A gaming approach to improving diabetes care

Giorgia Tani, Maxima Medical Center

Managing diabetes

Diabetes is undoubtedly one of the most challenging health problems in the 21st century, currently affecting more than 280 million people in the world. Diabetes is associated with impaired glucose homeostasis, which is influenced by insulin. Although diabetes cannot be cured, it can be kept under control with the right treatment. This includes carefully managing diet, exercising, injecting insulin and taking medications to keep the blood glucose level within the target range. The role of the patient in the treatment is thus central.

Objectives

For a good management of diabetes, the patients have to be aware of the factors which influence their metabolism. This is why we want to provide them with an educational diabetes simulator, a tool that, in response to external stimuli, predicts the glucose and insulin temporal profile in the blood, the two indicators of a correct therapy. The goal of the design project was to develop the core of the simulator: a physiology-based model of glucose metabolism in diabetes patients in response to food and insulin intake.

Design process

We first modeled a healthy-person and then modified it to get the diabetes-patient models:
type-1 and type-2. For each model we identified the main processes in production, usage and storage of glucose (Fig.1). The models were based on state-of-the-art literature models, coupled and modified according to our needs. The resulting model is a set of coupled differential equations characterized by fifteen parameters. For model identifiability, the number of unknown parameters was reduced to six by studying their sensitivity and correlations; the remaining parameters were estimated through optimization algorithms. To account for heterogeneity of
type-2 diabetes, we divided the patients in six classes and then estimated the parameters for each class separately.

Results

Three operative models were obtained: healthy-person, type-1 and type-2 diabetes patients. The simulations were in good agreement with clinical data (Fig.2) and fulfilled the expectations of the MMC diabetes team. These models provide a solid background for further development of the diabetes simulator. The project has now an industrial partner and the product will become a web-simulator for open access to diabetes patients.

Giorgia Tani MSc PDEng

Giorgia Tani obtained her MSc degree in biomedical engineering at Campus Biomedico University in Rome in 2007. Later she joined the DTI program working on a physiology-based model for diabetes at Máxima Medisch Centrum (MMC). She currently works at Phonak (Zurich) on digital signal processing for hearing aids.