Improved glucose monitoring during exercise using wearable devices

Continuous measurement of blood glucose can improve the quality of life of people with type 1 diabetes. Current continuous glucose monitors (CGM) show a greater measurement error during aerobic exercise periods. The use of increasingly accessible physical activity monitoring devices opens up new possibilities to improve accuracy during these periods. The method and system suggested helps to avoid hypoglycemic situations due to a greater accuracy of glucose measurement, allowing better glucose control during exercise. Type 1 diabetes is a disease that is projected to affect 63 million people in 2045 with an estimated annual cost in healthcare spending of 77 million dollars. It is characterized by the inability of the cells of the pancreas to partially or totally produce insulin in the bloodstream, which causes abnormally high levels of glucose in the blood that can cause serious complications in patients. Continuous Glucose Monitoring (CGM) has established a turning point in the monitoring of diabetes, making it possible to know the blood glucose concentration of people with diabetes with great accuracy. However, it has been shown that CGM decreases its accuracy during periods of exercise, and may put the patient's health at risk in a period of great relevance, since periods of exercise are prone to hypoglycemia, which can cause severe complications. Continuous glucose monitors are sensors that estimate blood glucose concentration from current intensity measurements of an electrochemical sensor in the subcutaneous tissue, which are affected by the transport of glucose between interstitium and plasma, which during exercise is changed by hemodynamic factors, thus affecting its accuracy. Through the use of biometric signals provided by physical activity monitoring devices (wearables) a system that is capable of reversing the CGM estimation error during exercise has been designed. This system is also able to restore the accuracy of glucose measurements to a similar magnitude to that taking place outside periods of activity. In the proposed measurement method, metabolic equivalent signals (MET) or normalized measure of energy expenditure, and skin temperature, are used to derive a static regression between the estimation error and said signals, which is subsequently used make corrections. In other possible applications, the regression model can be dynamic or other biometric signals can be used (even dispense with the wearable if additional sensors can be integrated into the CGM transmitter, for example with systems based solely on skin temperature). Thus, the system can be integrated with the current CGM generation, only by implementing the sensors present in the physical activity monitoring devices integrated in a single device together with the CGM. These improvements can benefit all those with insulinized diabetes, regardless of the technology on which their therapy is based, and especially those who exercise regularly, with the reduction of possible risks due to loss of accuracy of the CGM, especially in types of exercise with a high probability of inducing hypoglycemia.