Towards a personalised virtual diabetic patient simulator

The development of a diabetes simulator, an educational software tool which can help diabetic patients to better manage their disease, is described. Education of patients with diabetes mellitus is a fundamental part of diabetes care. One of the goals of diabetes education is to support the patients to understand the nature of their illness and its treatment. Diabetes patients should be aware of the impact of nutrition, physical activity and insulin injections on blood glucose excursions. They should be able to identify emerging health problems like severe hyperglycemia or hypoglycemia in early stages. This will enable them to adapt to and prevent these conditions. Personalised ICT services provide several possibilities to improve patient education and therefore diabetic care. In particular the development of an educational diabetes simulator is considered. Only one diabetes simulator, called AIDA, is currently available. AIDA is designed only for type-1 diabetes patients (5-10% of the whole diabetic population), and does not use patient-specific data. We are developing the Virtual Diabetic Patient Simulator (VDPS) for both type-1 and type-2 diabetes. The VDPS represents the key characteristics and behaviour of the main processes of glucose metabolism involved in diabetes for a period of one day (24 hours). Based on patient-specific information (parameters) and daily life situations the VDPS can be used to train the patient in a virtual environment with realistic scenarios. The core of the simulator is a computational model which calculates temporal glucose and insulin plasma profiles in response to known inputs (food, insulin, etc.). As a first step we have coupled and modified existing literature models. The current model can predict plasma glucose concentrations with acceptable accuracy for education purposes. In conclusion, the model provides a solid background for further developments of the diabetes simulator.