A Graphical Model of Nocturnal Therapy of Juvenile Diabetics

• Published in: Procedia computer science 2023, Vol.225, p.188-197
• Main Authors: Deja, Rafal, Deja, Grazyna
• Format: Article
• Language: English
• Subjects: decision support systems; diabetes mellitus; prediction model
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2023.10.003

Diabetes is one of the most dangerous diseases humanity faces. What is surprising is how widespread diabetes is among children, even those very young. Many of those children are admitted to hospitals and underlay intensive diabetic therapy. The goal of that therapy is to normalize and stabilize the patient's blood glucose level (BGL). This should be done as soon as possible by proper adjustment of insulin injections. It is worth noting that the problem is especially tough while dealing with children, those admitted to a hospital for the first time. At diabetes onset, the individual reaction of the human body to insulin is not known, therefore, establishing insulin therapy must be done very carefully. In that period, the decision support system for physicians is much appreciated. The research presented in this paper addresses that problem. Specifically, to complement our previous works, we develop a graphical model representing the nocturnal diabetic therapy which, due to the patient rest state, is very specific and requires special attention from the physicians. The proposed model helps to avoid excessive fluctuations of BGL during nocturnal periods, namely hypoglycemia and hyperglycemia. The model has been validated by a diabetology expert. The results of that validation reveal that the proposed graph-based approach is a valuable tool assisting physicians in their clinical practice.