Offline and online data assimilation for real-time blood glucose forecasting in type 2 diabetes

We evaluate the benefits of combining different offline and online data assimilation methodologies to improve personalized blood glucose prediction with type 2 diabetes self-monitoring data. We collect self-monitoring data (nutritional reports and pre- and post-prandial glucose measurements) from 4 individuals with diabetes and 2 individuals without diabetes. We write online to refer to methods that update state and parameters sequentially as nutrition and glucose data are received, and offline to refer to methods that estimate parameters over a fixed data set, distributed over a time window containing multiple nutrition and glucose measurements. We fit a model of ultradian glucose dynamics to the first half of each data set using offline (MCMC and nonlinear optimization) and online (unscented Kalman filter and an unfiltered model---a dynamical model driven by nutrition data that does not update states) data assimilation methods. Model parameters estimated over the first half of the data are used within online forecasting methods to issue forecasts over the second half of each data set. Offline data assimilation methods provided consistent advantages in predictive performance and practical usability in 4 of 6 patient data sets compared to online data assimilation methods alone; yet 2 of 6 patients were best predicted with a strictly online approach. Interestingly, parameter estimates generated offline led to worse predictions when fed to a stochastic filter than when used in a simple, unfiltered model that incorporates new nutritional information, but does not update model states based on glucose measurements. The relative improvements seen from the unfiltered model, when carefully trained offline, exposes challenges in model sensitivity and filtering applications, but also opens possibilities for improved glucose forecasting and relaxed patient self-monitoring requirements.

[1]  Cynthia R. Marling,et al.  A Machine Learning Approach to Predicting Blood Glucose Levels for Diabetes Management , 2014, AAAI Workshop: Modern Artificial Intelligence for Health Analytics.

[2]  R. Bergman,et al.  Physiologic evaluation of factors controlling glucose tolerance in man: measurement of insulin sensitivity and beta-cell glucose sensitivity from the response to intravenous glucose. , 1981, The Journal of clinical investigation.

[3]  R. Bhushan Gopaluni,et al.  A comprehensive compartmental model of blood glucose regulation for healthy and type 2 diabetic subjects , 2015, Medical & Biological Engineering & Computing.

[4]  E. Mosekilde,et al.  Computer model for mechanisms underlying ultradian oscillations of insulin and glucose. , 1991, The American journal of physiology.

[5]  Paul Poirier,et al.  Development and validation of a predictive model of acute glucose response to exercise in individuals with type 2 diabetes , 2013, Diabetology & Metabolic Syndrome.

[6]  Bhushan Gopaluni,et al.  A Feedback Glucose Control Strategy for Type II Diabetes Mellitus , 2011, 2011 International Symposium on Advanced Control of Industrial Processes (ADCONIP).

[7]  Daniela Calvetti,et al.  Parameter estimation for stiff deterministic dynamical systems via ensemble Kalman filter , 2014 .

[8]  Ezra Kwok,et al.  Detection of Abnormalities in Type II Diabetic Patients Using Particle Filters , 2015 .

[9]  Jeffrey K. Uhlmann,et al.  Unscented filtering and nonlinear estimation , 2004, Proceedings of the IEEE.

[10]  Lena Mamykina,et al.  Personalized glucose forecasting for type 2 diabetes using data assimilation , 2017, PLoS Comput. Biol..

[11]  B H Ginsberg,et al.  A new consensus error grid to evaluate the clinical significance of inaccuracies in the measurement of blood glucose. , 2000, Diabetes care.

[12]  Aditya Tulsyan,et al.  Assessment of type II diabetes mellitus using irregularly sampled measurements with missing data , 2014, Bioprocess and Biosystems Engineering.

[13]  J. Sturis,et al.  Abnormalities in the ultradian oscillations of insulin secretion and glucose levels in Type 2 (non-insulin-dependent) diabetic patients , 1992, Diabetologia.

[14]  E. Segal,et al.  Personalized Nutrition by Prediction of Glycemic Responses , 2015, Cell.

[15]  Helen X. Gao,et al.  American Diabetes Association Standards of Medical Care in Diabetes 2017 , 2017, Journal of diabetes.

[16]  Transl. D.D. Ivanov,et al.  American Diabetes Association. Standards of Medical Care in Diabetes — 2017 , 2017 .

[17]  Suzanne L Robertson,et al.  Identifiability and estimation of multiple transmission pathways in cholera and waterborne disease. , 2013, Journal of theoretical biology.

[18]  Claudio Cobelli,et al.  Meal Simulation Model of the Glucose-Insulin System , 2007, IEEE Transactions on Biomedical Engineering.

[19]  David C Klonoff,et al.  Technical Aspects of the Parkes Error Grid , 2013, Journal of diabetes science and technology.

[20]  Raed Abu Zitar,et al.  Towards Neural Network Model for Insulin/Glucose in Diabetics-II , 2005, Informatica.

[21]  Nicolette Meshkat,et al.  An algorithm for finding globally identifiable parameter combinations of nonlinear ODE models using Gröbner Bases. , 2009, Mathematical biosciences.

[22]  Y. Khader,et al.  Factors associated with poor glycemic control among patients with type 2 diabetes. , 2010, Journal of diabetes and its complications.

[23]  Lena Mamykina,et al.  Data-driven health management: reasoning about personally generated data in diabetes with information technologies , 2016, J. Am. Medical Informatics Assoc..

[24]  Marisa C Eisenberg,et al.  Determining identifiable parameter combinations using subset profiling. , 2014, Mathematical biosciences.

[25]  Dimitrios I. Fotiadis,et al.  Glucose Prediction in Type 1 and Type 2 Diabetic Patients Using Data Driven Techniques , 2011 .

[26]  Derrick K. Rollins,et al.  Free-living inferential modeling of blood glucose level using only noninvasive inputs , 2010 .

[27]  Richard M Hoffman,et al.  A probabilistic model for predicting hypoglycemia in type 2 diabetes mellitus: The Diabetes Outcomes in Veterans Study (DOVES). , 2004, Archives of internal medicine.

[28]  Malinda Peeples,et al.  Hypoglycemia Prediction Using Machine Learning Models for Patients With Type 2 Diabetes , 2014, Journal of diabetes science and technology.

[29]  B. Paterson,et al.  Expert decision making in relation to unanticipated blood glucose levels. , 2000, Research in nursing & health.

[30]  Ali Cinar,et al.  Multiple-Input Subject-Specific Modeling of Plasma Glucose Concentration for Feedforward Control , 2014, Industrial & engineering chemistry research.

[31]  Derrick K. Rollins,et al.  An algorithm for optimally fitting a wiener model , 2011 .

[32]  John Joseph Valletta,et al.  Gaussian process modelling of blood glucose response to free-living physical activity data in people with type 1 diabetes , 2009, 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[33]  Georgios D. Mitsis,et al.  Data-driven Modeling for Diabetes , 2014 .

[34]  S. E. Ahmed,et al.  Markov Chain Monte Carlo: Stochastic Simulation for Bayesian Inference , 2008, Technometrics.