A strategy based on integer programming for optimal dosing and timing of preventive hypoglycemic treatments in type 1 diabetes management

BACKGROUND AND OBJECTIVES

One of the major problems related to type 1 diabetes (T1D) management is hypoglycemia, a condition characterized by low blood glucose levels and responsible for reduced quality of life and increased mortality. Fast-acting carbohydrates, also known as hypoglycemic treatments (HT), can counteract this event. In the literature, dosage and timing of HT are usually based on heuristic rules. In the present work, we propose an algorithm for mitigating hypoglycemia by suggesting preventive HT consumption, with dosages and timing determined by solving an optimization problem.

METHODS

By leveraging integer programming and linear inequality constraints, the algorithm can bind the amount of suggested carbohydrates to standardized quantities (i.e., those available in "off-the-shelf" HT) and the minimal distance between consecutive suggestions (to reduce the nuisance for patients).

RESULTS

The proposed method was tested in silico and compared with competitor algorithms using the UVa/Padova T1D simulator. At the cost of a slight increase of HT consumed per day, the proposed algorithm produces the lowest median and interquartile range of the time spent in hypoglycemia, with a statistically significant improvement over most competitor algorithms. Also, the average number of hypoglycemic events per day is reduced to 0 in median.

CONCLUSIONS

Thanks to its positive performances and reduced computational burden, the proposed algorithm could be a candidate tool for integration in a DSS aimed at improving T1D management.

Risk of hypoglycemia in type 1 diabetes management: An in-silico sensitivity analysis to assess and rank the quantitative impact of different behavioral factors

BACKGROUND AND OBJECTIVE

In type 1 diabetes (T1D), a quantitative evaluation of the impact on hypoglycemia of suboptimal therapeutic decision (e.g. incorrect estimation of the ingested carbohydrates, inaccurate insulin timing, etc) is unavailable. Clinical trials to measure sensitivity to patient actions would be expensive, exposed to confounding factors and risky for the participants. In this work, a T1D patient decision simulator (T1D-PDS), realistically reproducing blood glucose dynamics in a large virtual population, is used to perform extensive in-silico trials and the so-derived data employed to implement a sensitivity analysis that ranks different behavioral factors based on their impact on a clinically meaningful parameter, the time below range (TBR).

METHODS

Eleven behavioral factors impacting on hypoglycemia are considered. The T1D-PDS was used to perform multiple 2-week simulations involving 100 adults, by testing about 3500 different perturbations for nominal behavior. A local linear approximation of the function linking the TBR and the factors were computed to derive sensitivity indices (SIs), quantifying the impact of each factor on TBR variations.

RESULTS

The obtained ranking quantifies importance of factors w.r.t. the others. Factors apparently related to hypoglycemia were correctly placed on the top of the ranking, including systematic (SI=2.05%) and random (SI=1.35%) carb-counting error, hypotreatment dose (SI=-1.21%), insulin bolus time w.r.t. mealtime (SI=1.09%).

CONCLUSIONS

The obtained SIs allowed to rank behavioral factors based on their impact on TBR. The behavioral factors identified as most influential can be prioritized in patient training.

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

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

A Markov Model of Gap Occurrence in Continuous Glucose Monitoring Data for Realistic in Silico Clinical Trials

BACKGROUND AND OBJECTIVE

Continuous glucose monitoring (CGM) sensors measure interstitial glucose concentration every 1-5 min for days or weeks. New CGM-based diabetes therapies are often tested in in silico clinical trials (ISCTs) using diabetes simulators. Accurate models of CGM sensor inaccuracies and failures could help improve the realism of ISCTs. However, the modeling of CGM failures has not yet been fully addressed in the literature. This work aims to develop a mathematical model of CGM gaps, i.e., occasional portions of missing data generated by temporary sensor errors (e.g., excessive noise or artifacts).

METHODS

Two datasets containing CGM traces collected in 167 adults and 205 children, respectively, using the Dexcom G6 sensor (Dexcom Inc., San Diego, CA) were used. Four Markov models, of increasing complexity, were designed to describe three main characteristics: number of gaps for each sensor, gap distribution in the monitoring days, and gap duration. Each model was identified on a portion of each dataset (training set). The remaining portion of each dataset (real test set) was used to evaluate model performance through a Monte Carlo simulation approach. Each model was used to generate 100 simulated test sets with the same size as the real test set. The distributions of gap characteristics on the simulated test sets were compared with those observed on the real test set, using the two-sample Kolmogorov-Smirnov test and the Jensen-Shannon divergence.

RESULTS

A six-state Markov model, having two states to describe normal sensor operation and four states to describe gap occurrence, achieved the best results. For this model, the Kolmogorov-Smirnov test found no significant differences between the distribution of simulated and real gap characteristics. Moreover, this model obtained significantly lower Jensen-Shannon divergence values than the other models.

CONCLUSIONS

A Markov model describing CGM gaps was developed and validated on two real datasets. The model describes well the number of gaps for each sensor, the gap distribution over monitoring days, and the gap durations. Such a model can be integrated into existing diabetes simulators to realistically simulate CGM gaps in ISCTs and thus enable the development of more effective and robust diabetes management strategies.

Impact of Carbohydrate Counting Error on Glycemic Control in Open-Loop Management of Type 1 Diabetes: Quantitative Assessment Through an In Silico Trial

BACKGROUND

In the management of type 1 diabetes (T1D), systematic and random errors in carb-counting can have an adverse effect on glycemic control. In this study, we performed an in silico trial aiming at quantifying the impact of different levels of carb-counting error on glycemic control.

METHODS

The T1D patient decision simulator was used to simulate 7-day glycemic profiles of 100 adults using open-loop therapy. The simulation was repeated for different values of systematic and random carb-counting errors, generated with Gaussian distribution varying the error mean from -10% to +10% and standard deviation (SD) from 0% to 50%. The effect of the error was evaluated by computing the difference of time inside (∆TIR), above (∆TAR) and below (∆TBR) the target glycemic range (70-180mg/dl) compared to the reference case, that is, absence of error. Finally, 3 linear regression models were developed to mathematically describe how error mean and SD variations result in ∆TIR, ∆TAR, and ∆TBR changes.

RESULTS

Random errors globally deteriorate the glycemic control; systematic underestimations lead to, on average, up to 5.2% more TAR than the reference case, while systematic overestimation results in up to 0.8% more TBR. The different time in range metrics were linearly related with error mean and SD (R²>0.95), with slopes of <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>M</mml:mi><mml:mi>E</mml:mi><mml:mi>A</mml:mi><mml:mi>N</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>21</mml:mn></mml:mrow></mml:math>, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>S</mml:mi><mml:mi>D</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mo>-</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>07</mml:mn></mml:mrow></mml:math> for ∆TIR, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>M</mml:mi><mml:mi>E</mml:mi><mml:mi>A</mml:mi><mml:mi>N</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mo>-</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>25</mml:mn></mml:mrow></mml:math>, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>S</mml:mi><mml:mi>D</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mo>+</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>06</mml:mn></mml:mrow></mml:math> for ∆TAR, and <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>M</mml:mi><mml:mi>E</mml:mi><mml:mi>A</mml:mi><mml:mi>N</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>05</mml:mn></mml:mrow></mml:math>, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>S</mml:mi><mml:mi>D</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mo>+</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>01</mml:mn></mml:mrow></mml:math> for ∆TBR.

CONCLUSIONS

The quantification of carb-counting error impact performed in this work may be useful understanding causes of glycemic variability and the impact of possible therapy adjustments or behavior changes in different glucose metrics.

Minimal and Maximal Models to Quantitate Glucose Metabolism: Tools to Measure, to Simulate and to Run in Silico Clinical Trials

Several models have been proposed to describe the glucose system at whole-body, organ/tissue and cellular level, designed to measure non-accessible parameters (minimal models), to simulate system behavior and run in silico clinical trials (maximal models). Here, we will review the authors' work, by putting it into a concise historical background. We will discuss first the parametric portrait provided by the oral minimal models-building on the classical intravenous glucose tolerance test minimal models-to measure otherwise non-accessible key parameters like insulin sensitivity and beta-cell responsivity from a physiological oral test, the mixed meal or the oral glucose tolerance tests, and what can be gained by adding a tracer to the oral glucose dose. These models were used in various pathophysiological studies, which we will briefly review. A deeper understanding of insulin sensitivity can be gained by measuring insulin action in the skeletal muscle. This requires the use of isotopic tracers: both the classical multiple-tracer dilution and the positron emission tomography techniques are discussed, which quantitate the effect of insulin on the individual steps of glucose metabolism, that is, bidirectional transport plasma-interstitium, and phosphorylation. Finally, we will present a cellular model of insulin secretion that, using a multiscale modeling approach, highlights the relations between minimal model indices and subcellular secretory events. In terms of maximal models, we will move from a parametric to a flux portrait of the system by discussing the triple tracer meal protocol implemented with the tracer-to-tracee clamp technique. This allows to arrive at quasi-model independent measurement of glucose rate of appearance (Ra), endogenous glucose production (EGP), and glucose rate of disappearance (Rd). Both the fast absorbing simple carbs and the slow absorbing complex carbs are discussed. This rich data base has allowed us to build the UVA/Padova Type 1 diabetes and the Padova Type 2 diabetes large scale simulators. In particular, the UVA/Padova Type 1 simulator proved to be a very useful tool to safely and effectively test in silico closed-loop control algorithms for an artificial pancreas (AP). This was the first and unique simulator of the glucose system accepted by the U.S. Food and Drug Administration as a substitute to animal trials for in silico testing AP algorithms. Recent uses of the simulator have looked at glucose sensors for non-adjunctive use and new insulin molecules.

Generation of post-meal insulin correction boluses in type 1 diabetes simulation models for in-silico clinical trials: More realistic scenarios obtained using a decision tree approach

BACKGROUND AND OBJECTIVE

In type 1 diabetes (T1D) research, in-silico clinical trials (ISCTs) notably facilitate the design/testing of new therapies. Published simulation tools embed mathematical models of blood glucose (BG) and insulin dynamics, continuous glucose monitoring (CGM) sensors, and insulin treatments, but lack a realistic description of some aspects of patient lifestyle impacting on glucose control. Specifically, to effectively simulate insulin correction boluses, required to treat post-meal hyperglycemia (BG > 180 mg/dL), the timing of the bolus may be influenced by subjects' behavioral attitudes. In this work, we develop an easily interpretable model of the variability of correction bolus timing observed in real data, and embed it into a popular simulation tool for ISCTs.

METHODS

Using data collected in 196 adults with T1D monitored in free-living conditions, we trained a decision tree (DT) model to classify whether a correction bolus is injected in a future time window, based on predictors collected back in time, related to CGM data, previous insulin boluses and subject's characteristics. The performance was compared to that of a logistic regression classifier with LASSO regularization (LC), trained on the same dataset. After validation, the DT was embedded within a popular T1D simulation tool and an ISCT was performed to compare the simulated correction boluses against those observed in a subset of data not used for model training.

RESULTS

The DT provided better classification performance (accuracy: 0.792, sensitivity: 0.430, specificity: 0.878, precision: 0.455) than the LC and presented good interpretability. The most predictive features were related to CGM (and its temporal variations), time since the last insulin bolus, and time of the day. The correction boluses simulated by the DT, after implementation in the simulation tool, showed a good agreement with real-world data.

CONCLUSIONS

The DT developed in this work represents a simple set of rules to mimic the same timing of correction boluses observed on real data. The inclusion of the model in simulation tools allows investigators to perform ISCTs that more realistically represent the patient behavior in taking correction boluses and the post-prandial BG response. In the future, more complex models can be investigated.

Interval Safety Layer Coupled With an Impulsive MPC for Artificial Pancreas to Handle Intrapatient Variability

The aim of control strategies for artificial pancreas systems is to calculate the insulin doses required by a subject with type 1 diabetes to regulate blood glucose levels by reducing hyperglycemia and avoiding the induction of hypoglycemia. Several control formulations developed for this end involve a safety constraint given by the insulin on board (IOB) estimation. This constraint has the purpose of reducing hypoglycemic episodes caused by insulin stacking. However, intrapatient variability constantly changes the patient's response to insulin, and thus, an adaptive method is required to restrict the control action according to the current situation of the subject. In this work, the control action computed by an impulsive model predictive controller is modulated with a safety layer to satisfy an adaptive IOB constraint. This constraint is established with two main steps. First, upper and lower IOB bounds are generated with an interval model that accounts for parameter uncertainty, and thus, define the possible system responses. Second, the constraint is selected according to the current value of glycemia, an estimation of the plant-model mismatch, and their corresponding first and second time derivatives to anticipate the changes of both glucose levels and physiological variations. With this strategy satisfactory results were obtained in an adult cohort where random circadian variability and sensor noise were considered. A 92% time in normoglycemia was obtained, representing an increase of time in range compared to previous MPC strategies, and a reduction of time in hypoglycemia to 0% was achieved without dangerously increasing the time in hyperglycemia.

Data Gap Modeling in Continuous Glucose Monitoring Sensor Data

Continuous glucose monitoring (CGM) sensors are minimally-invasive sensors used in diabetes therapy to monitor interstitial glucose concentration. The measurements are collected almost continuously (e.g. every 5 min) and permit the detection of dangerous hypo/hyperglycemic episodes. Modeling the various error components affecting CGM sensors is very important (e.g., to generate realistic scenarios for developing and testing CGM-based applications in type 1 diabetes simulators). In this work we focus on data gaps, which are portions of missing data due to a disconnection or a temporary sensor error. A dataset of 167 adults monitored with the Dexcom (San Diego, CA) G6 sensor is considered. After the evaluation of some statistics (the number of gaps for each sensor, the gap distribution over the monitoring days and the data gap durations), we develop a two-state Markov model to describe such statistics about data gap occurrence. Statistics about data gaps are compared between real data and simulated data generated by the model with a Monte Carlo simulation. Results show that the model describes quite accurately the occurrence and the duration of data gaps observed in real data.

A New Decision Support System for Type 1 Diabetes Management

Type 1 diabetes (T1D) is a chronic life-threatening metabolic condition which needs to be accurately and continuously managed with care by multiple daily exogenous insulin injections, frequent blood glucose concentration monitoring, ad-hoc diet, and physical activity. In the last decades, new technologies, such as continuous glucose monitoring sensors, eased the burden for T1D patients and opened new therapy perspectives by fostering the development of decision support systems (DSS). A DSS for T1D should be able to provide patients with advice aimed at improving metabolic control and reducing the number of actions related to therapy handling. Major challenges are the vast intra-/inter-subject physiological variability and the many factors that impact glucose metabolism. The present work illustrates a new DSS for T1D management. The algorithmic core includes a module for optimal, personalized, insulin dose calculation and a module that triggers the assumption of rescue carbohydrates to avoid/mitigate impending hypoglycemic events. The algorithms are integrated within a prototype communication platform that comprises a mobile app, a real-time telemonitoring interface, and a cloud server to safely store patients' data. Tests made in silico show that the use of the new algorithms lead to metabolic control indices significantly better than those obtained by the standard care for T1D. The preliminary test of the prototype platform suggests that it is robust, performant, and well-accepted by both patients and clinicians. Future work will focus on the refinement of the communication platform and the design of a clinical trial to assess the system effectiveness in real-life conditions.Clinical Relevance- The presented DSS is a promising tool to facilitate T1D daily management and improve therapy efficacy.

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

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

Linear Model Identification for Personalized Prediction and Control in Diabetes

OBJECTIVE

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

METHODS

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

RESULTS

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

CONCLUSION

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

SIGNIFICANCE

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

Stacked LSTM based deep recurrent neural network with kalman smoothing for blood glucose prediction

BACKGROUND

Blood glucose (BG) management is crucial for type-1 diabetes patients resulting in the necessity of reliable artificial pancreas or insulin infusion systems. In recent years, deep learning techniques have been utilized for a more accurate BG level prediction system. However, continuous glucose monitoring (CGM) readings are susceptible to sensor errors. As a result, inaccurate CGM readings would affect BG prediction and make it unreliable, even if the most optimal machine learning model is used.

METHODS

In this work, we propose a novel approach to predicting blood glucose level with a stacked Long short-term memory (LSTM) based deep recurrent neural network (RNN) model considering sensor fault. We use the Kalman smoothing technique for the correction of the inaccurate CGM readings due to sensor error.

RESULTS

For the OhioT1DM (2018) dataset, containing eight weeks' data from six different patients, we achieve an average RMSE of 6.45 and 17.24 mg/dl for 30 min and 60 min of prediction horizon (PH), respectively.

CONCLUSIONS

To the best of our knowledge, this is the leading average prediction accuracy for the ohioT1DM dataset. Different physiological information, e.g., Kalman smoothed CGM data, carbohydrates from the meal, bolus insulin, and cumulative step counts in a fixed time interval, are crafted to represent meaningful features used as input to the model. The goal of our approach is to lower the difference between the predicted CGM values and the fingerstick blood glucose readings-the ground truth. Our results indicate that the proposed approach is feasible for more reliable BG forecasting that might improve the performance of the artificial pancreas and insulin infusion system for T1D diabetes management.

Mathematical Models of Meal Amount and Timing Variability With Implementation in the Type-1 Diabetes Patient Decision Simulator

BACKGROUND

In type 1 diabetes (T1D) research, in-silico clinical trials (ISCTs) have proven effective in accelerating the development of new therapies. However, published simulators lack a realistic description of some aspects of patient lifestyle which can remarkably affect glucose control. In this paper, we develop a mathematical description of meal carbohydrates (CHO) amount and timing, with the aim to improve the meal generation module in the T1D Patient Decision Simulator (T1D-PDS) published in Vettoretti et al.

METHODS

Data of 32 T1D subjects under free-living conditions for 4874 days were used. Univariate probability density function (PDF) parametric models with different candidate shapes were fitted, individually, against sample distributions of: CHO amounts of breakfast (CHOB), lunch (CHOL), dinner (CHOD), and snack (CHOS); breakfast timing (TB); and time between breakfast-lunch (TBL) and between lunch-dinner (TLD). Furthermore, a support vector machine (SVM) classifier was developed to predict the occurrence of a snack in future fixed-length time windows. Once embedded inside the T1D-PDS, an ISCT was performed.

RESULTS

Resulting PDF models were: gamma (CHOB, CHOS), lognormal (CHOL, TB), loglogistic (CHOD), and generalized-extreme-values (TBL, TLD). The SVM showed a classification accuracy of 0.8 over the test set. The distributions of simulated meal data were not statistically different from the distributions of the real data used to develop the models (α = 0.05).

CONCLUSIONS

The models of meal amount and timing variability developed are suitable for describing real data. Their inclusion in modules that describe patient behavior in the T1D-PDS can permit investigators to perform more realistic, reliable, and insightful ISCTs.

Performance Analysis of Different Embedded Systems and Open-Source Optimization Packages Towards an Impulsive MPC Artificial Pancreas

Current technological advances have brought closer to reality the project of a safe, portable, and efficient artificial pancreas for people with type 1 diabetes (T1D). Among the developed control strategies for T1D, model predictive control (MPC) has been emphasized in literature as a promising control for glucose regulation. However, these control strategies are commonly designed in a computer environment, regardless of the limitations of a portable device. In this paper, the performances of six embedded platforms and three open-source optimization solver algorithms are assessed for T1D treatment. Their advantages and limitations are clarified using four MPC formulations of increasing complexity and a hardware-in-the-loop methodology to evaluate glucose control in virtual adult subjects. The performance comparison includes the execution time, the difference concerning the evolution obtained in MATLAB, the processor temperature, energy consumption, time percentage in normoglycemia, and the number of hypo- and hyperglycemic events. Results show that Quadprog is the package that faithfully follows the results obtained with control strategies designed and tuned on a computer with the MATLAB software. In addition, the Raspberry Pi 3 and the Tinker Board S embedded systems present the appropriate characteristics to be implemented as portable devices in the artificial pancreas application according to the criteria set out in this work.

The Padova Type 2 Diabetes Simulator from Triple-Tracer Single-Meal Studies: In Silico Trials Also Possible in Rare but Not-So-Rare Individuals

Background:In silico trials in type 2 diabetes (T2D) would be useful for testing diabetes treatments and accelerating the development of new antidiabetic drugs. In this study, we present a T2D simulator able to reproduce the variability observed in a T2D population. The simulator also allows to safely experiment on virtual subjects with severe (and possibly rare) pathological conditions. Methods: A meal simulation model of glucose, insulin, and C-peptide systems, made of 15 differential equations and 39 parameters, has been identified using a system decomposition and forcing function Bayesian strategy on data of 51 T2D subjects undergoing a single triple-tracer mixed meal. One hundred T2D in silico subjects have been generated from the joint distribution of estimated model parameters. A case study is presented to illustrate the simulator use for testing a virtual drug (improving insulin action and secretion) in a subpopulation of rare, extremely impaired, T2D subjects. Results: The model well fitted T2D data and parameters were estimated with precision. Simulated plasma glucose, insulin, and C-peptide well matched the data (e.g., median [25^th-75^th percentile] glucose area under the curves of 6.9 [6.1-8.5] 10⁴ mg/dL·min in silico vs. 7.0 [5.6-8.2] 10⁴ mg/dL·min in vivo). The potential use of the simulator was shown in a case study, in which the (virtual) antidiabetic drug dose was optimized for very insulin-resistant T2D subjects. Conclusions: We have developed a T2D simulator that captures the behavior of T2D population during a meal, both in terms of average and intersubject variability. The simulator represents a cost-effective way to test new antidiabetic drugs, before moving to human trials.

Nonlinear Machine Learning Models for Insulin Bolus Estimation in Type 1 Diabetes Therapy

Type 1 diabetes (T1D) therapy requires multiple daily insulin injections to compensate the lack of endogenous insulin production due to β-cells destruction. An empirical standard formula (SF) is commonly used for such a task. Unfortunately, SF does not include information on glucose dynamics, e.g. the glucose rate-of-change (ROC) provided by continuous glucose monitoring (CGM) sensor. Hence, SF can sometimes lead to under/overestimations that can cause critical hypo/hyperglycemic episodes during/after the meal. Recently, to overcome this limitation, we proposed new linear regression models, integrating ROC information and personalized features. Despite the first encouraging results, the nonlinear nature of the problem calls for the application of nonlinear models. In this work, random forest (RF) and gradient boosting tree (GBT), nonlinear machine learning methodologies, were investigated. A dataset of 100 virtual subjects, opportunely divided into training and testing sets, was used. For each individual, a single-meal scenario with different meal conditions (preprandial ROC, BG and meal amounts) was simulated. The assessment was performed both in terms of accuracy in estimating the optimal bolus and glycemic control. Results were compared to the best performing linear model previously developed. The two tree-based models proposed lead to a statistically significant improvement of glycemic control compared to the linear approach, reducing the time spent in hypoglycemia (from 32.49% to 27.57-25.20% for RF and GBT, respectively). These results represent a preliminary step to prove that nonlinear machine learning techniques can improve the estimation of insulin bolus in T1D therapy. Particularly, RF and GBT were shown to outperform the previously linear models proposed.Clinical Relevance- Insulin bolus estimation with nonlinear machine learning techniques reduces the risk of adverse events in T1D therapy.

Modeling Carbohydrate Counting Error in Type 1 Diabetes Management

Background: The error in estimating meal carbohydrates (CHO) amount is a critical mistake committed by type 1 diabetes (T1D) subjects. The aim of this study is both to investigate which factors, related to meals and subjects, affect the CHO counting error most and to develop a mathematical model of CHO counting error embeddable in T1D patient decision simulators to conduct in silico clinical trials. Methods: A published dataset of 50 T1D adults is used, which includes a patient's CHO count of 692 meals, dietitian's estimates of meal composition (used as reference), and several potential explanatory factors. The CHO counting error is modeled by multiple linear regression, with stepwise variable selection starting from 10 candidate predictors, that is, education level, insulin treatment duration, age, body weight, meal type, CHO, lipid, energy, protein, and fiber content. Inclusion of quadratic and interaction terms is also evaluated. Results: Larger errors correspond to larger meals, and most of the large meals are underestimated. The linear model selects CHO (P < 0.00001), meal type (P < 0.00001), and body weight (P = 0.047), whereas its extended version embeds a quadratic term of CHO (P < 0.00001) and interaction terms of meal type with CHO (P = 0.0001) and fiber amount (P = 0.001). The extended model explains 34.9% of the CHO counting error variance. Comparison with the CHO counting error description previously used in the T1D patient decision simulator shows that the proposed models return more credible realizations. Conclusions: The most important predictors of CHO counting errors are CHO and meal type. The mathematical models proposed improve the description of patients' behavior in the T1D patient decision simulator.

In Silico Head-to-Head Comparison of Insulin Glargine 300 U/mL and Insulin Degludec 100 U/mL in Type 1 Diabetes

Background: Second-generation long-acting insulin glargine 300 U/mL (Gla-300) and degludec 100 U/mL (Deg-100) provide novel basal insulin therapies for the treatment of type 1 diabetes (T1D). Both offer a flatter pharmacokinetic (PK) profile than the previous generation of long-acting insulins, thus improving glycemic control while reducing hypoglycemic events. This work describes an in silico head-to-head comparison of the two basal insulins on 24-h glucose profiles and was used to guide the design of a clinical trial. Materials and Methods: The Universities of Virginia (UVA)/Padova T1D simulator describes the intra-/interday variability of glucose-insulin dynamics and thus provides a robust bench-test for assessing glucose control for basal insulin therapies. A PK model describing subcutaneous absorption of Deg-100, in addition to the one already available for Gla-300, has been developed based on T1D clinical data and incorporated into the simulator. One hundred in silico T1D subjects received a basal insulin dose (Gla-300 or Deg-100) for 12 weeks (8 weeks uptitration, 4 weeks stable dosing) by morning or evening administration in a basal/bolus regimen. The virtual patients were uptitrated to their individual doses with two different titration rules. Results: The last 2-week simulated continuous glucose monitoring data were used to calculate various outcome metrics for both basal insulin treatments, with primary outcome being the percent time in glucose target (70-140 mg/dL). The simulations show no statistically significant difference for Gla-300 versus Deg-100 in the main endpoints. Conclusions: This work suggests comparable glucose control using either Gla-300 or Deg-100 and was used to guide the design of a clinical trial intended to compare second-generation long-acting insulin analogues.

Altered directed functional connectivity of the right amygdala in depression: high-density EEG study

The cortico-striatal-pallidal-thalamic and limbic circuits are suggested to play a crucial role in the pathophysiology of depression. Stimulation of deep brain targets might improve symptoms in treatment-resistant depression. However, a better understanding of connectivity properties of deep brain structures potentially implicated in deep brain stimulation (DBS) treatment is needed. Using high-density EEG, we explored the directed functional connectivity at rest in 25 healthy subjects and 26 patients with moderate to severe depression within the bipolar affective disorder, depressive episode, and recurrent depressive disorder. We computed the Partial Directed Coherence on the source EEG signals focusing on the amygdala, anterior cingulate, putamen, pallidum, caudate, and thalamus. The global efficiency for the whole brain and the local efficiency, clustering coefficient, outflow, and strength for the selected structures were calculated. In the right amygdala, all the network metrics were significantly higher (p < 0.001) in patients than in controls. The global efficiency was significantly higher (p < 0.05) in patients than in controls, showed no correlation with status of depression, but decreased with increasing medication intake (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\bf{R}}}^{{\bf{2}}}{\boldsymbol{=}}{\bf{0.59}}\,{\bf{and}}\,{\bf{p}}{\boldsymbol{=}}{\bf{1.52}}{\bf{e}}{\boldsymbol{ \mbox{-} }}{\bf{05}}$$\end{document}R2=0.59andp=1.52e‐05). The amygdala seems to play an important role in neurobiology of depression. Practical treatment studies would be necessary to assess the amygdala as a potential future DBS target for treating depression.

Modeling the error of factory-calibrated continuous glucose monitoring sensors: application to Dexcom G6 sensor data

Minimally-invasive continuous glucose monitoring (CGM) sensors are used in diabetes therapy to monitor interstitial glucose (IG) concentration almost continuously (e.g. every 5 min) and detect/predict dangerous hypo/hyperglycemic episodes. When compared with frequent blood glucose (BG) concentration references, CGM measurements are unavoidably affected by error. Models of the CGM error can be important in several applications, e.g. for testing in simulation the safety and effectiveness of CGM-based artificial pancreas algorithms. In this work, we model the error of the Dexcom G6, a CGM sensor that recently entered the market and does not require in vivo calibrations. The dataset includes CGM and BG data collected in 11 subjects wearing two Dexcom G6 sensors in parallel. The model is derived applying a methodology to dissect and model 3 main CGM error components: BG-to-IG kinetics, calibration error and measurement noise. An aspect of novelty of the method is its capability of handling factory-calibrated CGM sensor data. Results of model identification show that the time-variability of sensor calibration error during the sensor lifetime (10 days) can be well represented by a regression model with time-variant parameters described by 2^nd-order polynomials in time.

In-silico Assessment of Preventive Hypotreatment Efficacy and Development of a Continuous Glucose Monitoring Based Algorithm to Prevent/Mitigate Hypoglycemia in Type 1 Diabetes

In Type 1 diabetes (T1D) standard treatment, the mitigation of hypoglycemia is achieved by the assumption of small amounts of carbohydrates (CHO), called hypotreatments (HTs), as soon as hypoglycemia is revealed. However, since CHO takes time to reach the blood stream, hypoglycemia cannot be totally avoided. Our purpose is to evaluate in-silico the effectiveness of preventive HTs and to propose a new real-time algorithm for the mitigation/avoidance of hypoglycemia, based on continuous glucose monitoring (CGM) sensor data. To such a purpose, the algorithm exploits the "dynamic risk" non linear-function that, by combining CGM value and trend, allows predicting the forthcoming hypoglycemic event. The algorithm is tested in an ideal noise-free environment on 100 virtual subjects (VSs) generated by the UVA/Padova T1D simulator and undergoing a single-meal experiment, with induced post-meal hypoglycemia. Compared to a reference HT rule, which suggest to assume HTs when hypoglycemia is detected, the algorithm reduces, on median [25^th - 75^th percentiles], both the time spent in hypoglycemia (from 36 [29 - 43] min to 10 [0 - 20] min) and the post-treatment rebound (from 136 [121 - 148] mg/dl to 114 [98 - 130] mg/dl). In conclusion, the proposed real-time algorithm efficiently generates preventive HTs that allow to almost totally avoid hypoglycemia. Future work will concern to modify the algorithm for detecting in advance the severity of the hypoglycemic episode -since performance are influenced on the hypoglycemic episode aggressiveness level- and to assess algorithm in a more challenging environment, including CGM measurement error.

Development of an Error Model for a Factory-Calibrated Continuous Glucose Monitoring Sensor with 10-Day Lifetime

Factory-calibrated continuous glucose monitoring (FC-CGM) sensors are new devices used in type 1 diabetes (T1D) therapy to measure the glucose concentration almost continuously for 10-14 days without requiring any in vivo calibration. Understanding and modelling CGM errors is important when designing new tools for T1D therapy. Available literature CGM error models are not suitable to describe the FC-CGM sensor error, since their domain of validity is limited to 12-h time windows, i.e., the time between two consecutive in vivo calibrations. The aim of this paper is to develop a model of the error of FC-CGM sensors. The dataset used contains 79 FC-CGM traces collected by the Dexcom G6 sensor. The model is designed to dissect the error into its three main components: effect of plasma-interstitium kinetics, calibration error, and random measurement noise. The main novelties are the model extension to cover the entire sensor lifetime and the use of a new single-step identification procedure. The final error model, which combines a first-order linear dynamic model to describe plasma-interstitium kinetics, a second-order polynomial model to describe calibration error, and an autoregressive model to describe measurement noise, proved to be suitable to describe FC-CGM sensor errors, in particular improving the estimation of the physiological time-delay.

A Real-Time Continuous Glucose Monitoring-Based Algorithm to Trigger Hypotreatments to Prevent/Mitigate Hypoglycemic Events

Background: The standard treatment for hypoglycemia recommended by the American Diabetes Association (ADA) suggests patients with diabetes to take small amounts of carbohydrates, the so-called hypotreatments (HTs), as soon as blood glucose concentration goes below 70 mg/dL. However, prevention, or at least mitigation, of hypoglycemic events could be achieved by triggering HTs ahead of time thanks to the use of the predictive capabilities of suitable real-time algorithms fed by continuous glucose monitoring (CGM) sensor data. Materials and Methods: The algorithm proposed in this article to trigger HTs for preventing forthcoming hypoglycemic events is based on the computation of the "dynamic risk", there is a nonlinear function combining current glycemia with its rate-of-change, both provided by CGM. A comparison of performance of the proposed algorithm against the ADA guidelines is made, in silico, on datasets of 100 virtual patients undergoing a single-meal experiment, with induced postmeal hypoglycemia, generated by the UVA/Padova type 1 diabetes simulator. Results: On noise-free CGM data, the proposed algorithm reduces the time spent in hypoglycemia, on median [25th-75th percentiles] from 36 [29-43] to 0 [0-11] min (P < 0.0001), with a concomitant decrease of the post-treatment rebound (PTR) in glucose concentration, on median [25th-75th percentiles] from 136 [121-148] to 121 [116-127] mg/dL (P < 0.0001). On noisy CGM data, there is still a reduction of both time spent in hypoglycemia from 41 [28-49] min to 25 [0-41] min (P < 0.0001) and PTR from 174 [146-189] mg/dL to 137 [123-151] mg/dL (P < 0.0001). Conclusions: The potentiality of the new algorithm in generating preventive HTs, which can allow significant reduction of hypoglycemia without concomitant increase of hyperglycemia, suggests its further development and test in silico, for example, simulating both insulin pump and multiple-daily-injection therapies.

Detection of Insulin Pump Malfunctioning to Improve Safety in Artificial Pancreas Using Unsupervised Algorithms

BACKGROUND

Recent development of automated closed-loop (CL) insulin delivery systems, the so-called artificial pancreas (AP), improved the quality of type 1 diabetes (T1D) therapy. As new technologies emerge, patients put increasing trust in their therapeutic devices; therefore, it becomes increasingly important to detect malfunctioning affecting such devices. In this work, we explore a new paradigm to detect insulin pump faults (IPFs) that use unsupervised anomaly detection.

METHODS

We generated CL data corrupted with IPFs using the latest version of the T1D Padova/UVA simulator. From the data, we extracted several features capable to describe the patient dynamics and making more apparent suspicious data portions. Then, a feature selection is performed to determine the optimal feature set. Finally, the performance of several popular unsupervised anomaly detection algorithms is analyzed and compared on the identified optimal feature set.

RESULTS

Using the identified optimal configuration, the best performance is obtained by the Histogram-Based Outlier Score (HBOS) algorithm, which detected 87% of the IPF with only 0.08 false positives per day on average. Isolation forest is the best algorithm that offers more conservative performances, detection of 85% of the faults but only 0.06 false positives per day on average.

CONCLUSION

Unsupervised anomaly detection algorithms can be used effectively to detect IPFs and improve the safety of the AP. Future studies will be dedicated to test the presented method inside dedicated clinical trials.

GluNet: A Deep Learning Framework for Accurate Glucose Forecasting

For people with Type 1 diabetes (T1D), forecasting of blood glucose (BG) can be used to effectively avoid hyperglycemia, hypoglycemia and associated complications. The latest continuous glucose monitoring (CGM) technology allows people to observe glucose in real-time. However, an accurate glucose forecast remains a challenge. In this work, we introduce GluNet, a framework that leverages on a personalized deep neural network to predict the probabilistic distribution of short-term (30-60 minutes) future CGM measurements for subjects with T1D based on their historical data including glucose measurements, meal information, insulin doses, and other factors. It adopts the latest deep learning techniques consisting of four components: data pre-processing, label transform/recover, multi-layers of dilated convolution neural network (CNN), and post-processing. The method is evaluated in-silico for both adult and adolescent subjects. The results show significant improvements over existing methods in the literature through a comprehensive comparison in terms of root mean square error (RMSE) ([Formula: see text] mg/dL) with short time lag ([Formula: see text] minutes) for prediction horizons (PH) = 30 mins (minutes), and RMSE ([Formula: see text] mg/dL) with time lag ([Formula: see text] mins) for PH = 60 mins for virtual adult subjects. In addition, GluNet is also tested on two clinical data sets. Results show that it achieves an RMSE ([Formula: see text] mg/dL) with time lag ([Formula: see text] mins) for PH = 30 mins and an RMSE ([Formula: see text] mg/dL) with time lag ([Formula: see text] mins) for PH = 60 mins. These are the best reported results for glucose forecasting when compared with other methods including the neural network for predicting glucose (NNPG), the support vector regression (SVR), the latent variable with exogenous input (LVX), and the auto regression with exogenous input (ARX) algorithm.

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

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

Combining continuous glucose monitoring and insulin pumps to automatically tune the basal insulin infusion in diabetes therapy: a review

For individuals affected by Type 1 diabetes (T1D), a chronic disease in which the pancreas does not produce any insulin, maintaining the blood glucose (BG) concentration as much as possible within the safety range (70–180 mg/dl) allows avoiding short- and long-term complications. The tuning of exogenous insulin infusion can be difficult, especially because of the inter- and intra-day variability of physiological and behavioral factors. Continuous glucose monitoring (CGM) sensors, which monitor glucose concentration in the subcutaneous tissue almost continuously, allowed improving the detection of critical hypo- and hyper-glycemic episodes. Moreover, their integration with insulin pumps for continuous subcutaneous insulin infusion allowed developing algorithms that automatically tune insulin dosing based on CGM measurements in order to mitigate the incidence of critical episodes. In this work, we aim at reviewing the literature on methods for CGM-based automatic attenuation or suspension of basal insulin with a focus on algorithms, their implementation in commercial devices and clinical evidence of their effectiveness and safety.

In Silico Assessment of Literature Insulin Bolus Calculation Methods Accounting for Glucose Rate of Change

BACKGROUND

The standard formula (SF) used in bolus calculators (BCs) determines meal insulin bolus using "static" measurement of blood glucose concentration (BG) obtained by self-monitoring of blood glucose (SMBG) fingerprick device. Some methods have been proposed to improve efficacy of SF using "dynamic" information provided by continuous glucose monitoring (CGM), and, in particular, glucose rate of change (ROC). This article compares, in silico and in an ideal framework limiting the exposition to possibly confounding factors (such as CGM noise), the performance of three popular techniques devised for such a scope, that is, the methods of Buckingham et al (BU), Scheiner (SC), and Pettus and Edelman (PE).

METHOD

Using the UVa/Padova Type 1 diabetes simulator we generated data of 100 virtual subjects in noise-free, single-meal scenarios having different preprandial BG and ROC values. Meal insulin bolus was computed using SF, BU, SC, and PE. Performance was assessed with the blood glucose risk index (BGRI) on the 9 hours after meal.

RESULTS

On average, BU, SC, and PE improve BGRI compared to SF. When BG is rapidly decreasing, PE obtains the best performance. In the other ROC scenarios, none of the considered methods prevails in all the preprandial BG conditions tested.

CONCLUSION

Our study showed that, at least in the considered ideal framework, none of the methods to correct SF according to ROC is globally better than the others. Critical analysis of the results also suggests that further investigations are needed to develop more effective formulas to account for ROC information in BCs.

Modeling of Diabetes and Its Clinical Impact

Understanding all aspects of diabetes treatment is hindered by the complexity of this chronic disease and its multifaceted complications and comorbidities, including social and financial impacts. In vivo studies as well as clinical trials provided invaluable information for unraveling not only metabolic processes but also risk estimations of, for example, complications. These approaches are often time- and cost-consuming and have frequently been supported by simulation models. Simulation models provide the opportunity to investigate diabetes treatment from additional viewpoints and with alternative objectives. This review presents selected models focusing either on metabolic processes or risk estimations and financial outcomes to provide a basic insight into this complex subject. It also discusses opportunities and challenges of modeling diabetes.

Prediction of Adverse Glycemic Events From Continuous Glucose Monitoring Signal

The most important objective of any diabetes therapy is to maintain the blood glucose concentration within the euglycemic range, avoiding or at least mitigating critical hypo/hyperglycemic episodes. Modern continuous glucose monitoring (CGM) devices bear the promise of providing the patients with an increased and timely awareness of glycemic conditions as these get dangerously near to hypo/hyperglycemia. The challenge is to detect, with reasonable advance, the patterns leading to risky situations, allowing the patient to make therapeutic decisions on the basis of future (predicted) glucose concentration levels. We underline that a technically sound performance comparison of the approaches proposed in recent years has yet to be done, thus it is unclear which one is preferred. The aim of this study is to fill this gap by carrying out a comparative analysis among the most common methods for glucose event prediction. Both regression and classification algorithms have been implemented and analyzed, including static and dynamic training approaches. The dataset consists of 89 CGM time series measured in diabetic subjects for 7 subsequent days. Performance metrics, specifically defined to assess and compare the event-prediction capabilities of the methods, have been introduced and analyzed. Our numerical results show that a static training approach exhibits better performance, in particular when regression methods are considered. However, classifiers show some improvement when trained for a specific event category, such as hyperglycemia, achieving performance comparable to the regressors, with the advantage of predicting the events sooner.

A Neural-Network-Based Approach to Personalize Insulin Bolus Calculation Using Continuous Glucose Monitoring

BACKGROUND

In type 1 diabetes (T1D) therapy, the calculation of the meal insulin bolus is performed according to a standard formula (SF) exploiting carbohydrate intake, carbohydrate-to-insulin ratio, correction factor, insulin on board, and target glucose. Recently, some approaches were proposed to account for preprandial glucose rate of change (ROC) in the SF, including those by Scheiner and by Pettus and Edelman. Here, the aim is to develop a new approach, based on neural networks (NN), to optimize and personalize the bolus calculation using continuous glucose monitoring information and some easily accessible patient parameters.

METHOD

The UVa/Padova T1D Simulator was used to simulate data of 100 virtual adults in a single-meal noise-free scenario with different conditions in terms of meal amount and preprandial blood glucose and ROC values. An NN was trained to learn the optimal insulin dose using the SF parameters, ROC, body weight, insulin pump basal infusion rate and insulin sensitivity as features. The performance of the NN for meal bolus calculation was assessed by blood glucose risk index (BGRI) and compared to the methods by Scheiner and by Pettus and Edelman.

RESULTS

The NN approach brings to a small but statistically significant ( P < .001) reduction of BGRI value, equal to 0.37, 0.23, and 0.20 versus SF, Scheiner, and Pettus and Edelman, respectively.

CONCLUSION

This preliminary study showed the potentiality of using NNs for the personalization and optimization of the meal insulin bolus calculation. Future work will deal with more realistic scenarios including technological and physiological/behavioral sources of variability.

The UVA/Padova Type 1 Diabetes Simulator Goes From Single Meal to Single Day

BACKGROUND

A new version of the UVA/Padova Type 1 Diabetes (T1D) Simulator is presented which provides a more realistic testing scenario. The upgrades to the previous simulator, which was accepted by the Food and Drug Administration in 2013, are described.

METHOD

Intraday variability of insulin sensitivity (S_I) has been modeled, based on clinical T1D data, accounting for both intra- and intersubject variability of daily S_I. Thus, time-varying distributions of both subject's basal insulin infusion and insulin-to-carbohydrate ratio were calculated and made available to the user. A model of "dawn" phenomenon based on clinical T1D data has been also included. Moreover, the model of subcutaneous insulin delivery has been updated with a recently developed model of commercially available fast-acting insulin analogs. Models of both intradermal and inhaled insulin pharmacokinetics have been included. Finally, new models of error affecting continuous glucose monitoring and self-monitoring of blood glucose devices have been added.

RESULTS

One hundred in silico adults, adolescent, and children have been generated according to the above modifications. The new simulator reproduces the intraday glucose variability observed in clinical data, also describing the nocturnal glucose increase, and the simulated insulin profiles reflect real life data.

CONCLUSIONS

The new modifications introduced in the T1D simulator allow to extend its domain of validity from "single-meal" to "single-day" scenarios, thus enabling a more realistic framework for in silico testing of advanced diabetes technologies including glucose sensors, new insulin molecules and artificial pancreas.

Next-generation, personalised, model-based critical care medicine: a state-of-the art review of in silico virtual patient models, methods, and cohorts, and how to validation them

Critical care, like many healthcare areas, is under a dual assault from significantly increasing demographic and economic pressures. Intensive care unit (ICU) patients are highly variable in response to treatment, and increasingly aging populations mean ICUs are under increasing demand and their cohorts are increasingly ill. Equally, patient expectations are growing, while the economic ability to deliver care to all is declining. Better, more productive care is thus the big challenge. One means to that end is personalised care designed to manage the significant inter- and intra-patient variability that makes the ICU patient difficult. Thus, moving from current "one size fits all" protocolised care to adaptive, model-based "one method fits all" personalised care could deliver the required step change in the quality, and simultaneously the productivity and cost, of care. Computer models of human physiology are a unique tool to personalise care, as they can couple clinical data with mathematical methods to create subject-specific models and virtual patients to design new, personalised and more optimal protocols, as well as to guide care in real-time. They rely on identifying time varying patient-specific parameters in the model that capture inter- and intra-patient variability, the difference between patients and the evolution of patient condition. Properly validated, virtual patients represent the real patients, and can be used in silico to test different protocols or interventions, or in real-time to guide care. Hence, the underlying models and methods create the foundation for next generation care, as well as a tool for safely and rapidly developing personalised treatment protocols over large virtual cohorts using virtual trials. This review examines the models and methods used to create virtual patients. Specifically, it presents the models types and structures used and the data required. It then covers how to validate the resulting virtual patients and trials, and how these virtual trials can help design and optimise clinical trial. Links between these models and higher order, more complex physiome models are also discussed. In each section, it explores the progress reported up to date, especially on core ICU therapies in glycemic, circulatory and mechanical ventilation management, where high cost and frequency of occurrence provide a significant opportunity for model-based methods to have measurable clinical and economic impact. The outcomes are readily generalised to other areas of medical care.

Personalized blood glucose prediction: A hybrid approach using grammatical evolution and physiological models

The large patient variability in human physiology and the effects of variables such as exercise or meals challenge current prediction modeling techniques. Physiological models are very precise but they are typically complex and specific physiological knowledge is required. In contrast, data-based models allow the incorporation of additional inputs and accurately capture the relationship between these inputs and the outcome, but at the cost of losing the physiological meaning of the model. In this work, we designed a hybrid approach comprising physiological models for insulin and grammatical evolution, taking into account the clinical harm caused by deviations from the target blood glucose by using a penalizing fitness function based on the Clarke error grid. The prediction models were built using data obtained over 14 days for 100 virtual patients generated by the UVA/Padova T1D simulator. Midterm blood glucose was predicted for the 100 virtual patients using personalized models and different scenarios. The results obtained were promising; an average of 98.31% of the predictions fell in zones A and B of the Clarke error grid. Midterm predictions using personalized models are feasible when the configuration of grammatical evolution explored in this study is used. The study of new alternative models is important to move forward in the development of alarm-and-control applications for the management of type 1 diabetes and the customization of the patient's treatments. The hybrid approach can be adapted to predict short-term blood glucose values to detect continuous glucose-monitoring sensor errors and to estimate blood glucose values when the continuous glucose-monitoring system fails to provide them.