Prediction of Adverse Glycemic Events From Continuous Glucose Monitoring Signal

Impartial feature selection using multi-agent reinforcement learning for adverse glycemic event prediction

We developed an attention model to predict future adverse glycemic events 30 min in advance based on the observation of past glycemic values over a 35 min period. The proposed model effectively encodes insulin administration and meal intake time using Time2Vec (T2V) for glucose prediction. The proposed impartial feature selection algorithm is designed to distribute rewards proportionally according to agent contributions. Agent contributions are calculated by a step-by-step negation of updated agents. Thus, the proposed feature selection algorithm optimizes features from electronic medical records to improve performance. For evaluation, we collected continuous glucose monitoring data from 102 patients with type 2 diabetes admitted to Cheonan Hospital, Soonchunhyang University. Using our proposed model, we achieved F1-scores of 89.0%, 60.6%, and 89.8% for normoglycemia, hypoglycemia, and hyperglycemia, respectively.

Detection of Hypoglycemia and Hyperglycemia Using Noninvasive Wearable Sensors: Electrocardiograms and Accelerometry

BACKGROUND

Monitoring glucose excursions is important in diabetes management. This can be achieved using continuous glucose monitors (CGMs). However, CGMs are expensive and invasive. Thus, alternative low-cost noninvasive wearable sensors capable of predicting glycemic excursions could be a game changer to manage diabetes.

METHODS

In this article, we explore two noninvasive sensor modalities, electrocardiograms (ECGs) and accelerometers, collected on five healthy participants over two weeks, to predict both hypoglycemic and hyperglycemic excursions. We extract 29 features encompassing heart rate variability features from the ECG, and time- and frequency-domain features from the accelerometer. We evaluated two machine learning approaches to predict glycemic excursions: a classification model and a regression model.

RESULTS

The best model for both hypoglycemia and hyperglycemia detection was the regression model based on ECG and accelerometer data, yielding 76% sensitivity and specificity for hypoglycemia and 79% sensitivity and specificity for hyperglycemia. This had an improvement of 5% in sensitivity and specificity for both hypoglycemia and hyperglycemia when compared with using ECG data alone.

CONCLUSIONS

Electrocardiogram is a promising alternative not only to detect hypoglycemia but also to predict hyperglycemia. Supplementing ECG data with contextual information from accelerometer data can improve glucose prediction.

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.

Combined Use of Glucose-Specific Model Identification and Alarm Strategy Based on Prediction-Funnel to Improve Online Forecasting of Hypoglycemic Events

BACKGROUND

Advanced decision support systems for type 1 diabetes (T1D) management often embed prediction modules, which allow T1D people to take preventive actions to avoid critical episodes like hypoglycemia. Real-time prediction of blood glucose (BG) concentration relies on a subject-specific model of glucose-insulin dynamics. Model parameter identification is usually based on the mean square error (MSE) cost function, and the model is usually used to predict BG at a single prediction horizon (PH). Finally, a hypo-alarm is raised if the predicted BG crosses a threshold. This work aims to show that real-time hypoglycemia forecasting can be improved by leveraging: a glucose-specific mean square error (gMSE) cost function in model's parameters identification, and a "prediction-funnel," that is, confidence intervals (CIs) for multiple PHs, within the hypo-alarm-raising strategy.

METHODS

Autoregressive integrated moving average with exogenous input (ARIMAX) models are selected to illustrate the proposed solution (use of gMSE and prediction-funnel) and its assessment against the conventional approach (MSE and single PH). The gMSE penalizes the model misfit in unsafe BG ranges (e.g., hypoglycemia), and the prediction-funnel allows raising an alarm by monitoring if the CIs cross a suitable threshold. The algorithms were evaluated by measuring precision (P), recall (R), F1-score (F1), false positive per day (FP/day), and time gain (TG) on a real dataset collected in 11 T1D individuals.

RESULTS

The best performance is achieved exploiting both the gMSE and the prediction-funnel: P = 65%, R = 88%, F1 = 75%, FP/day = 0.29, and mean TG = 15 minutes.

CONCLUSIONS

The combined use of a glucose-specific metric and an alarm-raising strategy based on the prediction-funnel allows achieving a more effective and reliable hypoglycemia prediction algorithm.

DiaTrend: A dataset from advanced diabetes technology to enable development of novel analytic solutions

Objective digital data is scarce yet needed in many domains to enable research that can transform the standard of healthcare. While data from consumer-grade wearables and smartphones is more accessible, there is critical need for similar data from clinical-grade devices used by patients with a diagnosed condition. The prevalence of wearable medical devices in the diabetes domain sets the stage for unique research and development within this field and beyond. However, the scarcity of open-source datasets presents a major barrier to progress. To facilitate broader research on diabetes-relevant problems and accelerate development of robust computational solutions, we provide the DiaTrend dataset. The DiaTrend dataset is composed of intensive longitudinal data from wearable medical devices, including a total of 27,561 days of continuous glucose monitor data and 8,220 days of insulin pump data from 54 patients with diabetes. This dataset is useful for developing novel analytic solutions that can reduce the disease burden for people living with diabetes and increase knowledge on chronic condition management in outpatient settings.

A dual-attention based coupling network for diabetes classification with heterogeneous data

Diabetes Mellitus (DM) is a group of metabolic disorders characterized by hyperglycaemia in the absence of treatment. Classification of DM is essential as it corresponds to the respective diagnosis and treatment. In this paper, we propose a new coupling network with hierarchical dual-attention that utilizes heterogeneous data, including Flash Glucose Monitoring (FGM) data and biomarkers in electronic medical records. The long short-term memory-based FGM sub-network extracts the time-dependent features of dynamic FGM sequences, while the biomarkers sub-network learns the features of static biomarkers. The convolutional block attention module (CBAM) for dispersing the feature weights of the spatial and channel dimensions is implemented into the FGM sub-network to endure the variability of FGM and allows us to extract high-level discriminative features more accurately. To better adjust the importance weights of the characteristics of the two sub-networks, self-attention is introduced to integrate the characteristics of heterogeneous data. Based on the dataset provided by Peking University People's Hospital, the proposed method is evaluated through factorial experiments of multi-source heterogeneous data, ablation studies of various attention strategies, time consumption evaluation and quantitative evaluation. The benchmark tests reveal the proposed network achieves a type 1 and 2 diabetes classification accuracy of 95.835% and the comprehensive performance metrics, including Matthews correlation coefficient, F1-score and G-mean, are 91.333%, 94.939% and 94.937% respectively. In the factorial experiments, the proposed method reaches the maximum area under the receiver operating characteristic curve of 0.9428, which indicates the effectiveness of the coupling between the nominated sub-networks. The coupling network with a dual-attention strategy performs better than the one without or only with a single-attention strategy in the ablation study as well. In addition, the model is also tested on another data set, and the accuracy of the test reaches 94.286%, reflecting that the model is robust when it is transferred to untrained diabetes data. The experimental results show that the proposed method is feasible in the classification of diabetes types. The code is available at https://github.com/bitDalei/Diabetes-Classification-with-Heterogeneous-Data.

Data-based modeling for hypoglycemia prediction: Importance, trends, and implications for clinical practice

Background and objective

Hypoglycemia is a key barrier to achieving optimal glycemic control in people with diabetes, which has been proven to cause a set of deleterious outcomes, such as impaired cognition, increased cardiovascular disease, and mortality. Hypoglycemia prediction has come to play a role in diabetes management as big data analysis and machine learning (ML) approaches have become increasingly prevalent in recent years. As a result, a review is needed to summarize the existing prediction algorithms and models to guide better clinical practice in hypoglycemia prevention.

Materials and methods

PubMed, EMBASE, and the Cochrane Library were searched for relevant studies published between 1 January 2015 and 8 December 2022. Five hypoglycemia prediction aspects were covered: real-time hypoglycemia, mild and severe hypoglycemia, nocturnal hypoglycemia, inpatient hypoglycemia, and other hypoglycemia (postprandial, exercise-related).

Results

From the 5,042 records retrieved, we included 79 studies in our analysis. Two major categories of prediction models are identified by an overview of the chosen studies: simple or logistic regression models based on clinical data and data-based ML models (continuous glucose monitoring data is most commonly used). Models utilizing clinical data have identified a variety of risk factors that can lead to hypoglycemic events. Data-driven models based on various techniques such as neural networks, autoregressive, ensemble learning, supervised learning, and mathematical formulas have also revealed suggestive features in cases of hypoglycemia prediction.

Conclusion

In this study, we looked deep into the currently established hypoglycemia prediction models and identified hypoglycemia risk factors from various perspectives, which may provide readers with a better understanding of future trends in this topic.

Personalized Blood Glucose Prediction for Type 1 Diabetes Using Evidential Deep Learning and Meta-Learning

The availability of large amounts of data from continuous glucose monitoring (CGM), together with the latest advances in deep learning techniques, have opened the door to a new paradigm of algorithm design for personalized blood glucose (BG) prediction in type 1 diabetes (T1D) with superior performance. However, there are several challenges that prevent the widespread implementation of deep learning algorithms in actual clinical settings, including unclear prediction confidence and limited training data for new T1D subjects. To this end, we propose a novel deep learning framework, Fast-adaptive and Confident Neural Network (FCNN), to meet these clinical challenges. In particular, an attention-based recurrent neural network is used to learn representations from CGM input and forward a weighted sum of hidden states to an evidential output layer, aiming to compute personalized BG predictions with theoretically supported model confidence. The model-agnostic meta-learning is employed to enable fast adaptation for a new T1D subject with limited training data. The proposed framework has been validated on three clinical datasets. In particular, for a dataset including 12 subjects with T1D, FCNN achieved a root mean square error of 18.64±2.60 mg/dL and 31.07±3.62 mg/dL for 30 and 60-minute prediction horizons, respectively, which outperformed all the considered baseline methods with significant improvements. These results indicate that FCNN is a viable and effective approach for predicting BG levels in T1D. The well-trained models can be implemented in smartphone apps to improve glycemic control by enabling proactive actions through real-time glucose alerts.

Assessment of Seasonal Stochastic Local Models for Glucose Prediction without Meal Size Information under Free-Living Conditions

Accurate blood glucose (BG) forecasting is key in diabetes management, as it allows preventive actions to mitigate harmful hypoglycemic/hyperglycemic episodes. Considering the encouraging results obtained by seasonal stochastic models in proof-of-concept studies, this work assesses the methodology in two datasets (open-loop and closed-loop) recorded in free-living conditions. First, similar postprandial glycemic profiles are grouped together with fuzzy C-means clustering. Then, a seasonal stochastic model is identified for each cluster. Finally, real-time BG forecasting is performed by weighting each model’s prediction. The proposed methodology (named C-SARIMA) is compared to other linear and nonlinear black-box methods: autoregressive integrated moving average (ARIMA), its variant with input (ARIMAX), a feed-forward neural network (NN), and its modified version (NN-X) fed by BG, insulin, and carbohydrates (timing and dosing) information for several prediction horizons (PHs). In the open-loop dataset, C-SARIMA grants a median root-mean-squared error (RMSE) of 20.13 mg/dL (PH = 30) and 27.23 mg/dL (PH = 45), not significantly different from ARIMA and NN. Over a longer PH, C-SARIMA achieves an RMSE = 31.96 mg/dL (PH = 60) and RMSE = 33.91 mg/dL (PH = 75), significantly outperforming the ARIMA and NN, without significant differences from the ARIMAX for PH ≥ 45 and the NN-X for PH ≥ 60. Similar results hold on the closed-loop dataset: for PH = 30 and 45 min, the C-SARIMA achieves an RMSE = 21.63 mg/dL and RMSE = 29.67 mg/dL, not significantly different from the ARIMA and NN. On longer PH, the C-SARIMA outperforms the ARIMA for PH > 45 and the NN for PH > 60 without significant differences from the ARIMAX for PH ≥ 45. Although using less input information, the C-SARIMA achieves similar performance to other prediction methods such as the ARIMAX and NN-X and outperforming the CGM-only approaches on PH > 45min.

A computational framework for discovering digital biomarkers of glycemic control

Digital biomarkers can radically transform the standard of care for chronic conditions that are complex to manage. In this work, we propose a scalable computational framework for discovering digital biomarkers of glycemic control. As a feasibility study, we leveraged over 79,000 days of digital data to define objective features, model the impact of each feature, classify glycemic control, and identify the most impactful digital biomarkers. Our research shows that glycemic control varies by age group, and was worse in the youngest population of subjects between the ages of 2–14. In addition, digital biomarkers like prior-day time above range and prior-day time in range, as well as total daily bolus and total daily basal were most predictive of impending glycemic control. With a combination of the top-ranked digital biomarkers, we achieved an average F1 score of 82.4% and 89.7% for classifying next-day glycemic control across two unique datasets.

Investigating Temporal Patterns of Glycemic Control around Holidays

Maintaining good glycemic control is a central part of diabetes care. However, it can be a tedious task because many factors in daily living can affect glycemic control. To support management, a growing number of people living with diabetes are now being prescribed continuous glucose monitors (CGMs) for real-time tracking of their blood glucose levels. However, routine use of CGMs is also an invaluable source of patient-generated data for individual and population-level studies. Prior research has shown that festive periods such as holidays can be a notable contributor to overeating and weight gain. Thus, in this work, we sought to investigate patterns of glycemic control around the holidays, particularly Thanksgiving, Christmas, and New Year, by using 3-months of CGM data from 14 patients with Type 1 Diabetes. We leveraged clinically validated metrics for quantifying glycemic control from CGM data and well-established statistical tests to compare diabetes management on holiday weeks versus non-holiday weeks. Based on our analysis, we found that 86% of subjects (12 out of 14) had worse glycemic control (i.e., more ad-verse glycemic events) during holiday weeks compared to non-holiday weeks. This general trend was prevalent amongst most subjects, however, we also observed unique individual patterns of glycemic control. Our findings provide a basis for further research on temporal patterns in diabetes management and data-driven interventions to support patients and caregivers with maintaining good glycemic control all year round.

Enhancing self-management in type 1 diabetes with wearables and deep learning

People living with type 1 diabetes (T1D) require lifelong self-management to maintain glucose levels in a safe range. Failure to do so can lead to adverse glycemic events with short and long-term complications. Continuous glucose monitoring (CGM) is widely used in T1D self-management for real-time glucose measurements, while smartphone apps are adopted as basic electronic diaries, data visualization tools, and simple decision support tools for insulin dosing. Applying a mixed effects logistic regression analysis to the outcomes of a six-week longitudinal study in 12 T1D adults using CGM and a clinically validated wearable sensor wristband (NCT ID: NCT03643692), we identified several significant associations between physiological measurements and hypo- and hyperglycemic events measured an hour later. We proceeded to develop a new smartphone-based platform, ARISES (Adaptive, Real-time, and Intelligent System to Enhance Self-care), with an embedded deep learning algorithm utilizing multi-modal data from CGM, daily entries of meal and bolus insulin, and the sensor wristband to predict glucose levels and hypo- and hyperglycemia. For a 60-minute prediction horizon, the proposed algorithm achieved the average root mean square error (RMSE) of 35.28 ± 5.77 mg/dL with the Matthews correlation coefficients for detecting hypoglycemia and hyperglycemia of 0.56 ± 0.07 and 0.70 ± 0.05, respectively. The use of wristband data significantly reduced the RMSE by 2.25 mg/dL (p < 0.01). The well-trained model is implemented on the ARISES app to provide real-time decision support. These results indicate that the ARISES has great potential to mitigate the risk of severe complications and enhance self-management for people with T1D.

Multi-task disentangled autoencoder for time-series data in glucose dynamics

The objective of this study is to propose MD-VAE: a multi-task disentangled variational autoencoders (VAE) for exploring characteristics of latent representations (LR) and exploiting LR for diverse tasks including glucose forecasting, event detection, and temporal clustering.

Sense and Learn: Recent Advances in Wearable Sensing and Machine Learning for Blood Glucose Monitoring and Trend-Detection

Diabetes mellitus is characterized by elevated blood glucose levels, however patients with diabetes may also develop hypoglycemia due to treatment. There is an increasing demand for non-invasive blood glucose monitoring and trends detection amongst people with diabetes and healthy individuals, especially athletes. Wearable devices and non-invasive sensors for blood glucose monitoring have witnessed considerable advances. This review is an update on recent contributions utilizing novel sensing technologies over the past five years which include electrocardiogram, electromagnetic, bioimpedance, photoplethysmography, and acceleration measures as well as bodily fluid glucose sensors to monitor glucose and trend detection. We also review methods that use machine learning algorithms to predict blood glucose trends, especially for high risk events such as hypoglycemia. Convolutional and recurrent neural networks, support vector machines, and decision trees are examples of such machine learning algorithms. Finally, we address the key limitations and challenges of these studies and provide recommendations for future work.

Digital Solutions to Diagnose and Manage Postbariatric Hypoglycemia

Postbariatric hypoglycemia (PBH) is an increasingly recognized late metabolic complication of bariatric surgery, characterized by low blood glucose levels 1-3 h after a meal, particularly if the meal contains rapid-acting carbohydrates. PBH can often be effectively managed through appropriate nutritional measures, which remain the cornerstone treatment today. However, their implementation in daily life continues to challenge both patients and health care providers. Emerging digital technologies may allow for more informed and improved decision-making through better access to relevant data to manage glucose levels in PBH. Examples include applications for automated food analysis from meal images, digital receipts of purchased food items or integrated platforms allowing the connection of continuously measured glucose with food and other health-related data. The resulting multi-dimensional data can be processed with artificial intelligence systems to develop prediction algorithms and decision support systems with the aim of improving glucose control, safety, and quality of life of PBH patients. Digital innovations, however, face trade-offs between user burden vs. amount and quality of data. Further challenges to their development are regulatory non-compliance regarding data ownership of the platforms acquiring the required data, as well as user privacy concerns and compliance with regulatory requirements. Through navigating these trade-offs, digital solutions could significantly contribute to improving the management of PBH.

Communication Requirements in 5G-Enabled Healthcare Applications: Review and Considerations

Fifth generation (5G) mobile communication technology can enable novel healthcare applications and augment existing ones. However, 5G-enabled healthcare applications demand diverse technical requirements for radio communication. Knowledge of these requirements is important for developers, network providers, and regulatory authorities in the healthcare sector to facilitate safe and effective healthcare. In this paper, we review, identify, describe, and compare the requirements for communication key performance indicators in relevant healthcare use cases, including remote robotic-assisted surgery, connected ambulance, wearable and implantable devices, and service robotics for assisted living, with a focus on quantitative requirements. We also compare 5G-healthcare requirements with the current state of 5G capabilities. Finally, we identify gaps in the existing literature and highlight considerations for this space.

A predictive model incorporating the change detection and Winsorization methods for alerting hypoglycemia and hyperglycemia

This paper focuses on establishing an effective predictive model to quickly and accurately alert hypoglycemia and hyperglycemia for helping control blood glucose levels of people with diabetes. In general, a good predictive model is established on the features of data. Inspired by this, we first analyze the characteristics of continuous glucose monitoring (CGM) data by the equality of variances test and outlier detection, which show time-varying fluctuations and jump points in CGM data. Therefore, we incorporate the change detection method and the Winsorization method into the predictive model based on the autoregressive moving average (ARMA) model and the recursive least squares (RLS) method to fit the above characteristics. To the best of our knowledge, the proposed method is the first attempt to give a solution for matching the time-varying fluctuations and jump points of CGM data simultaneously. A case study using CGM data is given to validate the effectiveness of the proposed method under 30-min-ahead prediction. The results show that the proposed method can improve the true alarm ratio of hypoglycemia and hyperglycemia from 0.7983 to 0.8783, and lengthen the average advance detection time of hypoglycemia and hyperglycemia from 19.77 to 22.64 min.

Hypoglycaemia detection and prediction techniques: A systematic review on the latest developments

The main objective of diabetes control is to correct hyperglycaemia while avoiding hypoglycaemia, especially in insulin-treated patients. Fear of hypoglycaemia is a hurdle to effective correction of hyperglycaemia because it promotes under-dosing of insulin. Strategies to minimise hypoglycaemia include education and training for improved hypoglycaemia awareness and the development of technologies to allow their early detection and thus minimise their occurrence. Patients with impaired hypoglycaemia awareness would benefit the most from these technologies. The purpose of this systematic review is to review currently available or in-development technologies that support detection of hypoglycaemia or hypoglycaemia risk, and identify gaps in the research. Nanomaterial use in sensors is a promising strategy to increase the accuracy of continuous glucose monitoring devices for low glucose values. Hypoglycaemia is associated with changes on vital signs, so electrocardiogram and encephalogram could also be used to detect hypoglycaemia. Accuracy improvements through multivariable measures can make already marketed galvanic skin response devices a good noninvasive alternative. Breath volatile organic compounds can be detected by dogs and devices and alert patients at hypoglycaemia onset, while near-infrared spectroscopy can also be used as a hypoglycaemia alarms. Finally, one of the main directions of research are deep learning algorithms to analyse continuous glucose monitoring data and provide earlier and more accurate prediction of hypoglycaemia. Current developments for early identification of hypoglycaemia risk combine improvements of available 'needle-type' enzymatic glucose sensors and noninvasive alternatives. Patient usability will be essential to demonstrate to allow their implementation for daily use in diabetes management.

Data-based algorithms and models using diabetics real data for blood glucose and hypoglycaemia prediction - A systematic literature review

BACKGROUND AND AIM

Hypoglycaemia prediction play an important role in diabetes management being able to reduce the number of dangerous situations. Thus, it is relevant to present a systematic review on the currently available prediction algorithms and models for hypoglycaemia (or hypoglycemia in US English) prediction.

METHODS

This study aims to systematically review the literature on data-based algorithms and models using diabetics real data for hypoglycaemia prediction. Five electronic databases were screened for studies published from January 2014 to June 2020: ScienceDirect, IEEE Xplore, ACM Digital Library, SCOPUS, and PubMed.

RESULTS

Sixty-three eligible studies were retrieved that met the inclusion criteria. The review identifies the current trend in this topic: most of the studies perform short-term predictions (82.5%). Also, the review pinpoints the inputs and shows that information fusion is relevant for hypoglycaemia prediction. Regarding data-based models (80.9%) and hybrid models (19.1%) different predictive techniques are used: Artificial neural network (22.2%), ensemble learning (27.0%), supervised learning (20.6%), statistic/probabilistic (7.9%), autoregressive (7.9%), evolutionary (6.4%), deep learning (4.8%) and adaptative filter (3.2%). Artificial Neural networks and hybrid models show better results.

CONCLUSIONS

The data-based models for blood glucose and hypoglycaemia prediction should be able to provide a good balance between the applicability and performance, integrating complementary data from different sources or from different models. This review identifies trends and possible opportunities for research in this topic.

The challenge of predicting blood glucose concentration changes in patients with type I diabetes

Patients with Type I Diabetes (T1D) must take insulin injections to prevent the serious long term effects of hyperglycemia. They must also be careful not to inject too much insulin because this could induce (potentially fatal) hypoglycemia. Patients therefore follow a "regimen" that determines how much insulin to inject at each time, based on various measurements. We can produce an effective regimen if we can accurately predict a patient's future blood glucose (BG) values from his/her current features. This study explores the challenges of predicting future BG by applying a number of machine learning algorithms, as well as various data preprocessing variations (corresponding to 312 [learner, preprocessed-dataset] combinations), to a new T1D dataset that contains 29,601 entries from 47 different patients. Our most accurate predictor, a weighted ensemble of two Gaussian Process Regression models, achieved a (cross-validation) errL1 loss of 2.7 mmol/L (48.65 mg/dl). This result was unexpectedly poor given that one can obtain an errL1 of 2.9 mmol/L (52.43 mg/dl) using the naive approach of simply predicting the patient's average BG. These results suggest that the diabetes diary data that is typically collected may be insufficient to produce accurate BG prediction models; additional data may be necessary to build accurate BG prediction models over hours.

Linear Model Identification for Personalized Prediction and Control in Diabetes

OBJECTIVE

Type-1 diabetes (T1D) is a metabolic disease, characterized by impaired blood glucose (BG) regulation, which forces patients to multiple daily therapeutic actions, the most critical of which is exogenous insulin administration. T1D management can considerably benefit of mathematical models enabling accurate BG predictions and effective/safe automated insulin delivery. In building these models, dealing with large inter- and intra-patient variability in glucose-insulin dynamics represents a major challenge. The aim of the present work is to assess linear black-box methods, including a novel non-parametric methodology, for learning individualized models of glucose response to insulin and meal, suitable for model-based prediction and control.

METHODS

We focus on data-driven techniques for linear model-learning and compare the state-of-art parametric pipeline, exploring all its degrees of freedom (including population vs. individualized parameter identification, model class chosen among ARX/ARMAX/ARIMAX/Box-Jenkins, model order selection criteria, etc.), with a novel non-parametric approach based on Gaussian regression and stable spline kernel. By using data collected in 11 T1D individuals, we evaluate effectiveness of the different models by measuring root mean squared error (RMSE), coefficient of determination (COD), and time gain of the associated BG predictors.

RESULTS

Among the tested approaches, the non-parametric technique results in the best prediction performance: median RMSE=29.8mg/dL, and median COD=57.4%, for a prediction horizon (PH) of 60 min. With respect to the state-of-the-art parametric techniques, the non-parametric approach grants a COD improvement of about 2%, 7%, 21%, and 41% for PH = 30, 60, 90, and 120 min (paired-sample t-test p 0.001, p=0.003, p=0.03, and p=0.07 respectively).

CONCLUSION

Non-parametric linear model-learning grants statistically significant improvement with respect to the state-of-art parametric approach. The higher PH, the more pronounced the improvement.

SIGNIFICANCE

The use of a linear non-parametric model-learning approach for model-based prediction and control could bring to a more prompt, safe and effective T1D management.

Profiles of Intraday Glucose in Type 2 Diabetes and Their Association with Complications: An Analysis of Continuous Glucose Monitoring Data

Aims: To identify profiles of type 2 diabetes from continuous glucose monitoring (CGM) data using ambulatory glucose profile (AGP) indicators and examine the association with prevalent complications. Methods: Two weeks of CGM data, collected between 2015 and 2019, from 5901 adult type 2 diabetes patients were retrieved from a clinical database in Chennai, India. Non-negative matrix factorization was used to identify profiles as per AGP indicators. The association of profiles with existing complications was examined using multinomial and logistic regressions adjusted for glycated hemoglobin (HbA1c; %), sex, age at onset, and duration of diabetes. Results: Three profiles of glycemic variability (GV) were identified based on CGM data-Profile 1 ["TIR Profile"] (n = 2271), Profile 2 ["Hypo"] (n = 1471), and Profile 3 ["Hyper"] (n = 2159). Compared with time in range (TIR) profile, those belonging to Hyper had higher mean fasting plasma glucose (202.9 vs. 167.1, mg/dL), 2-h postprandial plasma glucose (302.1 vs. 255.6, mg/dL), and HbA1c (9.7 vs. 8.6; %). Both "Hypo profile" and "Hyper profile" had higher odds of nonproliferative diabetic retinopathy ("Hypo": 1.44, 1.20-1.73; "Hyper": 1.33, 1.11-1.58), macroalbuminuria ("Hypo": 1.58, 1.25-1.98; "Hyper": 1.37, 1.10-1.71), and diabetic kidney disease (DKD; "Hypo": 1.65, 1.18-2.31; "Hyper": 1.88, 1.37-2.58), compared with "TIR profile." Those in "Hypo profile" (vs. "TIR profile") had higher odds of proliferative diabetic retinopathy (PDR; 2.84, 1.65-2.88). Conclusions: We have identified three profiles of GV from CGM data. While both "Hypo profile" and "Hyper profile" had higher odds of prevalent DKD compared with "TIR profile," "Hypo profile" had higher odds of PDR. Our study emphasizes the clinical importance of recognizing and treating hypoglycemia (which is often unrecognized without CGM) in patients with type 2 Diabetes Mellitus.

A Multitask Learning Approach to Personalised Blood Glucose Prediction

Blood glucose prediction algorithms are key tools in the development of decision support systems and closed-loop insulin delivery systems for blood glucose control in diabetes. Deep learning models have provided leading results among machine learning algorithms to date in glucose prediction. However these models typically require large amounts of data to obtain best personalised glucose prediction results. Multitask learning facilitates an approach for leveraging data from multiple subjects while still learning accurate personalised models. In this work we present results comparing the effectiveness of multitask learning over sequential transfer learning, and learning only on subject-specific data with neural networks and support vector regression. The multitask learning approach shows consistent leading performance in predictive metrics at both short-term and long-term prediction horizons. We obtain a predictive accuracy (RMSE) of 18.8 2.3, 25.3 2.9, 31.8 3.9, 41.2 4.5, 47.2 4.6 mg/dL at 30, 45, 60, 90, and 120 min prediction horizons respectively, with at least 93\% clinically acceptable predictions using the Clarke Error Grid (EGA) at each prediction horizon. We also identify relevant prior information such as glycaemic variability that can be incorporated to improve predictive performance at long-term prediction horizons. Furthermore, we demonstrate consistent performance - 5% change in both RMSE and EGA (Zone A) - in rare cases of adverse glycaemic events with 1-6 weeks of training data. In conclusion, a multitask approach can allow for deploying personalised models even with significantly less subject-specific data without compromising performance.

Feature-Based Machine Learning Model for Real-Time Hypoglycemia Prediction

BACKGROUND

Hypoglycemia is a serious health concern in youth with type 1 diabetes (T1D). Real-time data from continuous glucose monitoring (CGM) can be used to predict hypoglycemic risk, allowing patients to take timely intervention measures.

METHODS

A machine learning model is developed for probabilistic prediction of hypoglycemia (<70 mg/dL) in 30- and 60-minute time horizons based on CGM datasets obtained from 112 patients over a range of 90 days consisting of over 1.6 million CGM values under normal living conditions. A comprehensive set of features relevant for hypoglycemia are developed and a parsimonious subset with most influence on predicting hypoglycemic risk is identified. Model performance is evaluated both with and without contextual information on insulin and carbohydrate intake.

RESULTS

The model predicted hypoglycemia with >91% sensitivity for 30- and 60-minute prediction horizons while maintaining specificity >90%. Inclusion of insulin and carbohydrate data yielded performance improvement for 60-minute but not for 30-minute predictions. Model performance was highest for nocturnal hypoglycemia (~95% sensitivity). Shortterm (less than one hour) and medium-term (one to four hours) features for good prediction performance are identified.

CONCLUSIONS

Innovative feature identification facilitated high performance for hypoglycemia risk prediction in pediatric youth with T1D. Timely alerts of impending hypoglycemia may enable proactive measures to avoid severe hypoglycemia and achieve optimal glycemic control. The model will be deployed on a patient-facing smartphone application in an upcoming pilot study.

Blood Glucose Level Forecasting on Type-1-Diabetes Subjects during Physical Activity: A Comparative Analysis of Different Learning Techniques

Background: Type 1 Diabetes Mellitus (T1DM) is a widespread chronic disease in industrialized countries. Preventing blood glucose levels from exceeding the euglycaemic range would reduce the incidence of diabetes-related complications and improve the quality of life of subjects with T1DM. As a consequence, in the last decade, many Machine Learning algorithms aiming to forecast future blood glucose levels have been proposed. Despite the excellent performance they obtained, the prediction of abrupt changes in blood glucose values produced during physical activity (PA) is still one of the main challenges. Methods: A Jump Neural Network was developed in order to overcome the issue of predicting blood glucose values during PA. Three learning configurations were developed and tested: offline training, online training, and online training with reinforcement. All configurations were tested on six subjects suffering from T1DM that held regular PA (three aerobic and three anaerobic) and exploited Continuous Glucose Monitoring (CGM). Results: The forecasting performance was evaluated in terms of the Root-Mean-Squared-Error (RMSE), according to a paradigm of Precision Medicine. Conclusions: The online learning configurations performed better than the offline configuration in total days but not on the only CGM associated with the PA; thus, the results do not justify the increased computational burden because the improvement was not significant.

Improved Low-Glucose Predictive Alerts Based on Sustained Hypoglycemia: Model Development and Validation Study

BACKGROUND

Predictive alerts for impending hypoglycemic events enable persons with type 1 diabetes to take preventive actions and avoid serious consequences.

OBJECTIVE

This study aimed to develop a prediction model for hypoglycemic events with a low false alert rate, high sensitivity and specificity, and good generalizability to new patients and time periods.

METHODS

Performance improvement by focusing on sustained hypoglycemic events, defined as glucose values less than 70 mg/dL for at least 15 minutes, was explored. Two different modeling approaches were considered: (1) a classification-based method to directly predict sustained hypoglycemic events, and (2) a regression-based prediction of glucose at multiple time points in the prediction horizon and subsequent inference of sustained hypoglycemia. To address the generalizability and robustness of the model, two different validation mechanisms were considered: (1) patient-based validation (model performance was evaluated on new patients), and (2) time-based validation (model performance was evaluated on new time periods).

RESULTS

This study utilized data from 110 patients over 30-90 days comprising 1.6 million continuous glucose monitoring values under normal living conditions. The model accurately predicted sustained events with >97% sensitivity and specificity for both 30- and 60-minute prediction horizons. The false alert rate was kept to <25%. The results were consistent across patient- and time-based validation strategies.

CONCLUSIONS

Providing alerts focused on sustained events instead of all hypoglycemic events reduces the false alert rate and improves sensitivity and specificity. It also results in models that have better generalizability to new patients and time periods.

A sensor based intelligent system for classification and assistance of diabetes patients in telemedicine technology

Diabetes is one of the chronic metabolic disorder. Under diabetic condition, blood glucose level should be properly maintained in order to avoid various major diseases. The condition will be worse when it is not controlled at an earlier stage. Even massive heart attack cannot be identified when the patient has been affected by diabetes. Early diagnosis is required for preventing fatal diseases like cardiac problem, asthma, heart attack etc. In the proposed system measurement of glucose level and Prediction/ diagnosis of diabetes is based on the real time low complexity neural network implemented on a wearable device. A larger network is required for the diagnosis which needs to be present far-off in cloud and initiated for diagnosis and classification process of diabetes whenever it is essential. People can be able to manage and monitor the required basic parameters like heart rate, glucose level, lung condition, pressure of blood using the corresponding light weight biosensors in the wearable device designed through telemedicine technology. The quality of the disease diagnosis and Prediction is improved in this way. Using neural network feed forward prediction model in conjugation with back propagation algorithm and given training data, the system predicts whether the patient is prone to diabetes or not. The proposed work was evaluated using physic sensor data from physio net data base and also tested for real time functioning. The Proposed system found to be efficient in accuracy, sensitivity and fast operative.

Forecasting of Glucose Levels and Hypoglycemic Events: Head-to-Head Comparison of Linear and Nonlinear Data-Driven Algorithms Based on Continuous Glucose Monitoring Data Only

In type 1 diabetes management, the availability of algorithms capable of accurately forecasting future blood glucose (BG) concentrations and hypoglycemic episodes could enable proactive therapeutic actions, e.g., the consumption of carbohydrates to mitigate, or even avoid, an impending critical event. The only input of this kind of algorithm is often continuous glucose monitoring (CGM) sensor data, because other signals (such as injected insulin, ingested carbs, and physical activity) are frequently unavailable. Several predictive algorithms fed by CGM data only have been proposed in the literature, but they were assessed using datasets originated by different experimental protocols, making a comparison of their relative merits difficult. The aim of the present work was to perform a head-to-head comparison of thirty different linear and nonlinear predictive algorithms using the same dataset, given by 124 CGM traces collected over 10 days with the newest Dexcom G6 sensor available on the market and considering a 30-min prediction horizon. We considered the state-of-the art methods, investigating, in particular, linear black-box methods (autoregressive; autoregressive moving-average; and autoregressive integrated moving-average, ARIMA) and nonlinear machine-learning methods (support vector regression, SVR; regression random forest; feed-forward neural network, fNN; and long short-term memory neural network). For each method, the prediction accuracy and hypoglycemia detection capabilities were assessed using either population or individualized model parameters. As far as prediction accuracy is concerned, the results show that the best linear algorithm (individualized ARIMA) provides accuracy comparable to that of the best nonlinear algorithm (individualized fNN), with root mean square errors of 22.15 and 21.52 mg/dL, respectively. As far as hypoglycemia detection is concerned, the best linear algorithm (individualized ARIMA) provided precision = 64%, recall = 82%, and one false alarm/day, comparable to the best nonlinear technique (population SVR): precision = 63%, recall = 69%, and 0.5 false alarms/day. In general, the head-to-head comparison of the thirty algorithms fed by CGM data only made using a wide dataset shows that individualized linear models are more effective than population ones, while no significant advantages seem to emerge when employing nonlinear methodologies.

Service Level Agreements for 5G and Beyond: Overview, Challenges and Enablers of 5G-Healthcare Systems

5G and beyond networks will transform the healthcare sector by opening possibilities for novel use cases and applications. Service level agreements (SLAs) can enable 5G-enabled medical device use cases by documenting how a medical device communication requirements are met by the unique characteristics of 5G networks and the roles and responsibilities of the stakeholders involved in offering safe and effective 5G-enabled healthcare to patients. However, there are gaps in this space that should be addressed to facilitate the efficient implementation of 5G technology in healthcare. Current literature is scarce regarding SLAs for 5G and is absent regarding SLAs for 5G-enabled medical devices. This paper aims to bridge these gaps by identifying key challenges, providing insight, and describing open research questions related to SLAs in 5G and specifically 5G-healthcare systems. This is helpful to network service providers, users, and regulatory authorities in developing, managing, monitoring, and evaluating SLAs in 5G-enabled medical systems.

Method to generate a large cohort in-silico for type 1 diabetes

BACKGROUND AND OBJECTIVE

In the last decade, several technological solutions have been proposed as artificial pancreas systems able to treat type 1 diabetes; most often they are built based on a control algorithm that needs to be validated before it is used with real patients. Control algorithms are usually tested with simulation tools that integrate mathematical models related mainly to the glucose-insulin dynamics, but other variables can be considered as well. In general, the simulators have a limited set of subjects. The main goal of this paper is to propose a new computational method to increase the number of virtual subjects, with physiological characteristics, included in the original mathematical models.

METHODS

A subject is defined by a set of parameters given by a mathematical model. From the available reduced number of subjects in the model, the covariance of each parameter of every subject is obtained to establish a mathematical relationship. Then, new sets of parameters are calculated using linear regression methods; this generates larger cohorts, which allows for testing insulin therapies in open-loop or closed-loop scenarios. The new method proposed here increases the number of subjects in a virtual cohort using two versions of Hovorka's mathematical model.

RESULTS

Two covariant cohorts are obtained with linear regression. Both cohorts are clustered to avoid overlapping in the glucose-insulin dynamics and are compared in terms of their qualitative and quantitative behaviours in the normoglycemic range. As a result, there have been generated two larger cohorts (256 subjects) than the original population, which contributes to improving the variability in in-silico tests. In addition, for analysing the characteristics of the covariant generation method, two random cohorts have been generated, where the parameters are obtained individually and independently from each other, exhibiting only distribution limitations so that these cohorts do not have physiological subjects.

CONCLUSIONS

The proposed methodology has enabled the generation of a large cohort of 256 subjects, with different characteristics that are plausible in the T1DM population, significantly increasing the number of available subjects in existing mathematical models. The proposed methodology does not limit the number of subjects that can be generated and thus, it can be used to increase the number of cohorts provided by other mathematical models in diabetes, or even other scientific problems.

A Glucose-Insulin Mixture Model and Application to Short-Term Hypoglycemia Prediction in the Night Time

OBJECTIVE

Insulin-induced hypoglycemia is recognized as a critical problem for diabetic patients, especially at night. To give glucose prediction and advance warning of hypoglycemia of at least 30 minutes, various glucose-insulin models have been proposed. Recognizing the complementary nature of the models, this research proposes a Glucose-Insulin Mixture (GIM) model to predict the glucose values for hypoglycemia detection, by optimally fusing different models with its adjusted parameters to address the inter- and intra-individual variability.

METHODS

Two types of classic glucose-insulin models, the Ruan model, with single-compartment glucose kinetics, and the Hovorka model, with two-compartment glucose kinetics, are selected as two candidate models. Based on Bayesian inference, GIM is introduced with quantified contributions from the models with the associated parameters. GIM is then applied to predict the glucose values and hypoglycemia events.

RESULTS

The proposed model is validated by the nocturnal glucose data collected from 12 participants with type 1 diabetes. The GIM model has promising fitting of RMSE within 0.3465 mmol/L and predicting of RMSE within 0.5571 mmol/L. According to the literature, the hypoglycemia is defined as 3.9 mmol/L, and the GIM model shows good short-term hypoglycemia prediction performance with the data collected within the last hour (accuracy: 95.97%, precision: 91.77%, recall: 95.60%). In addition, the probability of hypoglycemia event in 30 minutes is inferred.

CONCLUSION

GIM, by fusing various glucose-insulin models via Bayesian inference, has the promise to capture glucose dynamics and predict hypoglycemia.

SIGNIFICANCE

GIM based short-term hypoglycemia prediction has potential clinical utility for timely intervention.

Data-Driven Insights on Behavioral Factors that Affect Diabetes Management

The prevalence of personal health data from wearable devices enables new opportunities to understand the impact of behavioral factors on health. Unlike consumer devices that are often auxiliary, such as Fitbit and Garmin, wearable medical devices like continuous glucose monitoring (CGM) devices and insulin pumps are becoming critical in diabetes care to minimize the occurrence of adverse glycemic events. Joint analysis of CGM and insulin pump data can provide unparalleled insights on how to modify treatment regimen to improve diabetes management outcomes. In this paper, we employ a data-driven approach to study the relationship between key behavioral factors and proximal diabetic management indicators. Our dataset includes an average of 161 days of time-matched CGM and insulin pump data from 34 subjects with Type 1 Diabetes (T1D). By employing hypothesis testing and association mining, we observe that smaller meals and insulin doses are associated with better glycemic outcomes compared to larger meals and insulin doses. Meanwhile, the occurrence of interrupted sleep is associated with poorer glycemic outcomes. This paper introduces a method for inferring disrupted sleep from wearable diabetes-device data and provides a baseline for future research on sleep quality and diabetes. This work also provides insights for development of decision-support tools for improving short- and long-term outcomes in diabetes care.

A machine-learning approach to predict postprandial hypoglycemia

BACKGROUND

For an effective artificial pancreas (AP) system and an improved therapeutic intervention with continuous glucose monitoring (CGM), predicting the occurrence of hypoglycemia accurately is very important. While there have been many studies reporting successful algorithms for predicting nocturnal hypoglycemia, predicting postprandial hypoglycemia still remains a challenge due to extreme glucose fluctuations that occur around mealtimes. The goal of this study is to evaluate the feasibility of easy-to-use, computationally efficient machine-learning algorithm to predict postprandial hypoglycemia with a unique feature set.

METHODS

We use retrospective CGM datasets of 104 people who had experienced at least one hypoglycemia alert value during a three-day CGM session. The algorithms were developed based on four machine learning models with a unique data-driven feature set: a random forest (RF), a support vector machine using a linear function or a radial basis function, a K-nearest neighbor, and a logistic regression. With 5-fold cross-subject validation, the average performance of each model was calculated to compare and contrast their individual performance. The area under a receiver operating characteristic curve (AUC) and the F1 score were used as the main criterion for evaluating the performance.

RESULTS

In predicting a hypoglycemia alert value with a 30-min prediction horizon, the RF model showed the best performance with the average AUC of 0.966, the average sensitivity of 89.6%, the average specificity of 91.3%, and the average F1 score of 0.543. In addition, the RF showed the better predictive performance for postprandial hypoglycemic events than other models.

CONCLUSION

In conclusion, we showed that machine-learning algorithms have potential in predicting postprandial hypoglycemia, and the RF model could be a better candidate for the further development of postprandial hypoglycemia prediction algorithm to advance the CGM technology and the AP technology further.

Classification of Postprandial Glycemic Status with Application to Insulin Dosing in Type 1 Diabetes-An In Silico Proof-of-Concept

In the daily management of type 1 diabetes (T1D), determining the correct insulin dose to be injected at meal-time is fundamental to achieve optimal glycemic control. Wearable sensors, such as continuous glucose monitoring (CGM) devices, are instrumental to achieve this purpose. In this paper, we show how CGM data, together with commonly recorded inputs (carbohydrate intake and bolus insulin), can be used to develop an algorithm that allows classifying, at meal-time, the post-prandial glycemic status (i.e., blood glucose concentration being too low, too high, or within target range). Such an outcome can then be used to improve the efficacy of insulin therapy by reducing or increasing the corresponding meal bolus dose. A state-of-the-art T1D simulation environment, including intraday variability and a behavioral model, was used to generate a rich in silico dataset corresponding to 100 subjects over a two-month scenario. Then, an extreme gradient-boosted tree (XGB) algorithm was employed to classify the post-prandial glycemic status. Finally, we demonstrate how the XGB algorithm outcome can be exploited to improve glycemic control in T1D through real-time adjustment of the meal insulin bolus. The proposed XGB algorithm obtained good accuracy at classifying post-prandial glycemic status (AUROC = 0.84 [0.78, 0.87]). Consequently, when used to adjust, in real-time, meal insulin boluses obtained with a bolus calculator, the proposed approach improves glycemic control when compared to the baseline bolus calculator. In particular, percentage time in target [70, 180] mg/dL was improved from 61.98 (±13.89) to 67.00 (±11.54; p < 0.01) without increasing hypoglycemia.