T1D glossary

MARD (mean absolute relative difference)

Also called CGM accuracy.

MARD is the most common number used to describe CGM accuracy. It is the average gap, in percent, between CGM values and matched reference values from a lab or meter. A smaller MARD means the CGM sits closer to the reference. MARD figures from different studies do not compare well; head-to-head studies are the fair test. (International consensus on CGM use, 2017)

Formula

MARD (%) = average of |CGM value − reference value| ÷ reference value × 100, over all matched pairs

(International consensus on CGM use, 2017)

Worked example

Three matched pairs: CGM 110 vs reference 100 (10%), CGM 90 vs 100 (10%), CGM 120 vs 100 (20%). MARD = (10 + 10 + 20) ÷ 3 = 13.3%.

Related terms

• Continuous glucose monitor
• Compression low
• Sensor warm-up

Sources, quoted

• International consensus on CGM use, 2017: “The mean absolute relative difference (MARD) is currently the most common metric used to assess the performance of CGM systems.”
• International consensus on CGM use, 2017: “MARD is the average of the absolute error between all CGM values and matched reference values.”
• International consensus on CGM use, 2017: “A small percentage indicates that the CGM readings are close to the reference glucose value, whereas a larger MARD percentage indicates greater discrepancies between the CGM and reference glucose values.”
• International consensus on CGM use, 2017: “Comparing MARD values from different clinical studies has several limitations, and thus head-to-head studies should be performed.”

Every quote above is checked against its source page before each update of this site; last checked 2026-10-06. The worked example is our own arithmetic. This page explains a term; it is not advice for a person's treatment.