Page 143 - Glucose Monitoring Devices

144 CHAPTER 7 Clinical impact of CGM use

many users choosing not to engage as a conscious or subconscious avoidance
strategy [64]. Challenges for users include changing sensors, wirelessly pairing
transmitters, performing steady-state calibration, and analyzing large blood
glucose datasets, all of which may reinforce avoidant behavior. Conversely, will-
ingness to engage may not suffice in individuals where other factors such as
impaired visual acuity, reduced manual dexterity or inability to overcome the
learning curve would deem them unsuitable for CGM. The addition of nonnumer-
ical variables such as glucose trend arrows has ambiguous outcomes with regard to
the extent glucose levels will be affected. This, in turn, can introduce an element of
uncertainty when the user is making insulin dose decisions.

Healthcare provider dependent
CGM application remains relatively niche among the general diabetes population
and most physicians have limited experience interpreting CGM datasets. A standard
guideline or reference algorithm does not exist to assist healthcare professionals
when making CGM-driven treatment interventions resulting in varied proposed
interventions. A prospective observational study saw 2 days of CGM data from
20 pregnant women with T1DM presented to four physicians to give daily treatment
adjustment recommendations. Significant differences were observed when review-
ing the proposed interventions between the CGM days [65]. Although limited by
a small number of analyzing physicians, these results highlight the subjective
approach to CGM interpretation and call for a unified thinking process when
educating healthcare providers. The time requirements necessary to analyze CGM
data in a clinical setting pose a barrier in healthcare systems restricted by limited
trained professionals and consultation time constraints. Healthcare providers may
deem the additional time and resources required to educate healthcare professionals
to deliver CGM as economically nonviable and as a result limit CGM uptake.

Device dependent
CGMisdesignedtobeusedas anadjunctive tool alongside conventional SMBG
and not a direct replacement. Unfortunately, this is a common misconception
resulting in misplaced expectations among potential users. The application of
CGM for therapeutic decisions is assumed on interstitial glucose being inter-
changeable with traditional capillary blood glucose measurements. Although
glucose levels across both compartments are established by a process of diffusion,
CGM calibration against corresponding steady-state capillary glucose is recom-
mended to ensure accuracy. MARD (mean absolute relative difference) is the
most used metric for assessing CGM sensor accuracy with smaller percentages
reflecting measurements closer to reference glucose values. A value 10% is
regarded as the accuracy threshold based on in silico simulation demonstrating
insignificant hypoglycemia with lower percentages [66]. As a measure of accuracy,
MARD is a variable depending on glucose concentrations and rate of change, thus