Blood Glucose Levels Can Be Predicted By Math Model, Scientists Say

Blood glucose levels can be predicted by a mathematical model with 90 percent accuracy, according to researchers from Penn State University.

Peter Molenaar, professor of Human Development, Family Studies and Psychology, confirmed on the Penn State University website that blood glucose levels of individuals with type 1 diabetes can be determined by a predictive model 30 minutes before significant changes in the blood glucose occur, giving ample time for the patient to prepare for preventive adjustments.

According to Molenaar, glucose monitors that are available on the market today can pose many problems when tracking the blood glucose levels of a person. These monitors usually track fluids underneath a patient’s skin, but because the blood glucose from these fluids trail blood glucose levels from the body between eight and fifteen minutes, glucose monitors may not actually present an up-to-date analysis.

Molenaar added that this is especially problematic when the person is sleeping. People with hypoglycemia may be alerted too late by the glucose monitor alarm, which can lead to a patient’s death.

There are many factors that influence the blood glucose level of a person. Variables such as insulin dose, meal intake, physical activities and even psychological states can contribute to how high or low a person’s blood glucose can be.

Because of these variables, which may cause great fluctuations depending on the person, Molenaar and his team created a mathematical model that can account for time-varying changes in glucose kinetics caused by these different factors.The model is an extended version of the Kalman filtering technique, which is an algorithm of a series of measures that produces statistically optimal estimates from streams of inaccurate input data.

The team tested their model with 30 virtual patients and 5 real patients of type 1 diabetes. The results is set to be published this week in the Journal of Diabetes Science and Technology. According to Qian Wang, professor of mechanical engineering and an associate researcher in the study:

“We learned that the dynamic dependencies of blood glucose on insulin dose and meal intake vary substantially in time within each patient and between patients. The high prediction fidelity of our model over 30-minute intervals allows for the execution of optimal control of fast-acting insulin dose in real time because the initiation of insulin action has a delay of less than 30 minutes.”

The researchers claim that their model outperforms the commonly used methods because the model parameters are estimated real time.

Various scientific bodies supported the research on blood glucose levels, including the National Institute of Health and National Science Foundation.

[Image from Ruin Raider via Flickr]