Meal Detection and Carbohydrate Estimation Using Continuous Glucose Sensor Data

Mealtime prediction using wearable insulin pump data to support diabetes management

Many patients with diabetes struggle with post-meal high blood glucose due to missed or untimely meal-related insulin doses. To address this challenge, our research aims to: (1) study mealtime patterns in patients with type 1 diabetes using wearable insulin pump data, and (2) develop personalized models for predicting future mealtimes to support timely insulin dose administration. Using two independent datasets with over 45,000 meal logs from 82 patients with diabetes, we find that the majority of people ( ∼ 60%) have irregular and inconsistent mealtime patterns that change notably through the course of each day and across months in their own historical data. We also show the feasibility of predicting future mealtimes with personalized LSTM-based models that achieve an average F1 score of > 95% with less than 0.25 false positives per day. Our research lays the groundwork for developing a meal prediction system that can nudge patients with diabetes to administer bolus insulin doses before meal consumption to reduce the occurrence of post-meal high blood glucose.

Assessing Carbohydrate Counting Accuracy: Current Limitations and Future Directions

Diabetes mellitus is a prevalent chronic autoimmune disease with a high impact on global health, affecting millions of adults and resulting in significant morbidity and mortality. Achieving optimal blood glucose levels is crucial for diabetes management to prevent acute and long-term complications. Carbohydrate counting (CC) is widely used by patients with type 1 diabetes to adjust prandial insulin bolus doses based on estimated carbohydrate content, contributing to better glycemic control and improved quality of life. However, accurately estimating the carbohydrate content of meals remains challenging for patients, leading to errors in bolus insulin dosing. This review explores the current limitations and challenges in CC accuracy and emphasizes the importance of personalized educational programs to enhance patients' abilities in carbohydrate estimation. Existing tools for assessing patient learning outcomes in CC are discussed, highlighting the need for individualized approaches tailored to each patient's needs. A comprehensive review of the relevant literature was conducted to identify educational programs and assessment tools dedicated to training diabetes patients on carbohydrate counting. The research aims to provide insights into the benefits and limitations of existing tools and identifies future research directions to advance personalized CC training approaches. By adopting a personalized approach to CC education and assessment, healthcare professionals can empower patients to achieve better glycemic control and improve diabetes management. Moreover, this review identifies potential avenues for future research, paving the way for advancements in personalized CC training and assessment approaches and further enhancing diabetes management strategies.

An automatic deep reinforcement learning bolus calculator for automated insulin delivery systems

In hybrid automatic insulin delivery (HAID) systems, meal disturbance is compensated by feedforward control, which requires the announcement of the meal by the patient with type 1 diabetes (DM1) to achieve the desired glycemic control performance. The calculation of insulin bolus in the HAID system is based on the amount of carbohydrates (CHO) in the meal and patient-specific parameters, i.e. carbohydrate-to-insulin ratio (CR) and insulin sensitivity-related correction factor (CF). The estimation of CHO in a meal is prone to errors and is burdensome for patients. This study proposes a fully automatic insulin delivery (FAID) system that eliminates patient intervention by compensating for unannounced meals. This study exploits the deep reinforcement learning (DRL) algorithm to calculate insulin bolus for unannounced meals without utilizing the information on CHO content. The DRL bolus calculator is integrated with a closed-loop controller and a meal detector (both previously developed by our group) to implement the FAID system. An adult cohort of 68 virtual patients based on the modified UVa/Padova simulator was used for in-silico trials. The percentage of the overall duration spent in the target range of 70-180 mg/dL was 71.2 % and 76.2 % , < 70 mg/dL was 0.9 % and 0.1 % , and > 180 mg/dL was 26.7 % and 21.1 % , respectively, for the FAID system and HAID system utilizing a standard bolus calculator (SBC) including CHO misestimation. The proposed algorithm can be exploited to realize FAID systems in the future.

Continuous glucose monitoring as an objective measure of meal consumption in individuals with binge-spectrum eating disorders: A proof-of-concept study

OBJECTIVE

Going extended periods of time without eating increases risk for binge eating and is a primary target of leading interventions for binge-spectrum eating disorders (B-EDs). However, existing treatments for B-EDs yield insufficient improvements in regular eating and subsequently, binge eating. These unsatisfactory clinical outcomes may result from limitations in assessment and promotion of regular eating in therapy. Detecting the absence of eating using passive sensing may improve clinical outcomes by facilitating more accurate monitoring of eating behaviours and powering just-in-time adaptive interventions. We developed an algorithm for detecting meal consumption (and extended periods without eating) using continuous glucose monitor (CGM) data and machine learning.

METHOD

Adults with B-EDs (N = 22) wore CGMs and reported eating episodes on self-monitoring surveys for 2 weeks. Random forest models were run on CGM data to distinguish between eating and non-eating episodes.

RESULTS

The optimal model distinguished eating and non-eating episodes with high accuracy (0.82), sensitivity (0.71), and specificity (0.94).

CONCLUSIONS

These findings suggest that meal consumption and extended periods without eating can be detected from CGM data with high accuracy among individuals with B-EDs, which may improve clinical efforts to target dietary restriction and improve the field's understanding of its antecedents and consequences.

Recent advances in the precision control strategy of artificial pancreas

The scientific diagnosis and treatment of patients with diabetes require frequent blood glucose testing and insulin delivery to normoglycemia. Therefore, an artificial pancreas with a continuous blood glucose (BG) monitoring function is an urgent research target in the medical industry. The problem of closed-loop algorithmic control of the BG with a time delay is a key and difficult issue that needs to be overcome in the development of an artificial pancreas. Firstly, the composition, structure, and control characteristics of the artificial pancreas are introduced. Subsequently, the research progress of artificial pancreas control algorithms is reviewed, and the characteristics, advantages, and disadvantages of proportional-integral-differential control, model predictive control, and artificial intelligence control are compared and analyzed to determine whether they are suitable for the practical application of the artificial pancreas. Additionally, key advancements in areas such as blood glucose data monitoring, adaptive models, wearable devices, and fully automated artificial pancreas systems are also reviewed. Finally, this review highlights that meal prediction, control safety, integration, streamlining the optimization of control algorithms, constant temperature preservation of insulin, and dual-hormone artificial pancreas are issues that require further attention in the future.

An ensemble machine learning approach for the detection of unannounced meals to enhance postprandial glucose control

BACKGROUND

Hybrid automated insulin delivery systems enhance postprandial glucose control in type 1 diabetes, however, meal announcements are burdensome. To overcome this, we propose a machine learning-based automated meal detection approach; METHODS:: A heterogeneous ensemble method combining an artificial neural network, random forest, and logistic regression was employed. Trained and tested on data from two in-silico cohorts comprising 20 and 47 patients. It accounted for various meal sizes (moderate to high) and glucose appearance rates (slow and rapid absorbing). To produce an optimal prediction model, three ensemble configurations were used: logical AND, majority voting, and logical OR. In addition to the in-silico data, the proposed meal detector was also trained and tested using the OhioT1DM dataset. Finally, the meal detector is combined with a bolus insulin compensation scheme; RESULTS:: The ensemble majority voting obtained the best meal detector results for both the in-silico and OhioT1DM cohorts with a sensitivity of 77%, 94%, 61%, precision of 96%, 89%, 72%, F1-score of 85%, 91%, 66%, and with false positives per day values of 0.05, 0.19, 0.17, respectively. Automatic meal detection with insulin compensation has been performed in open-loop insulin therapy using the AND ensemble, chosen for its lower false positive rate. Time-in-range has significantly increased 10.48% and 16.03%, time above range was reduced by 5.16% and 11.85%, with a minimal time below range increase of 0.35% and 2.69% for both in-silico cohorts, respectively, compared to the results without a meal detector; CONCLUSION:: To increase the overall accuracy and robustness of the predictions, this ensemble methodology aims to take advantage of each base model's strengths. All of the results point to the potential application of the proposed meal detector as a separate module for the detection of meals in automated insulin delivery systems to achieve improved glycemic control.

Integrated image and sensor-based food intake detection in free-living

The first step in any dietary monitoring system is the automatic detection of eating episodes. To detect eating episodes, either sensor data or images can be used, and either method can result in false-positive detection. This study aims to reduce the number of false positives in the detection of eating episodes by a wearable sensor, Automatic Ingestion Monitor v2 (AIM-2). Thirty participants wore the AIM-2 for two days each (pseudo-free-living and free-living). The eating episodes were detected by three methods: (1) recognition of solid foods and beverages in images captured by AIM-2; (2) recognition of chewing from the AIM-2 accelerometer sensor; and (3) hierarchical classification to combine confidence scores from image and accelerometer classifiers. The integration of image- and sensor-based methods achieved 94.59% sensitivity, 70.47% precision, and 80.77% F1-score in the free-living environment, which is significantly better than either of the original methods (8% higher sensitivity). The proposed method successfully reduces the number of false positives in the detection of eating episodes.

Continuous glucose monitoring for automatic real-time assessment of eating events and nutrition: a scoping review

Background

Accurate dietary assessment remains a challenge, particularly in free-living settings. Continuous glucose monitoring (CGM) shows promise in optimizing the assessment and monitoring of ingestive activity (IA, i.e., consumption of calorie-containing foods/beverages), and it might enable administering dietary Just-In-Time Adaptive Interventions (JITAIs).

Objective

In a scoping review, we aimed to answer the following questions: (1) Which CGM approaches to automatically detect IA in (near-)real-time have been investigated? (2) How accurate are these approaches? (3) Can they be used in the context of JITAIs?

Methods

We systematically searched four databases until October 2023 and included publications in English or German that used CGM-based approaches for human (all ages) IA detection. Eligible publications included a ground-truth method as a comparator. We synthesized the evidence qualitatively and critically appraised publication quality.

Results

Of 1,561 potentially relevant publications identified, 19 publications (17 studies, total N = 311; for 2 studies, 2 publications each were relevant) were included. Most publications included individuals with diabetes, often using meal announcements and/or insulin boluses accompanying meals. Inpatient and free-living settings were used. CGM-only approaches and CGM combined with additional inputs were deployed. A broad range of algorithms was tested. Performance varied among the reviewed methods, ranging from unsatisfactory to excellent (e.g., 21% vs. 100% sensitivity). Detection times ranged from 9.0 to 45.0 min.

Conclusion

Several CGM-based approaches are promising for automatically detecting IA. However, response times need to be faster to enable JITAIs aimed at impacting acute IA. Methodological issues and overall heterogeneity among articles prevent recommending one single approach; specific cases will dictate the most suitable approach.

T1DiabetesGranada: a longitudinal multi-modal dataset of type 1 diabetes mellitus

Type 1 diabetes mellitus (T1D) patients face daily difficulties in keeping their blood glucose levels within appropriate ranges. Several techniques and devices, such as flash glucose meters, have been developed to help T1D patients improve their quality of life. Most recently, the data collected via these devices is being used to train advanced artificial intelligence models to characterize the evolution of the disease and support its management. Data scarcity is the main challenge for generating these models, as most works use private or artificially generated datasets. For this reason, this work presents T1DiabetesGranada, an open under specific permission longitudinal dataset that not only provides continuous glucose levels, but also patient demographic and clinical information. The dataset includes 257 780 days of measurements spanning four years from 736 T1D patients from the province of Granada, Spain. This dataset advances beyond the state of the art as one the longest and largest open datasets of continuous glucose measurements, thus boosting the development of new artificial intelligence models for glucose level characterization and prediction.

Data-driven meal events detection using blood glucose response patterns

BACKGROUND

In the Diabetes domain, events such as meals and exercises play an important role in the disease management. For that, many studies focus on automatic meal detection, specially as part of the so-called artificial [Formula: see text]-cell systems. Meals are associated to blood glucose (BG) variations, however such variations are not peculiar to meals, it mostly comes as a combination of external factors. Thus, general approaches such as the ones focused on glucose signal rate of change are not enough to detect personalized influence of such factors. By using a data-driven individualized approach for meal detection, our method is able to fit real data, detecting personalized meal responses even when such external factors are implicitly present.

METHODS

The method is split into model training and selection. In the training phase, we start observing meal responses for each individual, and identifying personalized patterns. Occurrences of such patterns are searched over the BG signal, evaluating the similarity of each pattern to each possible signal subsequence. The most similar occurrences are then selected as possible meal event candidates. For that, we include steps for excluding less relevant neighbors per pattern, and grouping close occurrences in time globally. Each candidate is represented by a set of time and response signal related qualitative variables. These variables are used as input features for different binary classifiers in order to learn to classify a candidate as MEAL or NON-MEAL. In the model selection phase, we compare all trained classifiers to select the one that performs better with the data of each individual.

RESULTS

The results show that the method is able to detect daily meals, providing a result with a balanced proportion between detected meals and false alarms. The analysis on multiple patients indicate that the approach achieves good outcomes when there is enough reliable training data, as this is reflected on the testing results.

CONCLUSIONS

The approach aims at personalizing the meal detection task by relying solely on data. The premise is that a model trained with data that contains the implicit influence of external factors is able to recognize the nuances of the individual that generated the data. Besides, the approach can also be used to improve data quality by detecting meals, opening opportunities to possible applications such as detecting and reminding users of missing or wrongly informed meal events.

Multivariable identification based MPC for closed-loop glucose regulation subject to individual variability

The controller is important for the artificial pancreas to guide insulin infusion in diabetic therapy. However, the inter- and intra-individual variability and time delay of glucose metabolism bring challenges to control glucose within a normal range. In this study, a multivariable identification based model predictive control (mi-MPC) is developed to overcome the above challenges. Firstly, an integrated glucose-insulin model is established to describe insulin absorption, glucose-insulin interaction under meal disturbance, and glucose transport. On this basis, an observable glucose-insulin dynamic model is formed, in which the individual parameters and disturbances can be identified by designing a particle filtering estimator. Next, embedded with the identified glucose-insulin dynamic model, a mi-MPC method is proposed. In this controller, plasma glucose concentration (PGC), an important variable and indicator of glucose regulation, is estimated and controlled directly. Finally, the method was tested on 30 in-silico subjects produced by the UVa/Padova simulator. The results show that the mi-MPC method including the model, individual identification, and the controller can regulate glucose with the mean value of 7.45 mmol/L without meal announcement.

Detection of Meals and Physical Activity Events From Free-Living Data of People With Diabetes

BACKGROUND

Predicting carbohydrate intake and physical activity in people with diabetes is crucial for improving blood glucose concentration regulation. Patterns of individual behavior can be detected from historical free-living data to predict meal and exercise times. Data collected in free-living may have missing values and forgotten manual entries. While machine learning (ML) can capture meal and exercise times, missing values, noise, and errors in data can reduce the accuracy of ML algorithms.

METHODS

Two recurrent neural networks (RNNs) are developed with original and imputed data sets to assess detection accuracy of meal and exercise events. Continuous glucose monitoring (CGM) data, insulin infused from pump data, and manual meal and exercise entries from free-living data are used to predict meals, exercise, and their concurrent occurrence. They contain missing values of various lengths in time, noise, and outliers.

RESULTS

The accuracy of RNN models range from 89.9% to 95.7% for identifying the state of event (meal, exercise, both, or neither) for various users. "No meal or exercise" state is determined with 94.58% accuracy by using the best RNN (long short-term memory [LSTM] with 1D Convolution). Detection accuracy with this RNN is 98.05% for meals, 93.42% for exercise, and 55.56% for concurrent meal-exercise events.

CONCLUSIONS

The meal and exercise times detected by the RNN models can be used to warn people for entering meal and exercise information to hybrid closed-loop automated insulin delivery systems. Reliable accuracy for event detection necessitates powerful ML and large data sets. The use of additional sensors and algorithms for detecting these events and their characteristics provides a more accurate alternative.

Multivariable Automated Insulin Delivery System for Handling Planned and Spontaneous Physical Activities

BACKGROUND

Hybrid closed-loop control of glucose levels in people with type 1 diabetes mellitus (T1D) is limited by the requirements on users to manually announce physical activity (PA) and meals to the artificial pancreas system. Multivariable automated insulin delivery (mvAID) systems that can handle unannounced PAs and meals without any manual announcements by the user can improve glycemic control by modulating insulin dosing in response to the occurrence and intensity of spontaneous physical activities.

METHODS

An mvAID system is developed to supplement the glucose measurements with additional physiological signals from a wristband device, with the signals analyzed using artificial intelligence algorithms to automatically detect the occurrence of PA and estimate its intensity. This additional information gained from the physiological signals enables more proactive insulin dosing adjustments in response to both planned exercise and spontaneous unanticipated physical activities.

RESULTS

In silico studies of the mvAID illustrate the safety and efficacy of the system. The mvAID is translated to pilot clinical studies to assess its performance, and the clinical experiments demonstrate an increased time in range and reduced risk of hypoglycemia following unannounced PA and meals.

CONCLUSIONS

The mvAID systems can increase the safety and efficacy of insulin delivery in the presence of unannounced physical activities and meals, leading to improved lives and less burden on people with T1D.

A Probabilistic Approach to Blood Glucose Prediction in Type 1 Diabetes Under Meal Uncertainties

Currently, most reliable and commercialized artificial pancreas systems for type 1 diabetes are hybrid closed-loop systems, which require the user to announce every meal and its size. However, estimating the amount of carbohydrates in a meal and announcing each and every meal is an error-prone process that introduces important uncertainties to the problem, which when not considered, lead to sub-optimal outcomes of the controller. To address this problem, we propose a novel deep-learning-based model for probabilistic glucose prediction, called the Input and State Recurrent Kalman Network (ISRKN), which consists in the incorporation of an input and state Kalman filter in the latent space of a deep neural network so that the posterior distributions can be computed in closed form and the uncertainty can be propagated using the Kalman equations. In addition, the proposed architecture allows explicit estimation of the meal uncertainty distribution, whose parameters are encoded in the filter parameters. Results using the UVA/Padova simulator and data from a clinical trial show that the proposed model outperforms other probabilistic models using several probabilistic metrics across different degrees of distributional shifts.

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.

Near-optimal feedback control for postprandial glucose regulation in type 1 diabetes

This paper is focused on feedback control of postprandial glucose levels for patients with type 1 Diabetes Mellitus. There are two important limitations that make this a challenging problem. First, the slow subcutaneous insulin pharmacokinetics that introduces a significant lag into the control loop. Second, the positivity constraint on the control action, meaning that it is not possible to remove insulin from the body. In this paper, both issues are explicitly considered in the design process using the internal model control framework, to derive a near-optimal feedback controller. Optimality is understood here as minimizing the blood glucose peak after a meal intake and, at the same time, preventing glucose values below a prescribed threshold. It is shown how the proposed controller approaches the optimal closed-loop performance as a limit case. The theoretical results are supported by a numerical example and the feasibility of the overall strategy under uncertainties is illustrated using an extended version UVa/Padova metabolic simulator.

Artificial Pancreas: A Review of Meal Detection and Carbohydrates Counting Techniques

OBJECTIVE

The development of an artificial pancreas is an open research problem that faces the challenge of creating a control algorithm capable of dosing insulin automatically and driving blood glucose to healthy levels. Many of these approaches, including artificial intelligence, are based on techniques that could result in and undesirable outcome because most of them include neither detect meal intake or meal size information. To overcome that issue, some meal count-detection algorithms reported in scientific publications have shown not only a good performance on blood glucose regulation but fewer hypoglicemia and hyperglycemia events too.

METHODS

We reviewed the most relevant authors and publications and main databases (particularly SCOPUS and Google Scholar), focusing on algorithms of detection and estimation of meal intake from multiple approaches.

RESULTS

A wide range of approaches and proposals have been found. The majority of them include trials on in silico patients rather than in vivo ones. Most of procedures require as inputs glucose samples from continuous glucose monitoring devices as basal insulin and bolus as well. Most of approaches could be grouped by 2 categories: mathematical model based and not model based.

CONCLUSION

A combination of methods seems to reach better results.

An overview of advancements in closed-loop artificial pancreas system

Type 1 diabetes (T1D) is one of the world's health problems with a prevalence of 1.1 million for children and young adults under the age of 20. T1D is a health problem characterized by autoimmunity and the destruction of pancreatic cells that produce insulin. The available treatment is to maintain blood glucose within the desired normal range. To meet bolus and basal requirements, T1D patients may receive multiple daily injections (MDI) of fast-acting and long-acting insulin once or twice daily. In addition, insulin pumps can deliver multiple doses a day without causing injection discomfort in individuals with T1D. T1D patients have also monitored their blood glucose levels along with insulin replacement treatment using a continuous glucose monitor (CGM). However, this CGM has some drawbacks, like the sensor needs to be replaced after being inserted under the skin for seven days and needs to be calibrated (for some CGMs). The treatments and monitoring devices mentioned creating a lot of workloads to maintain blood glucose levels in individuals with T1D. Therefore, to overcome these problems, closed-loop artificial pancreas (APD) devices are widely used to manage blood glucose in T1D patients. Closed-loop APD consists of a glucose sensor, an insulin infusion device, and a control algorithm. This study reviews the progress of closed-loop artificial pancreas systems from the perspective of device properties, uses, testing procedures, regulations, and current market conditions.

Digital Biomarkers for Personalized Nutrition: Predicting Meal Moments and Interstitial Glucose with Non-Invasive, Wearable Technologies

Digital health technologies may support the management and prevention of disease through personalized lifestyle interventions. Wearables and smartphones are increasingly used to continuously monitor health and disease in everyday life, targeting health maintenance. Here, we aim to demonstrate the potential of wearables and smartphones to (1) detect eating moments and (2) predict and explain individual glucose levels in healthy individuals, ultimately supporting health self-management. Twenty-four individuals collected continuous data from interstitial glucose monitoring, food logging, activity, and sleep tracking over 14 days. We demonstrated the use of continuous glucose monitoring and activity tracking in detecting eating moments with a prediction model showing an accuracy of 92.3% (87.2-96%) and 76.8% (74.3-81.2%) in the training and test datasets, respectively. Additionally, we showed the prediction of glucose peaks from food logging, activity tracking, and sleep monitoring with an overall mean absolute error of 0.32 (+/-0.04) mmol/L for the training data and 0.62 (+/-0.15) mmol/L for the test data. With Shapley additive explanations, the personal lifestyle elements important for predicting individual glucose peaks were identified, providing a basis for personalized lifestyle advice. Pending further validation of these digital biomarkers, they show promise in supporting the prevention and management of type 2 diabetes through personalized lifestyle recommendations.

Meal and Physical Activity Detection from Free-Living Data for Discovering Disturbance Patterns of Glucose Levels in People with Diabetes

Objective: The interpretation of time series data collected in free-living has gained importance in chronic disease management. Some data are collected objectively from sensors and some are estimated and entered by the individual. In type 1 diabetes (T1D), blood glucose concentration (BGC) data measured by continuous glucose monitoring (CGM) systems and insulin doses administered can be used to detect the occurrences of meals and physical activities and generate the personal daily living patterns for use in automated insulin delivery (AID). Methods: Two challenges in time-series data collected in daily living are addressed: data quality improvement and the detection of unannounced disturbances of BGC. CGM data have missing values for varying periods of time and outliers. People may neglect reporting their meal and physical activity information. In this work, novel methods for preprocessing real-world data collected from people with T1D and the detection of meal and exercise events are presented. Four recurrent neural network (RNN) models are investigated to detect the occurrences of meals and physical activities disjointly or concurrently. Results: RNNs with long short-term memory (LSTM) with 1D convolution layers and bidirectional LSTM with 1D convolution layers have average accuracy scores of 92.32% and 92.29%, and outperform other RNN models. The F1 scores for each individual range from 96.06% to 91.41% for these two RNNs. Conclusions: RNNs with LSTM and 1D convolution layers and bidirectional LSTM with 1D convolution layers provide accurate personalized information about the daily routines of individuals. Significance: Capturing daily behavior patterns enables more accurate future BGC predictions in AID systems and improves BGC regulation.

A Deep Learning Framework for Automatic Meal Detection and Estimation in Artificial Pancreas Systems

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, p< 0.01) and 77.8 ± 6.3% mean time between 70 and 180 mg/dL (+3.9%, p< 0.001). This improvement in control is realised without a significant increase in mean in hypoglycaemia (+0.1%, p= 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.

A New Meal Absorption Model for Artificial Pancreas Systems

BACKGROUND

Artificial pancreas (AP) systems reduce the treatment burden of Type 1 Diabetes by automatically regulating blood glucose (BG) levels. While many disturbances stand in the way of fully closed-loop (automated) control, unannounced meals remain the greatest challenge. Furthermore, different types of meals can have significantly different glucose responses, further increasing the uncertainty surrounding the meal.

METHODS

Effective attenuation of a meal requires quick and accurate insulin delivery because of slow insulin action relative to meal effects on BG. The proposed Variable Hump (VH) model adapts to meals of varying compositions by inferring both meal size and shape. To appropriately address the uncertainty of meal size, the model divides meal absorption into two disjoint regions: a region with coarse meal size predictions followed by a fine-grain region where predictions are fine-tuned by adapting to the meal shape.

RESULTS

Using gold-standard triple tracer meal data, the proposed VH model is compared to three simpler second-order response models. The proposed VH model increased model fit capacity by 22% and prediction accuracy by 12% relative to the next best models. A 47% increase in the accuracy of uncertainty predictions was also found. In a simple control scenario, the controller governed by the proposed VH model provided insulin just as fast or faster than the controller governed by the other models in four out of the six meals. While the controllers governed by the other models all delivered at least a 25% excess of insulin at their worst, the VH model controller only delivered 9% excess at its worst.

CONCLUSIONS

The VH Model performed best in accuracy metrics and succeeded over the other models in providing insulin quickly and accurately in a simple implementation. Use in an AP system may improve prediction accuracy and lead to better control around mealtimes.

An OGI model for personalized estimation of glucose and insulin concentration in plasma

Plasma glucose concentration (PGC) and plasma insulin concentration (PIC) are two essential metrics for diabetic regulation, but difficult to be measured directly. Often, PGC and PIC are estimated from continuous glucose monitoring and insulin delivery data. Nevertheless, the inter-individual variability and external disturbance (e.g. carbohydrate intake) bring challenges for accurate estimations. This study is to estimate PGC and PIC adaptively by identifying personalized parameters and external disturbances. An observable glucose-insulin (OGI) dynamic model is established to describe insulin absorption, glucose regulation, and glucose transport. The model parameters and disturbances can be extended to observable state variables and be identified dynamically by Bayesian filtering estimators. Two basic Gaussian noise based Bayesian filtering estimators, extended Kalman filtering (EKF) and unscented Kalman filtering (UKF), are implemented. Recognizing the prevalence of non-Gaussian noise, in this study, two new filtering estimators: particle filtering with Gaussian noise (PFG), and particle filtering with mixed non-Gaussian noise (PFM) are designed and implemented. The proposed OGI model in conjunction with the estimators is evaluated using the data from 30 in-silico subjects and 10 human participants. For in-silico subjects, the OGI with PFM estimator has the ability to estimate PIC and PGC adaptively, achieving RMSE of PIC 9.49±3.81 mU/L, and PGC 0.89±0.19 mmol/L. For human, the OGI with PFM has the promise to identify disturbances (95.46%±0.65% accurate rate of meal identification). OGI model provides a way to fully personalize the parameters and external disturbances in real time, and has potential clinical utility for artificial pancreas.

Diabetes Technology Meeting 2020

Diabetes Technology Society hosted its annual Diabetes Technology Meeting on November 12 to November 14, 2020. This meeting brought together speakers to cover various perspectives about the field of diabetes technology. The meeting topics included artificial intelligence, digital health, telemedicine, glucose monitoring, regulatory trends, metrics for expressing glycemia, pharmaceuticals, automated insulin delivery systems, novel insulins, metrics for diabetes monitoring, and discriminatory aspects of diabetes technology. A live demonstration was presented.

Physical Activity and Psychological Stress Detection and Assessment of Their Effects on Glucose Concentration Predictions in Diabetes Management

OBJECTIVE

Continuous glucose monitoring (CGM) enables prediction of the future glucose concentration (GC) trajectory for making informed diabetes management decisions. The glucose concentration values are affected by various physiological and metabolic variations, such as physical activity (PA) and acute psychological stress (APS), in addition to meals and insulin. In this work, we extend our adaptive glucose modeling framework to incorporate the effects of PA and APS on the GC predictions.

METHODS

A wristband conducive of use by free-living ambulatory people is used. The measured physiological variables are analyzed to generate new quantifiable input features for PA and APS. Machine learning techniques estimate the type and intensity of the PA and APS when they occur individually and concurrently. Variables quantifying the characteristics of both PA and APS are integrated as exogenous inputs in an adaptive system identification technique for enhancing the accuracy of GC predictions. Data from clinical experiments illustrate the improvement in GC prediction accuracy.

RESULTS

The average mean absolute error (MAE) of one-hour-ahead GC predictions with testing data decreases from 35.1 to 31.9 mg/dL (p-value = 0.01) with the inclusion of PA information, and it decreases from 16.9 to 14.2 mg/dL (p-value = 0.006) with the inclusion of PA and APS information.

CONCLUSION

The first-ever glucose prediction model is developed that incorporates measures of physical activity and acute psychological stress to improve GC prediction accuracy.

SIGNIFICANCE

Modeling the effects of physical activity and acute psychological stress on glucose concentration values will improve diabetes management and enable informed meal, activity and insulin dosing decisions.

Evaluation of Meal Carbohydrate Counting Errors in Patients with Type 1 Diabetes

AIM

Correct estimation of meal carbohydrate content is a prerequisite for successful intensified insulin therapy in patients with diabetes. In this survey, the counting error in adult patients with type 1 diabetes was investigated.

METHODS

Seventy-four patients with type 1 diabetes estimated the carbohydrate content of 24 standardized test meals. The test meals were categorized into 1 of 3 groups with different carbohydrate content: low, medium, and high. Estimation results were compared with the meals' actual carbohydrate content as determined by calculation based on weighing. A subgroup of the participants estimated the test meals for a second (n=35) and a third time (n=22) with a mean period of 11 months between the estimations.

RESULTS

During the first estimation, the carbohydrate content was underestimated by -28% (-50, 0) of the actual carbohydrate content. Particularly meals with high mean carbohydrate content were underestimated by -34% (-56, -13). Median counting error improved significantly when estimations were performed for a second time (p<0.001).

CONCLUSIONS

Participants generally underestimated the carbohydrate content of the test meals, especially in meals with higher carbohydrate content. Repetition of estimation resulted in significant improvements in estimation accuracy and is important for the maintenance of correct carbohydrate estimations. The ability to estimate the carbohydrate content of a meal should be checked and trained regularly in patients with diabetes.

Microcontroller Implementation of Support Vector Machine for Detecting Blood Glucose Levels Using Breath Volatile Organic Compounds

This paper presents an embedded system-based solution for sensor arrays to estimate blood glucose levels from volatile organic compounds (VOCs) in a patient's breath. Support vector machine (SVM) was trained on a general-purpose computer using an existing SVM library. A training model, optimized to achieve the most accurate results, was implemented in a microcontroller with an ATMega microprocessor. Training and testing was conducted using artificial breath that mimics known VOC footprints of high and low blood glucose levels. The embedded solution was able to correctly categorize the corresponding glucose levels of the artificial breath samples with 97.1% accuracy. The presented results make a significant contribution toward the development of a portable device for detecting blood glucose levels from a patient's breath.

Plasma-Insulin-Cognizant Adaptive Model Predictive Control for Artificial Pancreas Systems

An adaptive model predictive control (MPC) algorithm with dynamic adjustments of constraints and objective function weights based on estimates of the plasma insulin concentration (PIC) is proposed for artificial pancreas (AP) systems. A personalized compartment model that translates the infused insulin into estimates of PIC is integrated with a recursive subspace-based system identification to characterize the transient dynamics of glycemic measurements. The system identification approach is able to identify stable, reliable linear time-varying models from closed-loop data. An MPC algorithm using the adaptive models is designed to compute the optimal exogenous insulin delivery for AP systems without requiring any manually-entered meal information. A dynamic safety constraint derived from the estimation of PIC is incorporated in the adaptive MPC to improve the efficacy of the AP and prevent insulin overdosing. Simulation case studies demonstrate the performance of the proposed adaptive MPC algorithm.

Pattern Recognition Reveals Characteristic Postprandial Glucose Changes: Non-Individualized Meal Detection in Diabetes Mellitus Type 1

Accurate continuous glucose monitoring (CGM) is essential for fully automated glucose control in diabetes mellitus type 1. State-of-the-art glucose control systems automatically regulate the basal insulin infusion. Users still need to manually announce meals to dose the prandial insulin boluses. An automated meal detection could release the user and improve the glucose regulation. In this study, patterns in the postprandial CGM data are exploited for meal detection. Binary classifiers are trained to recognize the postprandial pattern in horizons of the estimated glucose rate of appearance and in CGM data. The appearance rate is determined by moving horizon estimation based on a simple model. Linear discriminant analysis (LDA) is used for classification. The proposed method is compared to methods that detect meals when thresholds are violated. Diabetes care data from 12 free-living pediatric patients was downloaded during regular screening. Experts identified meals and their start by retrospective evaluation. The classification was tested by cross-validation. Compared to the threshold-based methods, LDA showed higher sensitivity to meals with a low rate of false alarms. Classifying horizons outperformed the other methods also with respect to time of detection. The onset of meals can be detected by pattern recognition based on estimated model states and consecutive CGM measurements. No individual tuning is necessary. This makes the method easily adopted in the clinical practice.

Artificial Pancreas: Clinical Study in Latin America Without Premeal Insulin Boluses

BACKGROUND

Emerging therapies such as closed-loop (CL) glucose control, also known as artificial pancreas (AP) systems, have shown significant improvement in type 1 diabetes mellitus (T1DM) management. However, demanding patient intervention is still required, particularly at meal times. To reduce treatment burden, the automatic regulation of glucose (ARG) algorithm mitigates postprandial glucose excursions without feedforward insulin boluses. This work assesses feasibility of this new strategy in a clinical trial.

METHODS

A 36-hour pilot study was performed on five T1DM subjects to validate the ARG algorithm. Subjects wore a subcutaneous continuous glucose monitor (CGM) and an insulin pump. Insulin delivery was solely commanded by the ARG algorithm, without premeal insulin boluses. This was the first clinical trial in Latin America to validate an AP controller.

RESULTS

For the total 36-hour period, results were as follows: average time of CGM readings in range 70-250 mg/dl: 88.6%, in range 70-180 mg/dl: 74.7%, <70 mg/dl: 5.8%, and <50 mg/dl: 0.8%. Results improved analyzing the final 15-hour period of this trial. In that case, the time spent in range was 70-250 mg/dl: 94.7%, in range 70-180 mg/dl: 82.6%, <70 mg/dl: 4.1%, and <50 mg/dl: 0.2%. During the last night the time spent in range was 70-250 mg/dl: 95%, in range 70-180 mg/dl: 87.7%, <70 mg/dl: 5.0%, and <50 mg/dl: 0.0%. No severe hypoglycemia occurred. No serious adverse events were reported.

CONCLUSIONS

The ARG algorithm was successfully validated in a pilot clinical trial, encouraging further tests with a larger number of patients and in outpatient settings.

Incorporating Unannounced Meals and Exercise in Adaptive Learning of Personalized Models for Multivariable Artificial Pancreas Systems

BACKGROUND

Despite the recent advancements in the modeling of glycemic dynamics for type 1 diabetes mellitus, automatically considering unannounced meals and exercise without manual user inputs remains challenging.

METHOD

An adaptive model identification technique that incorporates exercise information and estimates of the effects of unannounced meals obtained automatically without user input is proposed in this work. The effects of the unknown consumed carbohydrates are estimated using an individualized unscented Kalman filtering algorithm employing an augmented glucose-insulin dynamic model, and exercise information is acquired from noninvasive physiological measurements. The additional information on meals and exercise is incorporated with personalized estimates of plasma insulin concentration and glucose measurement data in an adaptive model identification algorithm.

RESULTS

The efficacy of the proposed personalized and adaptive modeling algorithm is demonstrated using clinical data involving closed-loop experiments of the artificial pancreas system, and the results demonstrate accurate glycemic modeling with the average root-mean-square error (mean absolute error) of 25.50 mg/dL (18.18 mg/dL) for six-step (30 minutes ahead) predictions.

CONCLUSIONS

The approach presented is able to identify reliable time-varying individualized glucose-insulin models.

Artificial Intelligence for Diabetes Management and Decision Support: Literature Review

Background

Artificial intelligence methods in combination with the latest technologies, including medical devices, mobile computing, and sensor technologies, have the potential to enable the creation and delivery of better management services to deal with chronic diseases. One of the most lethal and prevalent chronic diseases is diabetes mellitus, which is characterized by dysfunction of glucose homeostasis.

Objective

The objective of this paper is to review recent efforts to use artificial intelligence techniques to assist in the management of diabetes, along with the associated challenges.

Methods

A review of the literature was conducted using PubMed and related bibliographic resources. Analyses of the literature from 2010 to 2018 yielded 1849 pertinent articles, of which we selected 141 for detailed review.

Results

We propose a functional taxonomy for diabetes management and artificial intelligence. Additionally, a detailed analysis of each subject category was performed using related key outcomes. This approach revealed that the experiments and studies reviewed yielded encouraging results.

Conclusions

We obtained evidence of an acceleration of research activity aimed at developing artificial intelligence-powered tools for prediction and prevention of complications associated with diabetes. Our results indicate that artificial intelligence methods are being progressively established as suitable for use in clinical daily practice, as well as for the self-management of diabetes. Consequently, these methods provide powerful tools for improving patients’ quality of life.

Fully Closed-Loop Multiple Model Probabilistic Predictive Controller Artificial Pancreas Performance in Adolescents and Adults in a Supervised Hotel Setting

BACKGROUND

Initial Food and Drug Administration-approved artificial pancreas (AP) systems will be hybrid closed-loop systems that require prandial meal announcements and will not eliminate the burden of premeal insulin dosing. Multiple model probabilistic predictive control (MMPPC) is a fully closed-loop system that uses probabilistic estimation of meals to allow for automated meal detection. In this study, we describe the safety and performance of the MMPPC system with announced and unannounced meals in a supervised hotel setting.

RESEARCH DESIGN AND METHODS

The Android phone-based AP system with remote monitoring was tested for 72 h in six adults and four adolescents across three clinical sites with daily exercise and meal challenges involving both three announced (manual bolus by patient) and six unannounced (no bolus by patient) meals. Safety criteria were predefined. Controller aggressiveness was adapted daily based on prior hypoglycemic events.

RESULTS

Mean 24-h continuous glucose monitor (CGM) was 157.4 ± 14.4 mg/dL, with 63.6 ± 9.2% of readings between 70 and 180 mg/dL, 2.9 ± 2.3% of readings <70 mg/dL, and 9.0 ± 3.9% of readings >250 mg/dL. Moderate hyperglycemia was relatively common with 24.6 ± 6.2% of readings between 180 and 250 mg/dL, primarily within 3 h after a meal. Overnight mean CGM was 139.6 ± 27.6 mg/dL, with 77.9 ± 16.4% between 70 and 180 mg/dL, 3.0 ± 4.5% <70 mg/dL, 17.1 ± 14.9% between 180 and 250 mg/dL, and 2.0 ± 4.5%> 250 mg/dL. Postprandial hyperglycemia was more common for unannounced meals compared with announced meals (4-h postmeal CGM 197.8 ± 44.1 vs. 140.6 ± 35.0 mg/dL; P < 0.001). No participants met safety stopping criteria.

CONCLUSIONS

MMPPC was safe in a supervised setting despite meal and exercise challenges. Further studies are needed in a less supervised environment.

Automatic Detection and Estimation of Unannounced Meals for Multivariable Artificial Pancreas System

BACKGROUND

Automatically attenuating the postprandial rise in the blood glucose concentration without manual meal announcement is a significant challenge for artificial pancreas (AP) systems. In this study, a meal module is proposed to detect the consumption of a meal and to estimate the amount of carbohydrate (CHO) intake.

METHODS

The meals are detected based on qualitative variables describing variation of continuous glucose monitoring (CGM) readings. The CHO content of the meals/snacks is estimated by a fuzzy system using CGM and subcutaneous insulin delivery data. The meal bolus amount is computed according to the patient's insulin to CHO ratio. Integration of the meal module into a multivariable AP system allows revision of estimated CHO based on knowledge about physical activity, sleep, and the risk of hypoglycemia before the final decision for a meal bolus is made.

RESULTS

The algorithm is evaluated by using 117 meals/snacks in retrospective data from 11 subjects with type 1 diabetes. Sensitivity, defined as the percentage of correctly detected meals and snacks, is 93.5% for meals and 68.0% for snacks. The percentage of false positives, defined as the proportion of false detections relative to the total number of detected meals and snacks, is 20.8%.

CONCLUSIONS

Integration of a meal detection module in an AP system is a further step toward an automated AP without manual entries. Detection of a consumed meal/snack and infusion of insulin boluses using an estimate of CHO enables the AP system to automatically prevent postprandial hyperglycemia.

Stochastic Seasonal Models for Glucose Prediction in the Artificial Pancreas

BACKGROUND

Linear empirical dynamic models have been widely used for glucose prediction. The extension of the concept of seasonality, characteristic of other domains, is explored here for the improvement of prediction accuracy.

METHODS

Twenty time series of 8-hour postprandial periods (PP) for a same 60g-carbohydrate meal were collected from a closed-loop controller validation study. A single concatenated time series was produced representing a collection of data from similar scenarios, resulting in seasonality. Variability in the resulting time series was representative of worst-case intrasubject variability. Following a leave-one-out cross-validation, seasonal and nonseasonal autoregressive integrated moving average models (SARIMA and ARIMA) were built to analyze the effect of seasonality in the model prediction accuracy. Further improvement achieved from the inclusion of insulin infusion rate as exogenous variable was also analyzed. Prediction horizons (PHs) from 30 to 300 min were considered.

RESULTS

SARIMA outperformed ARIMA revealing a significant role of seasonality. For a 5-h PH, average MAPE was reduced in 26.62%. Considering individual runs, the improvement ranged from 6.3% to 54.52%. In the best-performing case this reduction amounted to 29.45%. The benefit of seasonality was consistent among different PHs, although lower PHs benefited more, with MAPE reduction over 50% for PHs of 60 and 120 minutes, and over 40% for 180 min. Consideration of insulin infusion rate into the seasonal model further improved performance, with a 61.89% reduction in MAPE for 30-min PH and reductions over 20% for PHs over 180 min.

CONCLUSIONS

Seasonality improved model accuracy allowing for the extension of the PH significantly.

Multivariable Adaptive Artificial Pancreas System in Type 1 Diabetes

PURPOSE OF REVIEW

The review summarizes the current state of the artificial pancreas (AP) systems and introduces various new modules that should be included in future AP systems.

RECENT FINDINGS

A fully automated AP must be able to detect and mitigate the effects of meals, exercise, stress and sleep on blood glucose concentrations. This can only be achieved by using a multivariable approach that leverages information from wearable devices that provide real-time streaming data about various physiological variables that indicate imminent changes in blood glucose concentrations caused by meals, exercise, stress and sleep. The development of a fully automated AP will necessitate the design of multivariable and adaptive systems that use information from wearable devices in addition to glucose sensors and modify the models used in their model-predictive alarm and control systems to adapt to the changes in the metabolic state of the user. These AP systems will also integrate modules for controller performance assessment, fault detection and diagnosis, machine learning and classification to interpret various signals and achieve fault-tolerant control. Advances in wearable devices, computational power, and safe and secure communications are enabling the development of fully automated multivariable AP systems.