Universal Glucose Models for Predicting Subcutaneous Glucose Concentration in Humans

Abstract

Abstract-This paper tests the hypothesis that a "universal," data-driven model can be developed based on glucose data from one diabetic subject, and subsequently applied to predict subcutaneous glucose concentrations of other subjects, even of those with different types of diabetes. We employed three separate studies, each utilizing a different continuous glucose monitoring (CGM) device, to verify the model's universality. Two out of the three studies involved subjects with type 1 diabetes and the other one with type 2 diabetes. We first filtered the subcutaneous glucose concentration data by imposing constraints on their rate of change. Then, using the filtered data, we developed data-driven autoregressive models of order 30, and used them to make short-term, 30-min-ahead glucose-concentration predictions. We used same-subject model predictions as a reference for comparisons against cross-subject and cross-study model predictions, which were evaluated using the root-mean-squared error (RMSE) and Clarke error grid analysis (EGA). We found that, for each studied subject, the average crosssubject and cross-study RMSEs of the predictions were small and indistinguishable from those obtained with the same-subject models. These observations were corroborated by EGA, where better than 99.0% of the paired sensor-predicted glucose concentrations lay in the clinically acceptable zone A. In addition, the predictive capability of the models was found not to be affected by diabetes type, subject age, CGM device, and interindividual differences. We conclude that it is feasible to develop universal glucose models that allow for clinical use of predictive algorithms and CGM devices for proactive therapy of diabetic patients.

A Recurrent Neural Network Approach for Predicting Glucose Concentration in Type-1 Diabetic Patients

“…The RNN needs some fine tuning when it is used with a specific patient. This is one of the differences between our model and some others such as [5]. Our RNN can accurately predict the glucose values for PH=30 minutes without time delay.…”

“…The data was smoothed using low pass filter of order 11 before using it in training and testing the neural networks. The use of smoothed version of the CGM data reduces the time lag between the predicted glucose and measured glucose values [5].…”

“…Their model suffers from time delay between the original and predicted glucose concentration. Gani et al, [5] developed a glucose prediction autoregressive (AR) model of order 30 to make short term, 30-minutes ahead prediction time, without time lag. Pappada et al, [6] used a dataset of different patients obtained by CGM to construct a neural network using NeuroSolutions® software to predict glucose concentration while time varying predictive window from 50-180 minutes is used.…”

A Novel Adaptive-Weighted-Average Framework for Blood Glucose Prediction

“…Gani et al 5,9 clinically evaluated subject-specific AR models with 30 model orders to improve BG management. The results of Zhao et al 29 show that the best order of the AR model is 7.…”

“…First-order AR can produce acceptable predictions, but it introduces a significant delay between predicted and measured values. A high-order AR model was further studied, 8,9 but the prediction performance was not satisfactory. Allam et al 12 proposed a radial-basis-function neural network model to predict subcutaneous glucose concentrations.…”