Mean absolute relative difference

Mean Absolute Relative Difference (MARD) is a standard metric used to evaluate the accuracy of continuous glucose monitoring systems, which gives the average amount a CGM sensor reading varies from the actual blood glucose. It is calculated by taking the average of the absolute relative differences between the glucose readings reported by the CGM system and corresponding reference measurements, typically obtained through laboratory analysis or blood glucose meters. A lower MARD value indicates greater accuracy, and it is commonly used in clinical research and regulatory evaluations to compare the performance of different CGM devices. It is also of note that MARD percentages can vary by person, even while using the same device.

MARD by device

Device MARD Release year Notes
Dexcom STS 20.3% 2006
Dexcom Seven 17% 2007
Dexcom Seven Plus 16% 2010
Dexcom G4 Platinum 13.9% 2012
Dexcom G5 Mobile 9% 2015
Dexcom G6 9% 2018
Dexcom G7 8.2% 2022
Dexcom G7 15-Day 8% 2024
FreeStyle Libre 1 13.7% 2014
FreeStyle Libre 2 9.3% 2020
FreeStyle Libre 3 7.8% 2022