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

GARNN: An interpretable graph attentive recurrent neural network for predicting blood glucose levels via multivariate time series

Accurate prediction of future blood glucose (BG) levels can effectively improve BG management for people living with type 1 or 2 diabetes, thereby reducing complications and improving quality of life. The state of the art of BG prediction has been achieved by leveraging advanced deep learning methods to model multimodal data, i.e., sensor data and self-reported event data, organized as multi-variate time series (MTS). However, these methods are mostly regarded as "black boxes" and not entirely trusted by clinicians and patients. In this paper, we propose interpretable graph attentive recurrent neural networks (GARNNs) to model MTS, explaining variable contributions via summarizing variable importance and generating feature maps by graph attention mechanisms instead of post-hoc analysis. We evaluate GARNNs on four datasets, representing diverse clinical scenarios. Upon comparison with fifteen well-established baseline methods, GARNNs not only achieve the best prediction accuracy but also provide high-quality temporal interpretability, in particular for postprandial glucose levels as a result of corresponding meal intake and insulin injection. These findings underline the potential of GARNN as a robust tool for improving diabetes care, bridging the gap between deep learning technology and real-world healthcare solutions.

A data-driven personalized approach to predict blood glucose levels in type-1 diabetes patients exercising in free-living conditions

OBJECTIVE

The development of new technologies has generated vast amount of data that can be analyzed to better understand and predict the glycemic behavior of people living with type 1 diabetes. This paper aims to assess whether a data-driven approach can accurately and safely predict blood glucose levels in patients with type 1 diabetes exercising in free-living conditions.

METHODS

Multiple machine learning (XGBoost, Random Forest) and deep learning (LSTM, CNN-LSTM, Dual-encoder with Attention layer) regression models were considered. Each deep-learning model was implemented twice: first, as a personalized model trained solely on the target patient's data, and second, as a fine-tuned model of a population-based training model. The datasets used for training and testing the models were derived from the Type 1 Diabetes Exercise Initiative (T1DEXI). A total of 79 patients in T1DEXI met our inclusion criteria. Our models used various features related to continuous glucose monitoring, insulin pumps, carbohydrate intake, exercise (intensity and duration), and physical activity-related information (steps and heart rate). This data was available for four weeks for each of the 79 included patients. Three prediction horizons (10, 20, and 30 min) were tested and analyzed.

RESULTS

For each patient, there always exists either a machine learning or a deep learning model that conveniently predicts BGLs for up to 30 min. The best performing model differs from one patient to another. When considering the best performing model for each patient, the median and the mean Root Mean Squared Error (RMSE) values (across the 79 patients) for predictions made 10 min ahead were 6.99 mg/dL and 7.46 mg/dL, respectively. For predictions made 30 min ahead, the median and mean RMSE values were 16.85 mg/dL and 17.74 mg/dL, respectively. The majority of the predictions output by the best model of each patient fell within the clinically safe zones A and B of the Clarke Error Grid (CEG), with almost no predictions falling into the unsafe zone E. The most challenging patient to predict 30 min ahead achieved an RMSE value of 32.31 mg/dL (with the corresponding best performing model). The best-predicted patient had an RMSE value of 10.48 mg/dL. Predicting blood glucose levels was more difficult during and after exercise, resulting in higher RMSE values on average. Prediction errors during and after physical activity (two hours and four hours after) generally remained within the clinical safe zones of the CEG with less than 0.5% of predictions falling into the harmful zones D and E, regardless of the exercise category.

CONCLUSIONS

Data-driven approaches can accurately predict blood glucose levels in type 1 diabetes patients exercising in free-living conditions. The best-performing model varies across patients. Approaches in which a population-based model is initially trained and then fine-tuned for each individual patient generally achieve the best performance for the majority of patients. Some patients remain challenging to predict with no straightforward explanation of why a patient is more challenging to predict than another.

A deep learning approach for blood glucose monitoring and hypoglycemia prediction in glycogen storage disease

Glycogen storage disease (GSD) is a group of rare inherited metabolic disorders characterized by abnormal glycogen storage and breakdown. These disorders are caused by mutations in G6PC1, which is essential for proper glucose storage and metabolism. With the advent of continuous glucose monitoring systems, development of algorithms to analyze and predict glucose levels has gained considerable attention, with the aim of preemptively managing fluctuations before they become problematic. However, there is a lack of research focusing specifically on patients with GSD. Therefore, this study aimed to forecast glucose levels in patients with GSD using state-of-the-art deep-learning (DL) algorithms. This retrospective study utilized blood glucose data from patients with GSD who were either hospitalized or managed at Yonsei University Wonju Severance Christian Hospital, Korea, between August 2020 and February 2024. In this study, three state-of-the-art DL models for time-series forecasting were employed: PatchTST, LTSF N-Linear, and TS Mixer. First, the models were used to predict the patients' Glucose levels for the next hour. Second, a binary classification task was performed to assess whether hypoglycemia could be predicted alongside direct glucose levels. Consequently, this is the first study to demonstrate the capability of forecasting glucose levels in patients with GSD using continuous glucose-monitoring data and DL models. Our model provides patients with GSD with a more accessible tool for managing glucose levels. This study has a broader effect, potentially serving as a foundation for improving the care of patients with rare diseases using DL-based solutions.

Accounting for Hypoglycemia Treatments in Continuous Glucose Metrics

BACKGROUND

Continuous glucose monitoring (CGM) is increasingly used in the management of diabetes, providing dense data for patients and clinical providers to review and identify patterns and trends in blood glucose. However, behavioral factors like hypoglycemia treatments (HTs) are not captured in CGM data. Hypoglycemia treatments, by definition, reduce the visibility (frequency and duration) of hypoglycemia exposure recorded by CGM, which can lead to errors in treatment management when relying solely on CGM metrics.

METHODS

We propose a method to incorporate HTs into CGM-based metrics and standardize hypoglycemia exposure quantification for a variety of HT behaviors; specifically (1) treatment proactiveness and (2) potential severity of the avoided hypoglycemia. In addition, we introduce an HT detector to identify instances of HT using in CGM data that otherwise lack HT documentation. We then use the HT-modified hypoglycemia metrics in a previously published run-to-run treatment adaptation system using CGM-based metrics.

RESULTS

Using in-silico data to correct time-below-range with HT, we reconstruct the avoided hypoglycemia exposure with high fidelity (R2 = .94). Our HT detector has an F1 score of 0.72 on clinical data with labeled HT. In the example run-to-run application, we reduce the average number of HT per day from 3.3 in the HT-unaware system to 1.6, while maintaining 84% time in 70 to 180 mg/dL.

CONCLUSION

This new metric integrates HT behaviors into CGM-based analysis, offering a behavior-sensitive measure of hypoglycemia exposure for more robust T1D management. Our results show that HT can be seamlessly incorporated into existing CGM methods, enhancing treatment insights by accounting for HT variability.

Demystifying Infusion Pumps: Design of a Cost-Effective Platform for Education and Innovation

INTRODUCTION

This article presents a cost-effective, modular infusion platform to help diabetes specialists customize and understand infusion pump mechanics and control principles. Traditional insulin pumps are costly and inflexible, limiting accessibility, and particularly in low-resource settings. Inspired by open-source initiatives like OpenAPS, this platform engages specialists in device operation and customization, offering practical insights into infusion technology.

METHOD

An initial survey assessed technological literacy, customization interests, and feature preferences among Mexican diabetes specialists, followed by a hands-on engagement session with the platform's hardware. Core components are described and chosen for reliability, affordability, and integration ease. A follow-up survey evaluated specialists' confidence and interest in device customization, gathering feedback on usability and design.

RESULTS

Survey data showed strong specialist interest in understanding device mechanics and high confidence in customization after hands-on engagement. Most specialists found the hardware layout conducive to experimentation, with significant interest in closed-loop capabilities. Key valued features included safety, affordability, ease of use, customization, and integration of diverse continuous glucose monitors, with added suggestions for potential clinical certification, cost-effective supplies, and artificial intelligence integration.

CONCLUSION

This platform offers a promising educational and developmental tool in diabetes management, bridging clinical application, and customization. Its low-cost, modular design provides a feasible solution for low-resource settings, equipping specialists to tailor devices for specific patient needs. While the platform's educational potential is clear, further studies and validation are essential for a possible transition to a clinical-grade device. Continued development could democratize access to advanced diabetes technology, transforming specialist training, and patient care.

Physical Activity Integration in Blood Glucose Level Prediction: Different Levels of Data Fusion

Blood glucose level (BGL) prediction contributes to more effective management of diabetes. Physical activity (PA), which affects BGL, is a crucial factor in diabetes management. Due to the erratic nature of PA's impact on BGL inter- and intra-patients, deploying PA in BGL prediction is challenging. Hence, it is crucial to discover optimal approaches for utilising PA to improve the performance of BGL prediction. This work contributes to this gap by proposing several PA-informed BGL prediction models. Different approaches are developed to extract information from PA data and integrate this information with BGL data at signal, feature, and decision levels. For signal-level fusion, different automatically-recorded PA data are fused with BGL data. Also, three feature engineering approaches are developed for feature-level fusion: subjective assessments of PA, objective assessments of PA, and statistics of PA. Furthermore, in decision-level fusion, ensemble learning is used to combine predictions from models trained with different inputs. Then, a comparative investigation is performed between the developed PA-informed approaches and the no-fusion approach, as well as between themselves. The analyses are performed on the publicly available Ohio dataset with rigorous evaluation. The results show that among the developed approaches, fusing heart rate data at the signal-level and PA intensity categories at the feature-level with BGL data are effective ways of deploying PA in BGL prediction.

Integrating Bayesian Approaches and Expert Knowledge for Forecasting Continuous Glucose Monitoring Values in Type 2 Diabetes Mellitus

Precise and timely forecasting of blood glucose levels is essential for effective diabetes management. While extensive research has been conducted on Type 1 diabetes mellitus, Type 2 diabetes mellitus (T2DM) presents unique challenges due to its heterogeneity, underscoring the need for specialized blood glucose forecasting systems. This study introduces a novel blood glucose forecasting system, applied to a dataset of 100 patients from the ShanghaiT2DM study. Our study uniquely integrates knowledge-driven and data-driven approaches, leveraging expert knowledge to validate and interpret the relationships among diabetes-related variables and deploying the data-driven approach to provide accurate forecast blood glucose levels. The Bayesian network approach facilitates the analysis of dependencies among various diabetes-related variables, thus enabling the inference of continuous glucose monitoring (CGM) trajectories in similar individuals with T2DM. By incorporating past CGM data including inference CGM trajectories, dietary records, and individual-specific information, the Bayesian structural time series (BSTS) model effectively forecasts glucose levels across time intervals ranging from 15 to 60 minutes. Forecast results show a mean absolute error of mg/dL, a root mean square error of mg/dL, and a mean absolute percentage error of , for a 15-minute prediction horizon. This study makes the first application of the ShanghaiT2DM dataset for glucose level forecasting, considering the influences of diabetes-related variables. Its findings establish a foundational framework for developing personalized diabetes management strategies, potentially enhancing diabetes care through more accurate and timely interventions.

Optimizing hypoglycaemia prediction in type 1 diabetes with Ensemble Machine Learning modeling

BACKGROUND

Type 1 diabetes (T1D) is a chronic endocrine disorder characterized by high blood glucose levels, impacting millions of people globally. Its management requires intensive insulin therapy, frequent blood glucose monitoring, and lifestyle adjustments. The accurate prediction of the short-term course of glucose levels in the subcutaneous space in T1D people, as measured by a continuous glucose monitoring (CGM) system, is essential for improving glucose control by avoiding harmful hypoglycaemic and hyperglycaemic glucose swings, facilitating precise insulin management and individualized care and, in turn, minimizing long-term vascular complications.

METHODS

In this study, we propose an ensemble univariate short-term predictive model of the subcutaneous glucose concentration in T1D targeting at improving its error in the hypoglycaemic region. As such, the underlying basis functions are selected to minimize the percentage of erroneous predictions (EP) in the hypoglycaemic region, with EP being evaluated with continuous glucose error grid analysis (CG-EGA). The dataset comprises 29 individuals with T1D, who were monitored for 2 to 4 weeks during the GlucoseML prospective observational clinical study.

RESULTS

Among six different basis models (i.e., linear regression (LR), automatic relevance determination (ARD), support vector regression (SVR), Gaussian process regression (GPR), eXtreme gradient boosting (XGBoost), and long short-term memory (LSTM)), XGBoost and SVR showed a dominant performance in the hypoglycaemic region and were selected as the constituent basis models of the ensemble model. The results indicate that the ensemble model significantly reduces the percentage of EP in the hypoglycaemic region for a 30 min prediction horizon to 19% as compared with its individual basis models (i.e., XGBoost and SVR), whilst its errors over the entire glucose range (hypoglycaemia, euglycaemia, and hyperglycaemia) are similar to those of the basis models.

CONCLUSIONS

The consideration of the performance of the basis functions in the hypoglycaemic region during the construction of the ensemble model contributes to enhancing their joint performance in that specific area. This could lead to more precise insulin management and a reduced risk of short-term hypoglycaemic fluctuations.

Artificial Intelligence to Diagnose Complications of Diabetes

Artificial intelligence (AI) is increasingly being used to diagnose complications of diabetes. Artificial intelligence is technology that enables computers and machines to simulate human intelligence and solve complicated problems. In this article, we address current and likely future applications for AI to be applied to diabetes and its complications, including pharmacoadherence to therapy, diagnosis of hypoglycemia, diabetic eye disease, diabetic kidney diseases, diabetic neuropathy, diabetic foot ulcers, and heart failure in diabetes.Artificial intelligence is advantageous because it can handle large and complex datasets from a variety of sources. With each additional type of data incorporated into a clinical picture of a patient, the calculation becomes increasingly complex and specific. Artificial intelligence is the foundation of emerging medical technologies; it will power the future of diagnosing diabetes complications.

Continuous glucose monitoring using machine learning models and IoT device data: A meta-analysis

BACKGROUND

Machine learning offers diverse options for effectively managing blood glucose levels in diabetes patients. Selecting the right ML algorithm is critical given the array of available choices. Integrating data from IoT devices presents promising opportunities to enhance real-time blood glucose management models.

OBJECTIVE

This meta-analysis aims to evaluate the effectiveness of machine learning models utilizing IoT device data for predicting blood glucose levels.

METHODS

We systematically searched electronic databases for studies published between 2019 and 2023. We excluded studies lacking ML model derivation or performance metrics. The Quality Assessment of Diagnostic Accuracy Studies tool assessed study quality. Our primary outcomes compared ML models for BG level prediction across different prediction horizons (PHs).

RESULTS

We analyzed ten eligible studies across prediction horizons of 15, 30, 45, and 60 minutes. ML models exhibited mean absolute RMSE values of 15.02 (SD 1.45), 21.488 (SD 2.92), 30.094 (SD 3.245), and 35.89 (SD 6.4) mg/dL, respectively. Random Forest demonstrated superior performance across these PHs.

CONCLUSION

We observed significant heterogeneity across all subgroups, indicating diverse sources of variability. As the PH lengthened, the RMSE for blood glucose prediction by the ML model increased, with Random Forest showing the highest relative performance among the ML models.

Predicting Dysglycemia in Patients with Diabetes Using Electrocardiogram

Background: In this study, we explored the potential of predicting dysglycemia in patients who need to continuously manage blood glucose levels using a non-invasive method via electrocardiography (ECG). Methods: The data were collected from patients with diabetes, and heart rate variability (HRV) features were extracted via ECG processing. A residual block-based one-dimensional convolution neural network model was used to predict dysglycemia. Results: The dysglycemia prediction results at each time point, including at the time of blood glucose measurement, 15 min prior to measurement, and 30 min prior to measurement, exhibited no significant differences compared with the blood glucose measurement values. This result confirmed that the proposed artificial intelligence model for dysglycemia prediction performed well at each time point. Additionally, to determine the optimal number of features required for predicting dysglycemia, 77 HRV features were individually eliminated in the order of decreasing importance with respect to the prediction accuracy; the optimal number of features for the model to predict dysglycemia was determined to be 12. The dysglycemia prediction results obtained 30 min prior to measurement, which exhibited the highest prediction range in this study, were as follows: accuracy = 90.5, sensitivity = 87.52, specificity = 92.74, and precision = 89.86. Conclusions: Furthermore, we determined that no significant differences exist in the blood glucose prediction results reported in previous studies, wherein various vital signs and blood glucose values were used as model inputs, and the results obtained in this study, wherein only ECG data were used to predict dysglycemia.

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.

Enhancing the Capabilities of Continuous Glucose Monitoring With a Predictive App

BACKGROUND

Despite abundant evidence demonstrating the benefits of continuous glucose monitoring (CGM) in diabetes management, a significant proportion of people using this technology still struggle to achieve glycemic targets. To address this challenge, we propose the Accu-Chek® SmartGuide Predict app, an innovative CGM digital companion that incorporates a suite of advanced glucose predictive functionalities aiming to inform users earlier about acute glycemic situations.

METHODS

The app's functionalities, powered by three machine learning models, include a two-hour glucose forecast, a 30-minute low glucose detection, and a nighttime low glucose prediction for bedtime interventions. Evaluation of the models' performance included three data sets, comprising subjects with T1D on MDI (n = 21), subjects with type 2 diabetes (T2D) on MDI (n = 59), and subjects with T1D on insulin pump therapy (n = 226).

RESULTS

On an aggregated data set, the two-hour glucose prediction model, at a forecasting horizon of 30, 45, 60, and 120 minutes, achieved a percentage of data points in zones A and B of Consensus Error Grid of: 99.8%, 99.3%, 98.7%, and 96.3%, respectively. The 30-minute low glucose prediction model achieved an accuracy, sensitivity, specificity, mean lead time, and area under the receiver operating characteristic curve (ROC AUC) of: 98.9%, 95.2%, 98.9%, 16.2 minutes, and 0.958, respectively. The nighttime low glucose prediction model achieved an accuracy, sensitivity, specificity, and ROC AUC of: 86.5%, 55.3%, 91.6%, and 0.859, respectively.

CONCLUSIONS

The consistency of the performance of the three predictive models when evaluated on different cohorts of subjects with T1D and T2D on different insulin therapies, including real-world data, offers reassurance for real-world efficacy.

Nocturnal Hypoglycemia in the Era of Continuous Glucose Monitoring

Nocturnal hypoglycemia is a common acute complication of people with diabetes on insulin therapy. In particular, the inability to control glucose levels during sleep, the impact of external factors such as exercise, or alcohol and the influence of hormones are the main causes. Nocturnal hypoglycemia has several negative somatic, psychological, and social effects for people with diabetes, which are summarized in this article. With the advent of continuous glucose monitoring (CGM), it has been shown that the number of nocturnal hypoglycemic events was significantly underestimated when traditional blood glucose monitoring was used. The CGM can reduce the number of nocturnal hypoglycemia episodes with the help of alarms, trend arrows, and evaluation routines. In combination with CGM with an insulin pump and an algorithm, automatic glucose adjustment (AID) systems have their particular strength in nocturnal glucose regulation and the prevention of nocturnal hypoglycemia. Nevertheless, the problem of nocturnal hypoglycemia has not yet been solved completely with the technologies currently available. The CGM systems that use predictive models to warn of hypoglycemia, improved AID systems that recognize hypoglycemia patterns even better, and the increasing integration of artificial intelligence methods are promising approaches in the future to significantly minimize the risk of a side effect of insulin therapy that is burdensome for people with diabetes.

Clinical translation of nanomedicine with integrated digital medicine and machine learning interventions

Nanomaterials based therapeutics transform the ways of disease prevention, diagnosis and treatment with increasing sophistications in nanotechnology at a breakneck pace, but very few could reach to the clinic due to inconsistencies in preclinical studies followed by regulatory hinderances. To tackle this, integrating the nanomedicine discovery with digital medicine provide technologies as tools of specific biological activity measurement. Hence, overcome the redundancies in nanomedicine discovery by the on-site data acquisition and analytics through integrating intelligent sensors and artificial intelligence (AI) or machine learning (ML). Integrated AI/ML wearable sensors directly gather clinically relevant biochemical information from the subject's body and process data for physicians to make right clinical decision(s) in a time and cost-effective way. This review summarizes insights and recommend the infusion of actionable big data computation enabled sensors in burgeoning field of nanomedicine at academia, research institutes, and pharmaceutical industries, with a potential of clinical translation. Furthermore, many blind spots are present in modern clinically relevant computation, one of which could prevent ML-guided low-cost new nanomedicine development from being successfully translated into the clinic was also discussed.

Exploring Nutritional Influence on Blood Glucose Forecasting for Type 1 Diabetes Using Explainable AI

Type 1 diabetes mellitus (T1DM) is characterized by insulin deficiency and blood sugar control issues. The state-of-the-art solution is the artificial pancreas (AP), which integrates basal insulin delivery and glucose monitoring. However, APs are unable to manage postprandial glucose response (PGR) due to limited knowledge of its determinants, requiring additional information for accurate bolus delivery, such as estimated carbohydrate intake. This study aims to quantify the influence of various meal-related factors on predicting postprandial blood glucose levels (BGLs) at different time intervals (15 min, 60 min, and 120 min) after meals by using deep neural network (DNN) models. The prediction models incorporate preprandial blood glucose values, insulin dosage, and various meal-related nutritional factors such as intake of energy, carbohydrates, proteins, lipids, fatty acids, fibers, glycemic index, and glycemic load as input variables. The impact of input features was assessed by exploiting eXplainable Artificial Intelligence (XAI) methodologies, specifically SHapley Additive exPlanations (SHAP), which provide insights into each feature's contribution to the model predictions. By leveraging XAI methodologies, this study aims to enhance the interpretability and transparency of BGL prediction models and validate clinical literature hypotheses. The findings can aid in the development of decision-support tools for individuals with T1DM, facilitating PGR management and reducing the risks of adverse events. The improved understanding of PGR determinants may lead to advancements in AP technology and improve the overall quality of life for T1DM patients.

Population-Specific Glucose Prediction in Diabetes Care With Transformer-Based Deep Learning on the Edge

Leveraging continuous glucose monitoring (CGM) systems, real-time blood glucose (BG) forecasting is essential for proactive interventions, playing a crucial role in enhancing the management of type 1 diabetes (T1D) and type 2 diabetes (T2D). However, developing a model generalized to a population and subsequently embedding it within a microchip of a wearable device presents significant technical challenges. Furthermore, the domain of BG prediction in T2D remains under-explored in the literature. In light of this, we propose a population-specific BG prediction model, leveraging the capabilities of the temporal fusion Transformer (TFT) to adjust predictions based on personal demographic data. Then the trained model is embedded within a system-on-chip, integral to our low-power and low-cost customized wearable device. This device seamlessly communicates with CGM systems through Bluetooth and provides timely BG predictions using edge computing. When evaluated on two publicly available clinical datasets with a total of 124 participants with T1D or T2D, the embedded TFT model consistently demonstrated superior performance, achieving the lowest prediction errors when compared with a range of machine learning baseline methods. Executing the TFT model on our wearable device requires minimal memory and power consumption, enabling continuous decision support for more than 51 days on a single Li-Poly battery charge. These findings demonstrate the significant potential of the proposed TFT model and wearable device in enhancing the quality of life for people with diabetes and effectively addressing real-world challenges.

A Machine Learning Model for Week-Ahead Hypoglycemia Prediction From Continuous Glucose Monitoring Data

BACKGROUND

Remote patient monitoring (RPM) programs augment type 1 diabetes (T1D) care based on retrospective continuous glucose monitoring (CGM) data. Few methods are available to estimate the likelihood of a patient experiencing clinically significant hypoglycemia within one week.

METHODS

We developed a machine learning model to estimate the probability that a patient will experience a clinically significant hypoglycemic event, defined as CGM readings below 54 mg/dL for at least 15 consecutive minutes, within one week. The model takes as input the patient's CGM time series over a given week, and outputs the predicted probability of a clinically significant hypoglycemic event the following week. We used 10-fold cross-validation and external validation (testing on cohorts different from the training cohort) to evaluate performance. We used CGM data from three different cohorts of patients with T1D: REPLACE-BG (226 patients), Juvenile Diabetes Research Foundation (JDRF; 355 patients) and Tidepool (120 patients).

RESULTS

In 10-fold cross-validation, the average area under the receiver operating characteristic curve (ROC-AUC) was 0.77 (standard deviation [SD]: 0.0233) on the REPLACE-BG cohort, 0.74 (SD: 0.0188) on the JDRF cohort, and 0.76 (SD: 0.02) on the Tidepool cohort. In external validation, the average ROC-AUC across the three cohorts was 0.74 (SD: 0.0262).

CONCLUSIONS

We developed a machine learning algorithm to estimate the probability of a clinically significant hypoglycemic event within one week. Predictive algorithms may provide diabetes care providers using RPM with additional context when prioritizing T1D patients for review.

Deep Learning-Based Glucose Prediction Models: A Guide for Practitioners and a Curated Dataset for Improved Diabetes Management

Accurate short- and mid-term blood glucose predictions are crucial for patients with diabetes struggling to maintain healthy glucose levels, as well as for individuals at risk of developing the disease. Consequently, numerous efforts from the scientific community have focused on developing predictive models for glucose levels. This study harnesses physiological data collected from wearable sensors to construct a series of data-driven models based on deep learning approaches. We systematically compare these models to offer insights for practitioners and researchers venturing into glucose prediction using deep learning techniques. Key questions addressed in this work encompass the comparison of various deep learning architectures for this task, determining the optimal set of input variables for accurate glucose prediction, comparing population-wide, fine-tuned, and personalized models, and assessing the impact of an individual's data volume on model performance. Additionally, as part of our outcomes, we introduce a meticulously curated dataset inclusive of data from both healthy individuals and those with diabetes, recorded in free-living conditions. This dataset aims to foster research in this domain and facilitate equitable comparisons among researchers.

Fully integrated wearable microneedle biosensing platform for wide-range and real-time continuous glucose monitoring

Wearable microneedle sensors for continuous glucose monitoring (CGM) have great potential for clinical impact by allowing access to large data sets to provide individualized treatment plans. To date, their development has been challenged by the accurate wide linear range tracking of interstitial fluid (ISF) glucose (Glu) levels. Here, we present a CGM platform consisting of a three-electrode microneedle electrochemical biosensor and a fully integrated radio-chemical analysis system. The long-term performance of the robust CGM on diabetic rats was achieved by electrodepositing Prussian blue (PB), and crosslinking glucose oxidase (GOx) and chitosan to form a 3D network using glutaraldehyde (GA). After redox by GOx, PB rapidly decomposes hydrogen peroxide and mediates charge transfer, while the 3D network and graphite powder provide enrichment and release sites for Glu and catalytic products, enabling a sensing range of 0.25-35 mM. Microneedle CGM has high sensitivity, good stability, and anti-interference ability. In diabetic rats, CGM can accurately monitor Glu levels in the ISF in real-time, which are highly consistent with levels measured by commercial Glu meters. These results indicate the feasibility and application prospects of the PB-based CGM for the clinical management of diabetes. STATEMENT OF SIGNIFICANCE: This study addresses the challenge of continuous glucose monitoring system design where the narrow linear range of sensing due to the miniaturization of sensors fails to meet the monitoring needs of clinical diabetic patients. This was achieved by utilizing a three-dimensional network of glutaraldehyde cross-linked glucose oxidase and chitosan. The unique topology of the 3D network provides a large number of sites for glucose enrichment and anchors the enzyme to the sensing medium and the conductive substrate through covalent bonding, successfully blocking the escape of the enzyme and the sensing medium and shortening the electron transfer and transmission path.

Artificial Intelligence and Machine Learning for Improving Glycemic Control in Diabetes: Best Practices, Pitfalls, and Opportunities

OBJECTIVE

Artificial intelligence and machine learning are transforming many fields including medicine. In diabetes, robust biosensing technologies and automated insulin delivery therapies have created a substantial opportunity to improve health. While the number of manuscripts addressing the topic of applying machine learning to diabetes has grown in recent years, there has been a lack of consistency in the methods, metrics, and data used to train and evaluate these algorithms. This manuscript provides consensus guidelines for machine learning practitioners in the field of diabetes, including best practice recommended approaches and warnings about pitfalls to avoid.

METHODS

Algorithmic approaches are reviewed and benefits of different algorithms are discussed including importance of clinical accuracy, explainability, interpretability, and personalization. We review the most common features used in machine learning applications in diabetes glucose control and provide an open-source library of functions for calculating features, as well as a framework for specifying data sets using data sheets. A review of current data sets available for training algorithms is provided as well as an online repository of data sources.

SIGNIFICANCE

These consensus guidelines are designed to improve performance and translatability of new machine learning algorithms developed in the field of diabetes for engineers and data scientists.

Machine learning in precision diabetes care and cardiovascular risk prediction

Artificial intelligence and machine learning are driving a paradigm shift in medicine, promising data-driven, personalized solutions for managing diabetes and the excess cardiovascular risk it poses. In this comprehensive review of machine learning applications in the care of patients with diabetes at increased cardiovascular risk, we offer a broad overview of various data-driven methods and how they may be leveraged in developing predictive models for personalized care. We review existing as well as expected artificial intelligence solutions in the context of diagnosis, prognostication, phenotyping, and treatment of diabetes and its cardiovascular complications. In addition to discussing the key properties of such models that enable their successful application in complex risk prediction, we define challenges that arise from their misuse and the role of methodological standards in overcoming these limitations. We also identify key issues in equity and bias mitigation in healthcare and discuss how the current regulatory framework should ensure the efficacy and safety of medical artificial intelligence products in transforming cardiovascular care and outcomes in diabetes.

Uncovering personalized glucose responses and circadian rhythms from multiple wearable biosensors with Bayesian dynamical modeling

Wearable biosensors and smartphone applications can measure physiological variables over multiple days in free-living conditions. We measure food and drink ingestion, glucose dynamics, physical activity, heart rate (HR), and heart rate variability (HRV) in 25 healthy participants over 14 days. We develop a Bayesian inference framework to learn personal parameters that quantify circadian rhythms and physiological responses to external stressors. Modeling the effects of ingestion events on glucose levels reveals that slower glucose decay kinetics elicit larger postprandial glucose spikes, and we uncover a circadian baseline rhythm for glucose with high amplitudes in some individuals. Physical activity and circadian rhythms explain as much as 40%-65% of the HR variance, whereas the variance explained for HRV is more heterogeneous across individuals. A more complex model incorporating activity, HR, and HRV explains up to 15% of additional glucose variability, highlighting the relevance of integrating multiple biosensors to better predict glucose dynamics.

Machine Learning-Based Time in Patterns for Blood Glucose Fluctuation Pattern Recognition in Type 1 Diabetes Management: Development and Validation Study

BACKGROUND

Continuous glucose monitoring (CGM) for diabetes combines noninvasive glucose biosensors, continuous monitoring, cloud computing, and analytics to connect and simulate a hospital setting in a person's home. CGM systems inspired analytics methods to measure glycemic variability (GV), but existing GV analytics methods disregard glucose trends and patterns; hence, they fail to capture entire temporal patterns and do not provide granular insights about glucose fluctuations.

OBJECTIVE

This study aimed to propose a machine learning-based framework for blood glucose fluctuation pattern recognition, which enables a more comprehensive representation of GV profiles that could present detailed fluctuation information, be easily understood by clinicians, and provide insights about patient groups based on time in blood fluctuation patterns.

METHODS

Overall, 1.5 million measurements from 126 patients in the United Kingdom with type 1 diabetes mellitus (T1DM) were collected, and prevalent blood fluctuation patterns were extracted using dynamic time warping. The patterns were further validated in 225 patients in the United States with T1DM. Hierarchical clustering was then applied on time in patterns to form 4 clusters of patients. Patient groups were compared using statistical analysis.

RESULTS

In total, 6 patterns depicting distinctive glucose levels and trends were identified and validated, based on which 4 GV profiles of patients with T1DM were found. They were significantly different in terms of glycemic statuses such as diabetes duration (P=.04), glycated hemoglobin level (P<.001), and time in range (P<.001) and thus had different management needs.

CONCLUSIONS

The proposed method can analytically extract existing blood fluctuation patterns from CGM data. Thus, time in patterns can capture a rich view of patients' GV profile. Its conceptual resemblance with time in range, along with rich blood fluctuation details, makes it more scalable, accessible, and informative to clinicians.

Insulin detection in diabetes mellitus: challenges and new prospects

Tremendous progress has been made towards achieving tight glycaemic control in individuals with diabetes mellitus through the use of frequent or continuous glucose measurements. However, in patients who require insulin, accurate dosing must consider multiple factors that affect insulin sensitivity and modulate insulin bolus needs. Accordingly, an urgent need exists for frequent and real-time insulin measurements to closely track the dynamic blood concentration of insulin during insulin therapy and guide optimal insulin dosing. Nevertheless, traditional centralized insulin testing cannot offer timely measurements, which are essential to achieving this goal. This Perspective discusses the advances and challenges in moving insulin assays from traditional laboratory-based assays to frequent and continuous measurements in decentralized (point-of-care and home) settings. Technologies that hold promise for insulin testing using disposable test strips, mobile systems and wearable real-time insulin-sensing devices are discussed. We also consider future prospects for continuous insulin monitoring and for fully integrated multisensor-guided closed-loop artificial pancreas systems.

The role of patient-reported outcome measures in trials of artificial intelligence health technologies: a systematic evaluation of ClinicalTrials.gov records (1997-2022)

The extent to which patient-reported outcome measures (PROMs) are used in clinical trials for artificial intelligence (AI) technologies is unknown. In this systematic evaluation, we aim to establish how PROMs are being used to assess AI health technologies. We searched ClinicalTrials.gov for interventional trials registered from inception to Sept 20, 2022, and included trials that tested an AI health technology. We excluded observational studies, patient registries, and expanded access reports. We extracted data regarding the form, function, and intended use population of the AI health technology, in addition to the PROMs used and whether PROMs were incorporated as an input or output in the AI model. The search identified 2958 trials, of which 627 were included in the analysis. 152 (24%) of the included trials used one or more PROM, visual analogue scale, patient-reported experience measure, or usability measure as a trial endpoint. The type of AI health technologies used by these trials included AI-enabled smart devices, clinical decision support systems, and chatbots. The number of clinical trials of AI health technologies registered on ClinicalTrials.gov and the proportion of trials that used PROMs increased from registry inception to 2022. The most common clinical areas AI health technologies were designed for were digestive system health for non-PROM trials and musculoskeletal health (followed by mental and behavioural health) for PROM trials, with PROMs commonly used in clinical areas for which assessment of health-related quality of life and symptom burden is particularly important. Additionally, AI-enabled smart devices were the most common applications tested in trials that used at least one PROM. 24 trials tested AI models that captured PROM data as an input for the AI model. PROM use in clinical trials of AI health technologies falls behind PROM use in all clinical trials. Trial records having inadequate detail regarding the PROMs used or the type of AI health technology tested was a limitation of this systematic evaluation and might have contributed to inaccuracies in the data synthesised. Overall, the use of PROMs in the function and assessment of AI health technologies is not only possible, but is a powerful way of showing that, even in the most technologically advanced health-care systems, patients' perspectives remain central.

Modeling risk of hypoglycemia during and following physical activity in people with type 1 diabetes using explainable mixed-effects machine learning

BACKGROUND

Physical activity (PA) can cause increased hypoglycemia (glucose <70 mg/dL) risk in people with type 1 diabetes (T1D). We modeled the probability of hypoglycemia during and up to 24 h following PA and identified key factors associated with hypoglycemia risk.

METHODS

We leveraged a free-living dataset from Tidepool comprised of glucose measurements, insulin doses, and PA data from 50 individuals with T1D (6448 sessions) for training and validating machine learning models. We also used data from the T1Dexi pilot study that contains glucose management and PA data from 20 individuals with T1D (139 session) for assessing the accuracy of the best performing model on an independent test dataset. We used mixed-effects logistic regression (MELR) and mixed-effects random forest (MERF) to model hypoglycemia risk around PA. We identified risk factors associated with hypoglycemia using odds ratio and partial dependence analysis for the MELR and MERF models, respectively. Prediction accuracy was measured using the area under the receiver operating characteristic curve (AUROC).

RESULTS

The analysis identified risk factors significantly associated with hypoglycemia during and following PA in both MELR and MERF models including glucose and body exposure to insulin at the start of PA, low blood glucose index 24 h prior to PA, and PA intensity and timing. Both models showed overall hypoglycemia risk peaking 1 h after PA and again 5-10 h after PA, which is consistent with the hypoglycemia risk pattern observed in the training dataset. Time following PA impacted hypoglycemia risk differently across different PA types. Accuracy of hypoglycemia prediction using the fixed effects of the MERF model was highest when predicting hypoglycemia during the first hour following the start of PA (AUROCVALIDATION = 0.83 and AUROCTESTING = 0.86) and decreased when predicting hypoglycemia in the 24 h after PA (AUROCVALIDATION = 0.66 and AUROCTESTING = 0.68).

CONCLUSION

Hypoglycemia risk after the start of PA can be modeled using mixed-effects machine learning to identify key risk factors that may be used within decision support and insulin delivery systems. We published the population-level MERF model online for others to use.

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.

3D Printing Technology for Smart Clothing: A Topic Review

Clothing is considered to be an important element of human social activities. With the increasing maturity of 3D printing technology, functional 3D printing technology can realize the perfect combination of clothing and electronic devices while helping smart clothing to achieve specific functions. Furthermore, the application of functional 3D printing technology in clothing not only provides people with the most comfortable and convenient wearing experience, but also completely subverts consumers' perception of traditional clothing. This paper introduced the progress of the application of 3D printing from the aspect of traditional clothing and smart clothing through two mature 3D printing technologies normally used in the field of clothing, and summarized the challenges and prospects of 3D printing technology in the field of smart clothing. Finally, according to the analysis of the gap between 3D-printed clothing and traditionally made clothing due to the material limitations, this paper predicted that the rise in intelligent materials will provide a new prospect for the development of 3D-printed clothing. This paper will provide some references for the application research of 3D printing in the field of smart clothing.

Characteristics of Artificial Intelligence Clinical Trials in the Field of Healthcare: A Cross-Sectional Study on ClinicalTrials.gov

Artificial intelligence (AI) has driven innovative transformation in healthcare service patterns, despite a lack of understanding of its performance in clinical practice. We conducted a cross-sectional analysis of AI-related trials in healthcare based on ClinicalTrials.gov, intending to investigate the trial characteristics and AI's development status. Additionally, the Neo4j graph database and visualization technology were employed to construct an AI technology application graph, achieving a visual representation and analysis of research hotspots in healthcare AI. A total of 1725 eligible trials that were registered in ClinicalTrials.gov up to 31 March 2022 were included in this study. The number of trial registrations has dramatically grown each year since 2016. However, the AI-related trials had some design drawbacks and problems with poor-quality result reporting. The proportion of trials with prospective and randomized designs was insufficient, and most studies did not report results upon completion. Currently, most healthcare AI application studies are based on data-driven learning algorithms, covering various disease areas and healthcare scenarios. As few studies have publicly reported results on ClinicalTrials.gov, there is not enough evidence to support an assessment of AI's actual performance. The widespread implementation of AI technology in healthcare still faces many challenges and requires more high-quality prospective clinical validation.