DOI: 10.2337/db23-157-lb ISSN: 0012-1797

157-LB: A Randomized Trial of Continuous Glucose Monitoring upon Emergency Department Discharge

  • Endocrinology, Diabetes and Metabolism
  • Internal Medicine

Background: People with type 2 diabetes (T2D), particularly those with limited access to longitudinal care, frequently present to the emergency department (ED), leading to increased hospital utilization and disjointed care. Continuous glucose monitoring (CGM) has been shown to improve patient engagement and glycemic outcomes in outpatients, so we hypothesized that it would be beneficial if initiated upon ED discharge.

Methods: We randomized adults with T2D who were seen in the ED for hypo- or hyperglycemia to either flash CGM or care coordination alone. All participants were scheduled to follow up in our diabetes specialty clinic within two or three weeks. Outcomes included clinic attendance, patient-reported diabetes-related distress, satisfaction with CGM, and follow-up hemoglobin A1c.

Results: We recruited 30 participants, including 13 with newly diagnosed T2D. We found no significant difference between the CGM (n=16) and control (n=14) groups in terms of clinic attendance (75 versus 64%, P=0.69) or the change in diabetes-related distress as measured via the Diabetes Distress Scale. Among participants in the CGM group who returned for follow up, 83% found CGM helpful, as did 90% of their providers. The mean A1c reduction after approximately 3 months was greater in the CGM group (5.2 versus 1.9%, P=0.051), although the groups were not balanced in terms of diabetes duration. Among participants with newly diagnosed T2D, excluding those missing data, five out of five in the CGM group had a follow-up A1c under 7% (from a mean baseline of 12.9%) compared to zero out of two in the control group (P=0.048).

Discussion: Our study demonstrates the feasibility of starting CGM in the ED, an important setting for engaging difficult-to-reach patients. We did not see an effect of CGM on follow-up clinic attendance or diabetes-related distress, but there was a trend towards improved A1c at 3 months in the CGM group. Our study was limited by its small sample size, however, as recruitment in the ED can be challenging.


M. J. O'connor: None. X. Ding: None. C. Hernandez: None. L. M. Hubacz: None. R. J. Church: None. L. O'connor: None.


Herman G. Berkman Diabetes Clinical Innovation Fund

More from our Archive