Assessing the glucose time course in critically ill patients by a mathematical model

The idea of a close relationship between the tight glycemic control and the outcome of critically ill hospitalized patients (e.g., those in ICU therapy), has been consolidated with growing evidence in recent years. The results obtained by several international studies have motivated the clinical departments to pay an increasing attention to the management of hospitalized patients with hyperglycemia. Insulin therapy protocols, which nurses and doctors must follow, are nowadays routinely used to daily control glycemia. The optimization of these protocols, as well as their critical appraisal in comparison with what is proposed by the scientific community, can be conveniently performed relying on a mathematical model of the hospitalized patient behavior. This work aims to adapt an already published model and to evaluate it against the data of several hospitalized patients admitted to the Niguarda Ca' Granda hospital. The model structure and parameters are estimated on the basis of a real-life ICU data set and a re-estimation strategy to improve the performance of glycemia prediction in the ICU is proposed.