Page 61 - GHR_NursingCMU2024_Final.indd

Intermittently-scanned Continuous Glucose Monitoring to Motivate Diabetes
Self Care in Black Women with Type 2 Diabetes Who Do Not Use Insulin:
A Proof of Concept Study

1
C. Fritschi and C. Park 2

1 Department of Biobehavioral Nursing Science, University of Illinois, Chicago, College of Nursing, United
2
States, and Department of Population Health Nursing Science, University of Illinois, Chicago, College
of Nursing, United States

Background: Type 2 diabetes (T2DM) disproportionally affects low-income Black adults in the U.S.
They are less likely to have received diabetes education, have poorer glucose control, and seldom
monitor their blood glucose levels (BG) if they do not use insulin. Thus, they cannot connect their
dietary choices and physical activity (PA) to BG levels. Periodic use of intermittently scanned
continuous glucose monitoring (iCGM) may motivate changing eating and PA behaviors without
having to use fingersticks. Little is known about the relationship between the frequency of BG views
and BG or PA levels.

Objective: Use real-time iCGM and Fitbit monitors to examine temporal dynamic associations
Oral Presentation Abstracts
between the daily number of times iCGM glucose data were viewed, and daily mean BG and PA in
8 Black women with non-insulin-requiring T2DM.

Methods: We used a longitudinal, descriptive design. Participants wore a Freestyle Libre iCGM and
a Fitbit activity tracker 14 days. Participants could view BG throughout the day by swiping a reader
over the sensor. We used multilevel VAR (Vector Autoregressive Regression) models to estimate the
within-person, temporal dynamic associations between daily mean BG, number of daily swipes, and
daily mean PA. VAR is useful for capturing relationships between multiple variables as they change
over time. The estimation was done with R_mlVAR 0.5.2 version.

Results: Mean A1C was 6.8%. Estimation results showed individual-specific temporal patterns between
daily swipes, glucose, and PA levels. In some cases, more frequent swipes predicted lower mean BG,
either directly or indirectly through increased PA, while in other cases, more swipes predicted higher
mean BG. We found bidirectional temporal relationships between swipes and activity and between
swipes and mean glucose levels.

Conclusion: Diverse temporal patterns may help identify individuals for whom iCGM-based self-care
improves BG and PA levels. Effects would likely be more pronounced in a larger sample of women with
higher A1c levels.

Keywords: continuous glucose monitoring, diabetes, real-time data, active learning

_____________________________________________________________________________________________________
Correspondence: Cynthia Fritschi, Department of Biobehavioral Nursing Science, University of Illinois,
Chicago, College of Nursing, United States
E-mail: fritschi@uic.edu

59