Peculiarities of the continuous glucose monitoring data stream and their impact on developing closed-loop control technology

DEMIGOD: A Low-Cost Microcontroller-Based Closed-Loop System Integrating Nanoengineered Sweat-Based Glucose Monitoring and Controlled Transdermal Nanoemulsion Release of Hypoglycemic Treatment with a Software Application for Noninvasive Personalized Diabetes Care

This study endeavored to design and develop an innovative closed-loop diagnostic and therapeutic system with the following objectives: (a) the noninvasive detection of glucose concentration in sweat utilizing nanonengineered screen-printed biosensors; (b) the management of measured data through a specialized computer system comprising both hardware and software components, thereby enabling the precise control of therapeutic responses via a patch-based nanomedicine delivery system. This initiative addresses the significant challenges inherent in the management of diabetes mellitus, including the imperative need for glucose-level monitoring to optimize glycemic control. Leveraging chronoamperometric results as a foundational dataset and the in vivo hypoglycemic activity of nanoemulsion formulations, this research underscores the efficacy and accuracy of glucose concentration estimation, decision-making mechanism responses, and transdermal hypoglycemic treatment effects, within the proposed system.

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.

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.

Continuous Glucose Monitoring Time Series Data Analysis: A Time Series Analysis Package for Continuous Glucose Monitoring Data

The R package Continuous Glucose Monitoring Time Series Data Analysis (CGMTSA) was developed to facilitate investigations that examine the continuous glucose monitoring (CGM) data as a time series. Accordingly, novel time series functions were introduced to (1) enable more accurate missing data imputation and outlier identification; (2) calculate recommended CGM metrics as well as key time series parameters; (3) plot interactive and three-dimensional graphs that allow direct visualizations of temporal CGM data and time series model optimization. The software was designed to accommodate all popular CGM devices and support all common data processing steps. The program is available for Linux, Windows, and Mac at GitHub.

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.

Ability of Current Machine Learning Algorithms to Predict and Detect Hypoglycemia in Patients With Diabetes Mellitus: Meta-analysis

Background

Machine learning (ML) algorithms have been widely introduced to diabetes research including those for the identification of hypoglycemia.

Objective

The objective of this meta-analysis is to assess the current ability of ML algorithms to detect hypoglycemia (ie, alert to hypoglycemia coinciding with its symptoms) or predict hypoglycemia (ie, alert to hypoglycemia before its symptoms have occurred).

Methods

Electronic literature searches (from January 1, 1950, to September 14, 2020) were conducted using the Dialog platform that covers 96 databases of peer-reviewed literature. Included studies had to train the ML algorithm in order to build a model to detect or predict hypoglycemia and test its performance. The set of 2 × 2 data (ie, number of true positives, false positives, true negatives, and false negatives) was pooled with a hierarchical summary receiver operating characteristic model.

Results

A total of 33 studies (14 studies for detecting hypoglycemia and 19 studies for predicting hypoglycemia) were eligible. For detection of hypoglycemia, pooled estimates (95% CI) of sensitivity, specificity, positive likelihood ratio (PLR), and negative likelihood ratio (NLR) were 0.79 (0.75-0.83), 0.80 (0.64-0.91), 8.05 (4.79-13.51), and 0.18 (0.12-0.27), respectively. For prediction of hypoglycemia, pooled estimates (95% CI) were 0.80 (0.72-0.86) for sensitivity, 0.92 (0.87-0.96) for specificity, 10.42 (5.82-18.65) for PLR, and 0.22 (0.15-0.31) for NLR.

Conclusions

Current ML algorithms have insufficient ability to detect ongoing hypoglycemia and considerate ability to predict impeding hypoglycemia in patients with diabetes mellitus using hypoglycemic drugs with regard to diagnostic tests in accordance with the Users’ Guide to Medical Literature (PLR should be ≥5 and NLR should be ≤0.2 for moderate reliability). However, it should be emphasized that the clinical applicability of these ML algorithms should be evaluated according to patients’ risk profiles such as for hypoglycemia and its associated complications (eg, arrhythmia, neuroglycopenia) as well as the average ability of the ML algorithms. Continued research is required to develop more accurate ML algorithms than those that currently exist and to enhance the feasibility of applying ML in clinical settings.

Trial Registration

PROSPERO International Prospective Register of Systematic Reviews CRD42020163682; http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42020163682

Glycemic Variability: Risk Factors, Assessment, and Control

Glycemic variability (GV) a well-established risk factor for hypoglycemia and a suspected risk factor for diabetes complications. GV is also a marker of the instability of a person's metabolic system, expressed as frequent high and low glucose excursions and overall volatile glycemic control. In this review, the author discusses topics related to the assessment, quantification, and optimal control of diabetes, including (1) the notion that optimal control of diabetes, that is, lowering of HbA1c-the commonly accepted gold-standard outcome-can be achieved only if accompanied by simultaneous reduction of GV; (2) assessment and visualization of the two principal dimensions of GV, amplitude and time, which is now possible via continuous glucose monitoring (CGM) and various metrics quantifying GV and the risks associated with hypo- and hyperglycemic excursions; and (3) the evolution of diabetes science and technology beyond quantifying GV and into the realm of GV control via pharmacological agents, for example, GLP-1 receptor agonists and DPP-4 inhibitors, which have pronounced variability-reducing effect, or real-time automated closed-loop systems commonly referred to as the "artificial pancreas." The author concludes that CGM allows close tracking over time, and therefore precise quantification, of glycemic variability in diabetes. The next step-optimal control of glucose fluctuations-is also taken by medications with pronounced GV-lowering effect primarily in type 2 diabetes, and by automated insulin delivery in type 1 diabetes. Contemporary CGM-based artificial pancreas systems use specific GV representations as input signals, and thus their main objective is to minimize GV and, from there, optimize glycemic control.

A Century of Diabetes Technology: Signals, Models, and Artificial Pancreas Control

Arguably, diabetes mellitus is one of the best-quantified human conditions: elaborate in silico models describe the action of the human metabolic system; real-time signals such as continuous glucose monitoring are readily available; insulin delivery is being automated; and control algorithms are capable of optimizing blood glucose fluctuation in patients' natural environments. The transition of the artificial pancreas (AP) to everyday clinical use is happening now, and is contingent upon seamless concerted work of devices encompassing the patient in a digital treatment ecosystem. This review recounts briefly the story of diabetes technology, which began a century ago with the discovery of insulin, progressed through glucose monitoring and subcutaneous insulin delivery, and is now rapidly advancing towards fully automated clinically viable AP systems.

Automated closed-loop control of diabetes: the artificial pancreas

The incidence of Diabetes Mellitus is on the rise worldwide, which exerts enormous health toll on the population and enormous pressure on the healthcare systems. Now, almost hundred years after the discovery of insulin in 1921, the optimization problem of diabetes is well formulated as maintenance of strict glycemic control without increasing the risk for hypoglycemia. External insulin administration is mandatory for people with type 1 diabetes; various medications, as well as basal and prandial insulin, are included in the daily treatment of type 2 diabetes. This review follows the development of the Diabetes Technology field which, since the 1970s, progressed remarkably through continuous subcutaneous insulin infusion (CSII), mathematical models and computer simulation of the human metabolic system, real-time continuous glucose monitoring (CGM), and control algorithms driving closed-loop control systems known as the "artificial pancreas" (AP). All of these developments included significant engineering advances and substantial bioelectronics progress in the sensing of blood glucose levels, insulin delivery, and control design. The key technologies that enabled contemporary AP systems are CSII and CGM, both of which became available and sufficiently portable in the beginning of this century. This powered the quest for wearable home-use AP, which is now under way with prototypes tested in outpatient studies during the past 6 years. Pivotal trials of new AP technologies are ongoing, and the first hybrid closed-loop system has been approved by the FDA for clinical use. Thus, the closed-loop AP is well on its way to become the digital-age treatment of diabetes.

Short-term prediction of glucose in type 1 diabetes using kernel adaptive filters

This study aims at presenting a nonlinear, recursive, multivariate prediction model of the subcutaneous glucose concentration in type 1 diabetes. Nonlinear regression is performed in a reproducing kernel Hilbert space, by either the fixed budget quantized kernel least mean square (QKLMS-FB) or the approximate linear dependency kernel recursive least-squares (KRLS-ALD) algorithm, such that a sparse model structure is accomplished. A multivariate feature set (i.e., subcutaneous glucose, food carbohydrates, insulin regime and physical activity) is used and its influence on short-term glucose prediction is investigated. The method is evaluated using data from 15 patients with type 1 diabetes in free-living conditions. In the case when all the input variables are considered: (i) the average root mean squared error (RMSE) of QKLMS-FB increases from 13.1 mg dL^-1 (mean absolute percentage error (MAPE) 6.6%) for a 15-min prediction horizon (PH) to 37.7 mg dL^-1 (MAPE 20.8%) for a 60-min PH and (ii) the RMSE of KRLS-ALD, being predictably lower, increases from 10.5 mg dL^-1 (MAPE 5.2%) for a 15-min PH to 31.8 mg dL^-1 (MAPE 18.0%) for a 60-min PH. Multivariate data improve systematically both the regularity and the time lag of the predictions, reducing the errors in critical glucose value regions for a PH ≥ 30 min. Graphical abstract ᅟ.

Metrics for glycaemic control - from HbA1c to continuous glucose monitoring

As intensive treatment to lower levels of HbA_1c characteristically results in an increased risk of hypoglycaemia, patients with diabetes mellitus face a life-long optimization problem to reduce average levels of glycaemia and postprandial hyperglycaemia while simultaneously avoiding hypoglycaemia. This optimization can only be achieved in the context of lowering glucose variability. In this Review, I discuss topics that are related to the assessment, quantification and optimal control of glucose fluctuations in diabetes mellitus. I focus on markers of average glycaemia and the utility and/or shortcomings of HbA_1c as a 'gold-standard' metric of glycaemic control; the notion that glucose variability is characterized by two principal dimensions, amplitude and time; measures of glucose variability that are based on either self-monitoring of blood glucose data or continuous glucose monitoring (CGM); and the control of average glycaemia and glucose variability through the use of pharmacological agents or closed-loop control systems commonly referred to as the 'artificial pancreas'. I conclude that HbA_1c and the various available metrics of glucose variability reflect the management of diabetes mellitus on different timescales, ranging from months (for HbA_1c) to minutes (for CGM). Comprehensive assessment of the dynamics of glycaemic fluctuations is therefore crucial for providing accurate and complete information to the patient, physician, automated decision-support or artificial-pancreas system.

Analysis of the Accuracy and Performance of a Continuous Glucose Monitoring Sensor Prototype: An In-Silico Study Using the UVA/PADOVA Type 1 Diabetes Simulator

BACKGROUND

Computer simulation has been shown over the past decade to be a powerful tool to study the impact of medical devices characteristics on clinical outcomes. Specifically, in type 1 diabetes (T1D), computer simulation platforms have all but replaced preclinical studies and are commonly used to study the impact of measurement errors on glycemia.

METHOD

We use complex mathematical models to represent the characteristics of 3 continuous glucose monitoring systems using previously acquired data. Leveraging these models within the framework of the UVa/Padova T1D simulator, we study the impact of CGM errors in 6 simulation scenarios designed to generate a wide variety of glycemic conditions. Assessment of the simulated accuracy of each different CGM systems is performed using mean absolute relative deviation (MARD) and precision absolute relative deviation (PARD). We also quantify the capacity of each system to detect hypoglycemic events.

RESULTS

The simulated Roche CGM sensor prototype (RCGM) outperformed the 2 alternate systems (CGM-1 & CGM-2) in accuracy (MARD = 8% vs 11.4% vs 18%) and precision (PARD = 6.4% vs 9.4% vs 14.1%). These results held for all studied glucose and rate of change ranges. Moreover, it detected more than 90% of hypoglycemia, with a mean time lag less than 4 minutes (CGM-1: 86%/15 min, CGM-2: 57%/24 min).

CONCLUSION

The RCGM system model led to strong performances in these simulation studies, with higher accuracy and precision than alternate systems. Its characteristics placed it firmly as a strong candidate for CGM based therapy, and should be confirmed in large clinical studies.

Evaluation of short-term predictors of glucose concentration in type 1 diabetes combining feature ranking with regression models

Glucose concentration in type 1 diabetes is a function of biological and environmental factors which present high inter-patient variability. The objective of this study is to evaluate a number of features, which are extracted from medical and lifestyle self-monitoring data, with respect to their ability to predict the short-term subcutaneous (s.c.) glucose concentration of an individual. Random forests (RF) and RReliefF algorithms are first employed to rank the candidate feature set. Then, a forward selection procedure follows to build a glucose predictive model, where features are sequentially added to it in decreasing order of importance. Predictions are performed using support vector regression or Gaussian processes. The proposed method is validated on a dataset of 15 type diabetics in real-life conditions. The s.c. glucose profile along with time of the day and plasma insulin concentration are systematically highly ranked, while the effect of food intake and physical activity varies considerably among patients. Moreover, the average prediction error converges in less than d/2 iterations (d is the number of features). Our results suggest that RF and RReliefF can find the most informative features and can be successfully used to customize the input of glucose models.

Assessing sensor accuracy for non-adjunct use of continuous glucose monitoring

BACKGROUND

The level of continuous glucose monitoring (CGM) accuracy needed for insulin dosing using sensor values (i.e., the level of accuracy permitting non-adjunct CGM use) is a topic of ongoing debate. Assessment of this level in clinical experiments is virtually impossible because the magnitude of CGM errors cannot be manipulated and related prospectively to clinical outcomes.

MATERIALS AND METHODS

A combination of archival data (parallel CGM, insulin pump, self-monitoring of blood glucose [SMBG] records, and meals for 56 pump users with type 1 diabetes) and in silico experiments was used to "replay" real-life treatment scenarios and relate sensor error to glycemic outcomes. Nominal blood glucose (BG) traces were extracted using a mathematical model, yielding 2,082 BG segments each initiated by insulin bolus and confirmed by SMBG. These segments were replayed at seven sensor accuracy levels (mean absolute relative differences [MARDs] of 3-22%) testing six scenarios: insulin dosing using sensor values, threshold, and predictive alarms, each without or with considering CGM trend arrows.

RESULTS

In all six scenarios, the occurrence of hypoglycemia (frequency of BG levels ≤50 mg/dL and BG levels ≤39 mg/dL) increased with sensor error, displaying an abrupt slope change at MARD =10%. Similarly, hyperglycemia (frequency of BG levels ≥250 mg/dL and BG levels ≥400 mg/dL) increased and displayed an abrupt slope change at MARD=10%. When added to insulin dosing decisions, information from CGM trend arrows, threshold, and predictive alarms resulted in improvement in average glycemia by 1.86, 8.17, and 8.88 mg/dL, respectively.

CONCLUSIONS

Using CGM for insulin dosing decisions is feasible below a certain level of sensor error, estimated in silico at MARD=10%. In our experiments, further accuracy improvement did not contribute substantively to better glycemic outcomes.

Predicting Plasma Glucose From Interstitial Glucose Observations Using Bayesian Methods

One way of constructing a control algorithm for an artificial pancreas is to identify a model capable of predicting plasma glucose (PG) from interstitial glucose (IG) observations. Stochastic differential equations (SDEs) make it possible to account both for the unknown influence of the continuous glucose monitor (CGM) and for unknown physiological influences. Combined with prior knowledge about the measurement devices, this approach can be used to obtain a robust predictive model. A stochastic-differential-equation-based gray box (SDE-GB) model is formulated on the basis of an identifiable physiological model of the glucoregulatory system for type 1 diabetes mellitus (T1DM) patients. A Bayesian method is used to estimate robust parameters from clinical data. The models are then used to predict PG from IG observations from 2 separate study occasions on the same patient. First, all statistically significant diffusion terms of the model are identified using likelihood ratio tests, yielding inclusion of [Formula: see text], [Formula: see text], and [Formula: see text]. Second, estimates using maximum likelihood are obtained, but prediction capability is poor. Finally a Bayesian method is implemented. Using this method the identified models are able to predict PG using only IG observations. These predictions are assessed visually. We are also able to validate these estimates on a separate data set from the same patient. This study shows that SDE-GBs and a Bayesian method can be used to identify a reliable model for prediction of PG using IG observations obtained with a CGM. The model could eventually be used in an artificial pancreas.

A predictive model of subcutaneous glucose concentration in type 1 diabetes based on Random Forests

In this study, an individualized predictive model of the subcutaneous glucose concentration in type 1 diabetes is presented, which relies on the Random Forests regression technique. A multivariate dataset is utilized concerning the s.c. glucose profile, the plasma insulin concentration, the intestinal absorption of meal-derived glucose and the daily energy expenditure. In an attempt to capture daily rhythms in glucose metabolism, we also introduce a time feature in the predictive analysis. The dataset comes from the continuous multi-day recordings of 27 type 1 patients in free-living conditions. Evaluating the performance of the proposed method by 10-fold cross validation, an average RMSE of 6.60, 8.15, 9.25 and 10.83 mg/dl for 15, 30, 60 and 120 min prediction horizons, respectively, was attained.

A glucose model based on support vector regression for the prediction of hypoglycemic events under free-living conditions

BACKGROUND

The prevention of hypoglycemic events is of paramount importance in the daily management of insulin-treated diabetes. The use of short-term prediction algorithms of the subcutaneous (s.c.) glucose concentration may contribute significantly toward this direction. The literature suggests that, although the recent glucose profile is a prominent predictor of hypoglycemia, the overall patient's context greatly impacts its accurate estimation. The objective of this study is to evaluate the performance of a support vector for regression (SVR) s.c. glucose method on hypoglycemia prediction.

MATERIALS AND METHODS

We extend our SVR model to predict separately the nocturnal events during sleep and the non-nocturnal (i.e., diurnal) ones over 30-min and 60-min horizons using information on recent glucose profile, meals, insulin intake, and physical activities for a hypoglycemic threshold of 70 mg/dL. We also introduce herein additional variables accounting for recurrent nocturnal hypoglycemia due to antecedent hypoglycemia, exercise, and sleep. SVR predictions are compared with those from two other machine learning techniques.

RESULTS

The method is assessed on a dataset of 15 patients with type 1 diabetes under free-living conditions. Nocturnal hypoglycemic events are predicted with 94% sensitivity for both horizons and with time lags of 5.43 min and 4.57 min, respectively. As concerns the diurnal events, when physical activities are not considered, the sensitivity is 92% and 96% for a 30-min and 60-min horizon, respectively, with both time lags being less than 5 min. However, when such information is introduced, the diurnal sensitivity decreases by 8% and 3%, respectively. Both nocturnal and diurnal predictions show a high (>90%) precision.

CONCLUSIONS

Results suggest that hypoglycemia prediction using SVR can be accurate and performs better in most diurnal and nocturnal cases compared with other techniques. It is advised that the problem of hypoglycemia prediction should be handled differently for nocturnal and diurnal periods as regards input variables and interpretation of results.

On the use of continuous glucose monitoring systems to design optimal clinical tests for the identification of type 1 diabetes models

The identification of individual parameters of detailed physiological models of type 1 diabetes can be carried out by clinical tests designed optimally through model-based design of experiments (MBDoE) techniques. So far, MBDoE for diabetes models has been considered for discrete glucose measurement systems only. However, recent advances on sensor technology allowed for the development of continuous glucose monitoring systems (CGMSs), where glucose measurements can be collected with a frequency that is practically equivalent to continuous sampling. To specifically address the features of CGMSs, in this paper the optimal clinical test design problem is formulated and solved through a continuous, rather than discrete, approach. A simulated case study is used to assess the impact of CGMSs both in the optimal clinical test design problem and in the subsequent parameter estimation for the identification of a complex physiological model of glucose homeostasis. The results suggest that, although the optimal design of a clinical test is simpler if continuous glucose measurements are made available through a CGMS, the noise level and formulation may make continuous measurements less suitable for model identification than their discrete counterparts.

Diabetes technology: markers, monitoring, assessment, and control of blood glucose fluctuations in diabetes

People with diabetes face a life-long optimization problem: to maintain strict glycemic control without increasing their risk for hypoglycemia. Since the discovery of insulin in 1921, the external regulation of diabetes by engineering means has became a hallmark of this optimization. Diabetes technology has progressed remarkably over the past 50 years-a progress that includes the development of markers for diabetes control, sophisticated monitoring techniques, mathematical models, assessment procedures, and control algorithms. Continuous glucose monitoring (CGM) was introduced in 1999 and has evolved from means for retroactive review of blood glucose profiles to versatile reliable devices, which monitor the course of glucose fluctuations in real time and provide interactive feedback to the patient. Technology integrating CGM with insulin pumps is now available, opening the field for automated closed-loop control, known as the artificial pancreas. Following a number of in-clinic trials, the quest for a wearable ambulatory artificial pancreas is under way, with a first prototype tested in outpatient setting during the past year. This paper discusses key milestones of diabetes technology development, focusing on the progress in the past 10 years and on the artificial pancreas-still not a cure, but arguably the most promising treatment of diabetes to date.

Multivariate prediction of subcutaneous glucose concentration in type 1 diabetes patients based on support vector regression

Data-driven techniques have recently drawn significant interest in the predictive modeling of subcutaneous (s.c.) glucose concentration in type 1 diabetes. In this study, the s.c. glucose prediction is treated as a multivariate regression problem, which is addressed using support vector regression (SVR). The proposed method is based on variables concerning: (i) the s.c. glucose profile, (ii) the plasma insulin concentration, (iii) the appearance of meal-derived glucose in the systemic circulation, and (iv) the energy expenditure during physical activities. Six cases corresponding to different combinations of the aforementioned variables are used to investigate the influence of the input on the daily glucose prediction. The proposed method is evaluated using a dataset of 27 patients in free-living conditions. 10-fold cross validation is applied to each dataset individually to both optimize and test the SVR model. In the case where all the input variables are considered, the average prediction errors are 5.21, 6.03, 7.14 and 7.62 mg/dl for 15, 30, 60 and 120 min prediction horizons, respectively. The results clearly indicate that the availability of multivariable data and their effective combination can significantly increase the accuracy of both short-term and long-term predictions.

Real-time state estimation and long-term model adaptation: a two-sided approach toward personalized diagnosis of glucose and insulin levels

BACKGROUND

With continuous glucose sensors (CGSs), it is possible to obtain a dynamical signal of the patient's subcutaneous glucose concentration in real time. How could that information be exploited? We suggest a model-based diagnosis system with a twofold objective: real-time state estimation and long-term model parameter identification.

METHODS

To obtain a dynamical model, Bergman's nonlinear minimal model (considering plasma glucose G, insulin I, and interstitial insulin X) is extended by two states describing first and second insulin response. Furthermore, compartments for oral glucose and subcutaneous insulin inputs as well as for subcutaneous glucose measurement are added. The observability of states and external inputs as well as the identifiability of model parameters are assessed using the empirical observability Gramian. Signals are estimated for different nondiabetic and diabetic scenarios by unscented Kalman filter.

RESULTS

(1) Observability of different state subsets is evaluated, e.g., from CGSs, {G, I} or {G, X} can be observed and the set {G, I, X} cannot. (2) Model parameters are included, e.g., it is possible to estimate the second-phase insulin response gain k(G2) additionally. This can be used for model adaptation and as a diagnostic parameter that is almost zero for diabetes patients. (3) External inputs are considered, e.g., oral glucose is theoretically observable for nondiabetic patients, but estimation scenarios show that the time delay of 1 h limits application.

CONCLUSIONS

A real-time estimation of states (such as plasma insulin I) and parameters (such as k(G2)) is possible, which allows an improved real-time state prediction and a personalized model.

A meta-learning approach to the regularized learning-case study: blood glucose prediction

In this paper we present a new scheme of a kernel-based regularization learning algorithm, in which the kernel and the regularization parameter are adaptively chosen on the base of previous experience with similar learning tasks. The construction of such a scheme is motivated by the problem of prediction of the blood glucose levels of diabetic patients. We describe how the proposed scheme can be used for this problem and report the results of the tests with real clinical data as well as comparing them with existing literature.

Assessment of blood glucose predictors: the prediction-error grid analysis

BACKGROUND

Prediction of the future blood glucose (BG) evolution from continuous glucose monitoring (CGM) data is a promising direction in diabetes therapy management, and several glucose predictors have recently been proposed. This raises the problem of their assessment. There were attempts to use for such assessment the continuous glucose-error grid analysis (CG-EGA), originally developed for CGM devices. However, in the CG-EGA the BG rate of change is estimated from past BG readings, whereas predictors provide BG estimation ahead of time. Therefore, the original CG-EGA should be modified to assess predictors. Here we propose a new version of the CG-EGA, the Prediction-Error Grid Analysis (PRED-EGA).

METHODS

The analysis is based both on simulated data and on data from clinical trials, performed in the European FP7-project "DIAdvisor." Simulated data are used to test the ability of the analyzed CG-EGA modifications to capture erroneous predictions in controlled situation. Real data are used to show the impact of the different CG-EGA versions in the evaluation of a predictor.

RESULTS

Using the data of 10 virtual and 10 real subjects and analyzing two different predictors, we demonstrate that the straightforward application of the CG-EGA does not adequately classify the prediction performance. For example, we observed that up to 70% of 20 min ahead predictions in the hyperglycemia region that are classified by this application as erroneous are, in fact, accurate. Moreover, for predictions during hypoglycemia the assessments produced by the straightforward application of the CG-EGA are not only too pessimistic (in up to 60% of cases), but this version is not able to detect real erroneous predictions. In contrast, the proposed modification of the CG-EGA, where the rate of change is estimated on the predicted BG profile, is an adequate metric for the assessment of predictions.

CONCLUSIONS

We propose a new CG-EGA, the PRED-EGA, for the assessment of glucose predictors. The presented analysis shows that, compared with the straightforward application of the CG-EGA, the PRED-EGA gives a significant reduction of the misclassification cases. A reduction by a factor of at least 4 was observed in the study. Moreover, the PRED-EGA is much more robust against uncertainty in the input and references.

Glucose variability in diabetic pregnancy

BACKGROUND

Fetal overgrowth is the most important complication of gestational (GDM) and pregestational diabetes mellitus.

METHODS

We correlated maternal glucose profiles, as detected by continuous glucose monitoring (CGM), with fetal growth parameters for 80 pregnant women (32 with type 1 diabetes, 31 with GDM, and 17 healthy controls). Glucose profiles were monitored in the first, second, and third trimesters of pregnancy for type 1 diabetes women and in the second and third trimesters for GDM women and controls. To analyze glycemic variability, we considered the mean amplitude of glycemic excursion, mean glycemia, the continuous overlapping net glycemic action (CONGA), the SD, the High Blood Glucose Index (HBGI), the Low Blood Glucose Index, and the interquartile range (IQR).

RESULTS

Mean age was the same for the three groups. Prepregnancy body mass index was higher for the women with diabetes (GDM and type 1) than for controls. The newborn's mean birth weight and ponderal index were higher, although not significantly so, for the women with diabetes than for controls. For the type 1 diabetes patients, ponderal index correlated with the HBGI in the first trimester, CONGA1 and IQR in the second, and mean glycemia and SD in the third. For GDM patients, ponderal index correlated with mean glycemia and the HBGI in the second trimester.

CONCLUSIONS

Fetal exposure to glycemic variability and hyperglycemia seems to be important in determining fetal overgrowth in pregnant women with diabetes. Optimal glucose control and less glucose variability are needed as early as possible in both type 1 diabetes and GDM patients to ensure normal fetal growth.

FreeStyle Navigator Continuous Glucose Monitoring System with TRUstart algorithm, a 1-hour warm-up time

BACKGROUND

The first-generation FreeStyle Navigator® Continuous Glucose Monitoring System (FreeStyle Navigator CGM) requires a 10 h warm-up period to avoid inaccurate glucose readings caused by sensor insertion trauma and wound-healing processes. The performance of a second-generation FreeStyle Navigator CGM that begins reporting glucose 1 h after sensor insertion is described.

METHODS

Second-generation FreeStyle Navigator CGM performance was evaluated in an in-clinic study using the YSI Model 2300 STATPlus Glucose Analyzer as reference with 47 subjects with type 1 diabetes. The reference readings were taken at 15 min intervals, and the study was designed to emphasize the first 10 h of use.

RESULTS

The second-generation FreeStyle Navigator CGM exhibited continuous glucose error grid analysis ratings of 93.7% "clinically accurate," 3.6% "benign errors," and 2.8% "clinical errors" and a mean and median absolute relative difference of 14.5% and 10.7%, respectively. The second-generation algorithm detected signal instability in the first 10 h of use and suspended the reporting of 14.1% of first day continuous glucose readings. The clinical accuracy of the second-generation FreeStyle Navigator CGM was similar for the first 10 h versus subsequent hours, with 92.6% and 94.2% "clinically accurate" readings, respectively.

CONCLUSION

The warm-up period for the second-generation FreeStyle Navigator CGM was reduced from 10 to 1 h, with minimal interruption of glucose reporting and without sacrificing clinical performance.

Continuous glucose monitoring in subcutaneous tissue using factory-calibrated sensors: a pilot study

BACKGROUND

Commercial continuous subcutaneous glucose monitors require in vivo calibration using capillary blood glucose tests. Feasibility of factory calibration, i.e., sensor batch characterization in vitro with no further need for in vivo calibration, requires a predictable and stable in vivo sensor sensitivity and limited inter- and intra-subject variation of the ratio of interstitial to blood glucose concentration.

METHODS

Twelve volunteers wore two FreeStyle Navigator (Abbott Diabetes Care, Alameda, CA) continuous glucose monitoring systems for 5 days in parallel for two consecutive sensor wears (four sensors per subject, 48 sensors total). Sensors from a prototype sensor lot with a low variability in glucose sensitivity were used for the study. Median sensor sensitivity values based on capillary blood glucose were calculated per sensor and compared for inter- and intra-subject variation. Mean absolute relative difference (MARD) calculation and error grid analysis were performed using a single calibration factor for all sensors to simulate factory calibration and compared to standard fingerstick calibration.

RESULTS

Sensor sensitivity variation in vitro was 4.6%, which increased to 8.3% in vivo (P < 0.0001). Analysis of variance revealed no significant inter-subject differences in sensor sensitivity (P = 0.134). Applying a single universal calibration factor retrospectively to all sensors resulted in a MARD of 10.4% and 88.1% of values in Clarke Error Grid Zone A, compared to a MARD of 10.9% and 86% of values in Error Grid Zone A for fingerstick calibration.

CONCLUSIONS

Factory calibration of sensors for continuous subcutaneous glucose monitoring is feasible with similar accuracy to standard fingerstick calibration. Additional data are required to confirm this result in subjects with diabetes.

Essential elements of the native glucoregulatory system, which, if appreciated, may help improve the function of glucose controllers in the intensive care unit setting

In 2001, Van den Berghe and colleagues were able to show that tight glucose control decreases morbidity and mortality rates in the intensive care unit (ICU) setting. Several large, prospective, randomized controlled trials have failed to confirm these results. All of these studies attempted tight glucose control using expert-designed algorithms to adjust the rate of intravenous insulin. Unfortunately, these studies each had high rates of hypoglycemia, a high percentage of glucose values outside of the target range, and increased glucose variability. These three measurements have been shown to increase mortality rates in ICU patients. In order to achieve a high rate of success with regards to tight glucose control, a closed-loop system will need to be created. The two main elements of such a system are a continuous glucose sensor and a recursive glucose control algorithm. This review highlights the important elements of the native glucoregulatory system, which, if utilized, may help create a successful glucose control algorithm for a closed-loop system.

The accuracy of a new real-time continuous glucose monitoring algorithm: an analysis

In this issue of Journal of Diabetes Science and Technology, Keenan and colleagues used archival data from the STAR 1 clinical trial (Medtronic Diabetes) to support the claim that the new Veo calibration algorithm improves the accuracy of continuous glucose monitoring, particularly in the critical hypoglycemic range. Extensive data analyses are presented to support this claim; the results are convincing, and the estimated improvement in hypoglycemic detection from 55% for the standard calibration to 82% for the Veo is particularly impressive. We can therefore conclude that the Veo algorithm has the potential to improve the accuracy of hypoglycemia alarms and ultimately contribute to closed-loop control. However, the presented results should be interpreted cautiously because they are based on retrospective analysis and are heavily dependent on the distribution of blood glucose levels observed in a particular data set.

Hypoglycemia alarm enhancement using data fusion

BACKGROUND

The acceptance of closed-loop blood glucose (BG) control using continuous glucose monitoring systems (CGMS) is likely to improve with enhanced performance of their integral hypoglycemia alarms. This article presents an in silico analysis (based on clinical data) of a modeled CGMS alarm system with trained thresholds on type 1 diabetes mellitus (T1DM) patients that is augmented by sensor fusion from a prototype hypoglycemia alarm system (HypoMon). This prototype alarm system is based on largely independent autonomic nervous system (ANS) response features.

METHODS

Alarm performance was modeled using overnight BG profiles recorded previously on 98 T1DM volunteers. These data included the corresponding ANS response features detected by HypoMon (AiMedics Pty. Ltd.) systems. CGMS data and alarms were simulated by applying a probabilistic model to these overnight BG profiles. The probabilistic model developed used a mean response delay of 7.1 minutes, measurement error offsets on each sample of +/- standard deviation (SD) = 4.5 mg/dl (0.25 mmol/liter), and vertical shifts (calibration offsets) of +/- SD = 19.8 mg/dl (1.1 mmol/liter). Modeling produced 90 to 100 simulated measurements per patient. Alarm systems for all analyses were optimized on a training set of 46 patients and evaluated on the test set of 56 patients. The split between the sets was based on enrollment dates. Optimization was based on detection accuracy but not time to detection for these analyses. The contribution of this form of data fusion to hypoglycemia alarm performance was evaluated by comparing the performance of the trained CGMS and fused data algorithms on the test set under the same evaluation conditions.

RESULTS

The simulated addition of HypoMon data produced an improvement in CGMS hypoglycemia alarm performance of 10% at equal specificity. Sensitivity improved from 87% (CGMS as stand-alone measurement) to 97% for the enhanced alarm system. Specificity was maintained constant at 85%. Positive predictive values on the test set improved from 61 to 66% with negative predictive values improving from 96 to 99%. These enhancements were stable within sensitivity analyses. Sensitivity analyses also suggested larger performance increases at lower CGMS alarm performance levels.

CONCLUSION

Autonomic nervous system response features provide complementary information suitable for fusion with CGMS data to enhance nocturnal hypoglycemia alarms.

New and improved methods to characterize glycemic variability using continuous glucose monitoring

BACKGROUND

Glycemic variability is a possible risk factor for development of complications from diabetes. Numerous methods have been used to characterize glycemic variability.

METHODS

We propose several new methods to characterize glycemic variability. We evaluated these methods empirically and theoretically and compared them with previous methods.

RESULTS

We describe (1) extension and generalization of the mean amplitude of glycemic excursion (MAGE), i.e., "within day variability," (2) extension and generalization of the mean of daily differences (MODD), i.e., the "between day-within time points variability," (3) "between daily means variability," (4) "between time points variability" of the glucose profile averaged over several days, (5) "within series variability" for a time segment of any arbitrary length, (6) new measures of the stability of the daily glycemic patterns, (7) new types of graphical displays, including within day variability, between day-within time points variability, and between daily means variability versus total variability, and between daily means variability versus within day variability, and (8) new methods to evaluate whether within series and between day-within time points variability fluctuate systematically by time of day. We examined the new measures in relation to previous measures of glycemic variability using correlation analysis on a clinical dataset for 85 subjects. MAGE, MODD, and continuous overall net glycemic action (CONGA(n)) are directly proportional to total standard deviation (SD). MAGE is highly correlated with both total SD and within day variability but weakly correlated with measures of between day variability. MODD is highly correlated with between day-within time points variability and total SD but weakly correlated with measures of within day variability.

CONCLUSIONS

We provide a systematic, logical framework to characterize multiple aspects of glycemic variability and have implemented a simple, practical computing format. This approach can help clinical researchers and clinicians identify the major sources of variability for any given patient and monitor responses to interventions.

Graphical and numerical evaluation of continuous glucose sensing time lag

AIM

This study introduces a new method for graphical and numerical evaluation of time lags typically associated with subcutaneous glucose sensing, based on Poincaré-type plot and a maximum statistical agreement criterion.

METHODS

The proposed method is illustrated by retrospective analysis of 56 continuous glucose monitor (CGM) time series collected by the FreeStyle Navigator (Abbott Diabetes Care, Alameda, CA) from 28 patients with type 1 diabetes mellitus, each wearing simultaneously two sensors (on arm and abdomen) and parallel reference blood glucose (BG) collected with a reference YSI (Yellow Springs, OH) analyzer every 15 min. The average duration of a time series was 111 h; there were approximately 10,000 sensor-reference data pairs.

RESULTS

When sliding in time CGM readings versus BG, the point of minimal spread of a Poincaré-type plot marks visually the time of CGM delay. The same point is numerically estimated by minimizing the distance between BG and CGM readings. The average observed time lag between reference BG and CGM was 12.5 min. Stratified by BG rate of change, the time lag was longer (16.8 min) when BG was falling, compared to steady or rising BG (11.7 min and 9.9 min, respectively) (P < 0.005). The time lags at the two sensor locations were not significantly different: 12.4 min on the arm, 12.6 min on the abdomen.

CONCLUSIONS

In this data set, substantial blood-to-sensor time delays were observed, possibly because of both blood-to-interstitial glucose transport and instrumental delay. Analysis of BG-CGM co-dynamics that is free from mathematical approximation of glucose fluctuations resulted in convenient visualization and numerical estimation of these delays.

Diabetes: Models, Signals, and Control

The control of diabetes is an interdisciplinary endeavor, which includes a significant biomedical engineering component, with traditions of success beginning in the early 1960s. It began with modeling of the insulin-glucose system, and progressed to large-scale in silico experiments, and automated closed-loop control (artificial pancreas). Here, we follow these engineering efforts through the last, almost 50 years. We begin with the now classic minimal modeling approach and discuss a number of subsequent models, which have recently resulted in the first in silico simulation model accepted as substitute to animal trials in the quest for optimal diabetes control. We then review metabolic monitoring, with a particular emphasis on the new continuous glucose sensors, on the analyses of their time-series signals, and on the opportunities that they present for automation of diabetes control. Finally, we review control strategies that have been successfully employed in vivo or in silico, presenting a promise for the development of a future artificial pancreas and, in particular, discuss a modular architecture for building closed-loop control systems, including insulin delivery and patient safety supervision layers. We conclude with a brief discussion of the unique interactions between human physiology, behavioral events, engineering modeling and control relevant to diabetes.