Design and Evaluation of a Third-Order Regression Model for Estimating Glucose Levels from ECG Signal Features

Diabetes mellitus, also known simply as diabetes, is a chronic disease that affects the absorption of glucose from food, increasing levels in the blood. This disease mainly affects adults (type 2 diabetes) but can also occur in children (type 1 or childhood diabetes), as well as in pregnant women (gestational diabetes). Diabetes is one of the main diseases with the highest incidence and high mortality rate worldwide. Diabetes is a disease for which there is no cure; however, continuous monitoring and control of blood glucose levels reduce the risk of associated complications, such as gastrointestinal issues, vision loss, limb amputations (e.g., diabetic foot), and damage to vital organs such as the heart and kidneys. This article presents the results of the statistical and clinical evaluation for the estimation of blood glucose levels based on features derived from the ECG signal. The primary features used as estimating variables were heart rate, heart rate variability, amplitudes of the R, S, T waves, and the QT interval. For the electrocardiogram recording, lead VII was used. The ANOVA analysis showed a result of 0. 0.7236.The clinical validation using the Clarke error grid showed a validation data percentage of 58.54% in zone A, 39.04% in zone B, 1.21% in zone C, and 1.21% in zone D, with a correlation of 0. 8525. Smart citations: https://scite.ai/reports/10.61467/2007.1558.2026.v17i2.1110Dimensions.Open Alex.