Algorithm for offline recalibration of recorded CGM traces

Continuous glucose monitoring (CGM) systems are becoming increasingly popular for the management of type 1 diabetes mellitus (T1DM). Even though measurement performance has increased tremendously over the last couple of years, measured CGM values still have a significantly lower measurement accuracy than top-of-the-market blood glucose (BG) meters. Reasons for this difference in accuracy are the time delays of CGM recordings as compared to the BG dynamics, CGM misscalibrations and changes in sensor sensitivity over time. Against this background the current paper presents an algorithm for the offline recalibration of recorded CGM current signals based on regular BG point measurements by means of self-monitoring of blood glucose (SMBG). As a result an estimate of a continuous BG signal is computed together with an estimate for the change in sensor sensitivity over time and the CGM time delays. This algorithm is useful for analyzing characteristics of a new CGM sensor, but also to compute accurate numbers for the patients' glycemic control during a clinical trial.