An agent-based diabetic patient simulation

Abstract

This paper presents a new paradigm for modeling illness in the human population. In this work we propose the development of a patient model using a Mobile Software Agent. We concentrate on Diabetes Mellitus because of the prevalence of this disease and the reality that many citizens must learn to manage their disease through some simple guidelines on their diet, exercise and medication. This form of modeling illness has the potential to forecast outcomes for diabetic patients depending on their lifestyle. We further believe that the Patient Agent could be an effective tool in assisting patients to understand their prognosis if they are not meticulous in controlling their blood sugar and insulin levels. Additionally simulation results may be used to exercise physiological data collection and presentation systems. The Patient Agent is developed in accordance with the general parameters used in archetypal Diabetes medical tests. Conventional formulae have been applied to transform input variables such as Food, Exercise, and Medications, as well as other risk factors like Age, Ethnicity, and Gender, into output variables such as Blood Glucose and Blood Pressure. The time evolution of the Patient Agent is represented through the outputs which deteriorate over the long term period. © 2008 Springer-Verlag Berlin Heidelberg.