Design and implementation of tight glycemic control algorithms

OData support
Dr. Benyó Balázs István
Department of Control Engineering and Information Technology

Nowadays it is a difficult task in the field of intensive care unit to maintain metabolic balance of critically ill patients. Keeping the patients in the state of normal glycemic range can significantly reduce mortality and the number of other complications as well, according to clinical experience. However, implementing protocols to maintain this state (also known as TGC protocols) can be associated with many challenges. In recent years many models describing the blood sugar balance of patients in need of intensive treatment were made - and based on these models, many protocols capable of implementing TGC were prepared as well. From these protocols, only STAR (Stochastic TARget Protocol) managed to increase the patients' time spent in the normal glycemic range thus reduce the number of clinical complications.

My research was aimed at planning and implementing an application component for mobile devices which is capable of executing the algorithm used during the NICE-SUGAR Study. This component had to comply with the STAR (Stochastic Glycaemic Control) advisory system's latest version being developed at the Laboratory of Biomedical Engineering at the Department of Control Engineering and Information Technology. Before the planning phase I learned about the blood sugar control method used during the NICE-SUGAR Study, then I prepared the formal description for the commands implemented using natural programming language. I prepared flow diagrams for the parts of the algorithm, and solved the specification problems emerging during the designation of these diagrams. The result of the research was a controller-type application component, which could be used to prepare the clinical validation of the STAR protocol.

Another purpose of my work was to make and test an application component that is capable of implementing the TGC protocol and can be used in NICU.


Please sign in to download the files of this thesis.