Proportional-integral control in the evaluation of glucose homeostasis and Type 2 Diabetes

Abstract

Mathematical modelling of blood sugar dynamics has been explored since the early 1960s. For most of these models, measurements of hormones such as insulin and glucagon are required to find parameter values that yield plausible glucose behaviour. These hormone measurements require repeated blood tests and limit the feasibility of applying the model to “real world” medical applications such as disease diagnosis. We apply a proportional-integral control model to human glucose homeostasis. The model can be fit using glucose data obtained exclusively from a continuous glucose monitor, and measurements of hormone concentrations do not need to be taken. The resulting parameters from model tuning can be used as an evaluation of the effectiveness of an individual’s blood glucose homeostasis and as a biomarker of diabetic status. This research has the potential to be a diagnostic or intervention tool in type 2 diabetes and may aid in the early prediction of prediabetes.