Clinical Assessment and Mathematical Modeling of the Accuracy of Continuous Glucose Sensors (CGS)

In the treatment of diabetes, CGS calibration, time lag, and random errors influence the clinical glucose control decisions based on CGS output. These inaccuracies are determined by three main factors: quality of calibration, physiology, and sensor engineering. Simulated re-calibration and a diffusion model of blood-to-interstitial glucose transport allow the separation of these sources of inaccuracy. The methods are illustrated by data for 39 subjects with Type 1 diabetes collected during hyperinsulinemic euglycemic/hypoglycemic clamp by Minimed CGMStrade (Medtronic, Northridge, CA). The continuous glucose error-grid analysis (CG-EGA) was used to evaluate sensor inaccuracy from a clinical point of view