Statistical Modeling of Aggregated Lifestyle and Blood Glucose Data in Type 1 Diabetes Patients

Abstract

We describe how the mobile-phone based Few Touch application for self-management will be used by a significant number of patients with diabetes mellitus Type 1 to assemble a data set with blood glucose concentration time series and information about exercise, diet, and insulin medication. Currently work in progress, these data are intended to be utilized when building a statistical multi-scale model with the primary aim of making short-term forecasts of blood glucose concentration. The system can provide decision support regarding insulin injections in given situations in addition to describing and visualizing long-term variation and provide general information on the patient's medical situation.