Skip to main content
Annals of Intensive Care
Fig. 4 | Annals of Intensive Care

Fig. 4

From: Accuracy, reliability, feasibility and nurse acceptance of a subcutaneous continuous glucose management system in critically ill patients: a prospective clinical trial

Fig. 4

Confounding factors on MARD. a (left): Association between MARD and individual daily blood glucose variability shown in first and second standard deviation of reference glucose. First boxplot The CGM device shows acceptable accuracy* (MARD median 10.9 %) if the blood glucose variability is low (first standard deviation). Second boxplot Accuracy deteriorates (MARD median 24 %) during increased blood glucose variability (second standard deviation). b (right): Association between MARD and blood glucose ranges. Second boxplot The CGM device shows acceptable accuracy* (MARD median 8.8 %) in blood glucose ranges between 80 and 179 mg/dl. First and Third boxplots Accuracy deteriorates in the hypoglycemic range (MARD median 65.8 %) and during severe hyperglycemia (MARD median 16 %). *According to criteria specified within the consensus recommendations [20], MARD should be <14 %

Back to article page