In-Silico Analysis of Stochastic Modelling of Human Blood Glucose Regulatory System

Stress-induced hyperglycaemia, an elevated blood glucose level is a common complication in the intensive care which can be compensated by glycaemic control. However, due to the inter- and intra- patient variability which is the single greatest challenge in glycaemic control of critically ill patents it could not be implemented safely until the development of the model based control protocols. Model based protocols generally rely on the identification and prediction of the patient state. This study compares a deterministic and a stochastic differential equation based calculation method for patient state identification using a clinical data set of sixty patients from three independent intensive care patient cohorts. The accuracy of the two methods is found to be equivalent in the cohort level comparison and only insignificant differences are recognised in the patient level analysis.