Continuous glucose monitoring and hypo/hyperglycaemia prediction

Abstract We demonstrate that glucose can be predicted ahead in time by exploiting a mathematical model fitted against past values provided by continuous glucose monitoring (CGM). In particular, results, obtained in 13 diabetic volunteers monitored for 48h by the Glucoday, show that, even by using a simple model, the crossing of the hypoglycemic threshold can be predicted more than 20 min ahead in time, a margin sufficient to prevent the event by a sugar ingestion.