The physiological processes of the intensive care unit's patients such as blood glucose regulation are disturbed, and blood glucose rises. Due to its negative effects it should be normalized, but this is not easy because of the frequently changing state of the patient. There are many clinical protocols, varying from hospital to hospital, which regulate blood sugar by giving insulin or maybe by reducing glucose intake. Recently a protocol was developed, called STAR (Stochastic Targetted Glucose Conrol), which is proved to be efficient in practice, computerized and based on the human metabolic model.
In this study we investigated the possible usage of this protocol in different, not original circumstances. One question was whether the simultaneous usage of STAR and the continuous glucose monitor (CGM) can increase the quality of blood glucose regulation and decrease nurses' workload. The other question was whether STAR is safe during liver transplantation. We answered these questions by “in silico” simulations. In the first case we used a simple error model for modeling CGM's measurement error, and tried to find the best blood glucose alarm settings in order to achieve the previously mentioned aims. We concluded that in the case of appropriate alarm settings, there is point in using CGM with STAR, as it reduces the occurrence of low blood sugar, and it can also reduce nurses' workload.
We used 30 liver transplanted patients’ data to answer the second question. To do “in silico” trials, we had to change the metabolic model used during simulations, because the original one could not follow the pathophysiological changes of the liver transplantation. After this correction we examined the efficiency and safety of the original STAR and the modified ones. The original one improved the blood glucose control quality, and we could not improve it any further with modifications. We concluded that STAR can be used during liver transplantation.