete insulin suspension in real-time. This article investigates both supervised and unsupervised strategies and proposes a fair comparison under either population or personalized settings.

METHODS

Several algorithms are compared using data generated through the UVA/Padova T1D simulator, a computer simulator widely used to test control strategies in silico and accepted by the Food and Drugs Administration (FDA) as a substitute to animal pre-clinical trials. Two synthetic data sets, each consisting of 100 virtual subjects monitored for 1 month, were generated. Occasional faults of the insulin pump are simulated as complete occlusions by suspending the therapy administration. Personalized algorithms are investigated with unsupervised approaches only, since personalized labels are hardly available.

RESULTS

In the population scenario, the supervised approach outperforms the unsupervised strategy. In particular, logistic regression and random forest achieves a recall of 72% and 82%, with 0.12 and 0.21 false positives (FP) per day, respectively. In the personalized setting scenario, the unsupervised algorithms are tailored on each patient and outperform the population ones, in particular isolation forest achieves a recall 80% and 0.06 FPs per day.

CONCLUSIONS

This article suggests that unsupervised personalized approach, by addressing the large variability in glucose response among individuals with T1D, is superior to other one-fits-all approaches in detecting insulin suspensions caused by malfunctioning. Population methodologies can be effectively used while waiting to collect sufficient patient data, when the system is installed on a new patient.

An Automated Insulin Delivery System with Automatic Meal Bolus Based on a Hand-Gesturing Algorithm

Background: Carbohydrate counting (CC) and meal announcements, before eating, introduce a significant burden for individuals managing type 1 diabetes (T1D). An automated insulin delivery system with automatic bolusing that eliminates the need for CC and premeal bolusing (i.e., a hands-free closed-loop [HFCL] system) was assessed in a feasibility trial of adults with T1D. Methods: The system included the MiniMed™ 780G pump and a smartphone-paired smartwatch with the Klue application (Klue, Inc.) that detects eating and drinking gestures. A smartphone algorithm converted gestures into carb amounts that were transmitted to the pump for automatic bolusing. For 5 days, participants (N = 17, 18-75 years of age) used the system at home with meal announcements based on traditional CC, with the Klue application disabled (Home-stay phase). Thereafter, participants moved to a supervised hotel setting, where the Klue application was enabled for 5 days and meals were not announced (Hotel-stay phase). Participants consumed the same eight test meals (six solid and two liquid) of varying caloric and carb size at the same time and day of the week for both phases, and glycemic metrics were compared. Otherwise, there were no other meal restrictions. Results: The overall time in range (70-180 mg/dL) was 83.4% ± 7.0% and 80.6% ± 6.7% for the Home-stay and Hotel-stay, respectively (P = 0.08). The average time at <70 mg/dL was 3.1% and 3.0% (P = 0.9144), respectively, and the average time at >180 mg/dL was 13.5% and 16.3% (P = 0.1046), respectively. Postprandial glycemia following low-carb test meals was similar between the two phases. The system's ability to accommodate high-carb meals was somewhat limited. There were no episodes of severe hypoglycemia or diabetic ketoacidosis. Conclusion: Preliminary findings show that a HFCL system was safe and maintained overall glycemic control, similar to that observed with traditional CC and manual meal bolusing. By eliminating these daily T1D burdens, a HFCL system may improve quality of life for individuals with T1D. ClinicalTrials.gov number: NCT04964128.

ReplayBG: A Digital Twin-Based Methodology to Identify a Personalized Model From Type 1 Diabetes Data and Simulate Glucose Concentrations to Assess Alternative Therapies

OBJECTIVE

Design and assessment of new therapies for type 1 diabetes (T1D) management can be greatly facilitated by in silico simulations. The ReplayBG simulation methodology here proposed allows "replaying" the scenario behind data already collected by simulating the glucose concentration obtained in response to alternative insulin/carbohydrate therapies and evaluate their efficacy leveraging the concept of digital twin.

METHODS

ReplayBG is based on two steps. First, a personalized model of glucose-insulin dynamics is identified using insulin, carbohydrate, and continuous glucose monitoring (CGM) data. Then, this model is used to simulate the glucose concentration that would have been obtained by "replaying" the same portion of data using a different therapy. The validity of the methodology was evaluated on 100 virtual subjects using the UVa/Padova T1D Simulator (T1DS). In particular, the glucose concentration traces simulated by ReplayBG are compared with those provided by T1DS in five different scenarios of insulin and carbohydrate treatment modifications. Furthermore, we compared ReplayBG with a state-of-the-art methodology for the scope. Finally, two case studies using real data are also presented.

RESULTS

ReplayBG simulates with high accuracy the effect of the considered insulin and carbohydrate treatment alterations, performing significantly better than state-of-art method in almost all considered situations.

CONCLUSION

ReplayBG proved to be a reliable and robust tool to retrospectively explore the effect of new treatments for T1D on the glucose dynamics. It is freely available as open source software at https://github.com/gcappon/replay-bg.

SIGNIFICANCE

ReplayBG offers a new approach to preliminary evaluate new therapies for T1D management before clinical trials.

DLMBHCO: design of an augmented bioinspired deep learning-based multidomain body parameter analysis via heterogeneous correlative body organ analysis

Progressive organ-level disorders in the human body are often correlated with diseases in other body parts. For instance, liver diseases can be linked with heart issues, while cancers can be linked with brain diseases (or psychological conditions). Defining such correlations is a complex task, and existing deep learning models that perform this task either showcase lower accuracy or are non-comprehensive when applied to real-time scenarios. To overcome these issues, this text proposes design of an augmented bioinspired deep learning-based multidomain body parameter analysis via heterogeneous correlative body organ analysis. The proposed model initially collects temporal and spatial data scans for different body parts and uses a multidomain feature extraction engine to convert these scans into vector sets. These vectors are processed by a Bacterial Foraging Optimizer (BFO), which assists in identification of highly variant feature sets, which are individually classified into different disease categories. A fusion of Inception Net, XCeption Net, and GoogLeNet Models is used to perform these classifications. The classified categories are linked with other disease types via temporal analysis of blood reports. The temporal analysis engine uses Modified Analytical Hierarchical Processing (MAHP) Model for calculating inter-organ disease dependency probabilities. Based on these probabilities, the model is able to generate a patient-level correlation map, which can be used by clinical experts to suggest remedial treatments, due to which the model was able to identify correlations between brain disorders and kidneys, heart diseases and lungs, heart diseases and liver, brain diseases and different types of cancers with high efficiency when evaluated under clinical scenarios. When validated on MITBIH, DEAP, CT Kidney, RIDER, and PLCO data samples, it was observed that the proposed model was capable of improving accuracy of correlation by 8.5%, while improving precision and recall by 3.2% when compared with existing correlation models under similar clinical scenarios.

Enabling fully automated insulin delivery through meal detection and size estimation using Artificial Intelligence

We present a robust insulin delivery system that includes automated meal detection and carbohydrate content estimation using machine learning for meal insulin dosing called robust artificial pancreas (RAP). We conducted a randomized, single-center crossover trial to compare postprandial glucose control in the four hours following unannounced meals using a hybrid model predictive control (MPC) algorithm and the RAP system. The RAP system includes a neural network model to automatically detect meals and deliver a recommended meal insulin dose. The meal detection algorithm has a sensitivity of 83.3%, false discovery rate of 16.6%, and mean detection time of 25.9 minutes. While there is no significant difference in incremental area under the curve of glucose, RAP significantly reduces time above range (glucose >180 mg/dL) by 10.8% (P = 0.04) and trends toward increasing time in range (70-180 mg/dL) by 9.1% compared with MPC. Time below range (glucose <70 mg/dL) is not significantly different between RAP and MPC.

Data-Driven Supervised Compression Artifacts Detection on Continuous Glucose Sensors

Continuous Glucose Monitoring (CGM) sensors micro-invasively provide frequent glucose readings, improving the management of Type 1 diabetic patients' life and making available reach data-sets for retrospective analysis. Unlikely, CGM sensors are subject to failures, such as compression artifacts, that might impact on both real-time and respective CGM use. In this work is focused on retrospective detection of compression artifacts. An in-silico dataset is generated using the T1D UVa/Padova simulator and compression artifacts are subsequently added in known position, thus creating a dataset with perfectly accurate faulty/not-faulty labels. The problem of compression artifact detection is then faced with supervised data-driven techniques, in particular using Random Forest algorithm. The detection performance guaranteed by the method on in-silico data is satisfactory, opening the way for further analysis on real-data.

Super-twisting-based meal detector for type 1 diabetes management: Improvement and assessment in a real-life scenario

BACKGROUND AND OBJECTIVE

Hybrid automated insulin delivery systems rely on carbohydrate counting to improve postprandial control in type 1 diabetes. However, this is an extra burden on subjects, and it introduces a source of potential errors that could impact control performances. In fact, carbohydrates estimation is challenging, prone to errors, and it is known that subjects sometimes struggle to adhere to this requirement, forgetting to perform this task. A possible solution is the use of automated meal detection algorithms. In this work, we extended a super-twisting-based meal detector suggested in the literature and assessed it on real-life data.

METHODS

To reduce the false detections in the original meal detector, we implemented an implicit discretization of the super-twisting and replaced the Euler approximation of the glucose derivative with a Kalman filter. The modified meal detector is retrospectively evaluated in a challenging real-life dataset corresponding to a 2-week trial with 30 subjects using sensor-augmented pump control. The assessment includes an analysis of the nature and riskiness of false detections.

RESULTS

The proposed algorithm achieved a recall of 70 [13] % (median [interquartile range]), a precision of 73 [26] %, and had 1.4 [1.4] false positives-per-day. False positives were related to rising glucose conditions, whereas false negatives occurred after calibrations, missing samples, or hypoglycemia treatments.

CONCLUSIONS

The proposed algorithm achieves encouraging performance. Although false positives and false negatives were not avoided, they are related to situations with a low risk of hypoglycemia and hyperglycemia, respectively.

Machine Learning-based Meal Detection Using Continuous Glucose Monitoring on Healthy Participants: An Objective Measure of Participant Compliance to Protocol

Meal timing affects metabolic responses to diet, but participant compliance in time-restricted feeding and other diet studies is challenging to monitor and is a major concern for research rigor and reproducibility. To facilitate automated validation of participant self-reports of meal timing, the present study focuses on the creation of a meal detection algorithm using continuous glucose monitoring (CGM), physiological monitors and machine learning. While most CGM-related studies focus on participants who are diabetic, this study is the first to apply machine learning to meal detection using CGM in metabolically healthy adults. Furthermore, the results demonstrate a high area under the receiver operating characteristic curve (AUC-ROC) and precision-recall curve (AUC-PR). A cold-start simulation using a random forest algorithm yields .891 and .803 for AUC-ROC and AUC-PR respectively on 110-minutes data, and a non-cold start simulation using a gradient boosted tree model yields over .996 (AUC-ROC) and .964 (AUC-PR). Here it is demonstrated that CGM and physiological monitoring data is a viable tool for practitioners and scientists to objectively validate self-reports of meal consumption in healthy participants.

Fault Tolerant Strategies for Automated Insulin Delivery Considering the Human Component: Current and Future Perspectives

Automated Insulin Delivery (AID) are systems developed for daily use by people with type 1 diabetes (T1D). To ensure the safety of users, it is essential to consider how the human factor affects the performance and safety of these devices. While there are numerous publications on hardware-related failures of AID systems, there are few studies on the human component of the system. From a control point of view, people with T1D using AID systems are at the same time the plant to be controlled and the plant operator. Therefore, users may induce faults in the controller, sensors, actuators, and the plant itself. Strategies to cope with the human interaction in AID systems are needed for further development of the technology. In this paper, we present an analysis of potential faults introduced by AID users when the system is under normal operation. This is followed by a review of current fault tolerant control (FTC) approaches to identify missing areas of research. The paper concludes with a discussion on future directions for the new generation of FTC AID systems.

Wearable Glucose Monitoring and Implantable Drug Delivery Systems for Diabetes Management

The global cost of diabetes care exceeds $1 trillion each year with more than $327 billion being spent in the United States alone. Despite some of the advances in diabetes care including continuous glucose monitoring systems and insulin pumps, the technology associated with managing diabetes has largely remained unchanged over the past several decades. With the rise of wearable electronics and novel functional materials, the field is well-poised for the next generation of closed-loop diabetes care. Wearable glucose sensors implanted within diverse platforms including skin or on-tooth tattoos, skin-mounted patches, eyeglasses, contact lenses, fabrics, mouthguards, and pacifiers have enabled noninvasive, unobtrusive, and real-time analysis of glucose excursions in ambulatory care settings. These wearable glucose sensors can be integrated with implantable drug delivery systems, including an insulin pump, glucose responsive insulin release implant, and islets transplantation, to form self-regulating closed-loop systems. This review article encompasses the emerging trends and latest innovations of wearable glucose monitoring and implantable insulin delivery technologies for diabetes management with a focus on their advanced materials and construction. Perspectives on the current unmet challenges of these strategies are also discussed to motivate future technological development toward improved patient care in diabetes management.