Authors: Diana Ferro, PhD; David Williams, MPH; Brent Lockee; Colin Mullaney; Jacob Redel, MD; Lydia Skrabonja; Mark A. Clements MD, PhD; Ryan J. McDonough, DO
Children Mercy Research Institute, Kansas City, Missouri, USA
dferro@cmh.edu
+1 (816) 7317242
Background/Objective:
The use of continuous glucose monitors (CGM), insulin pumps (PUMP), and hybrid closed loop systems has been associated with improved outcomes in pediatric type 1 diabetes (T1D) care. We sought to characterize near-term outcomes occurring after disengagement from diabetes devices.
Methods:
We conducted a retrospective cohort study on youth, ages 0-21 years, in our diabetes clinics from 1/1/2018 through 12/1/2020. Using an IRB-approved data repository, we identified patients on a CGM or PUMP in two sequential clinical encounters (E1 and E2). For each engagement category we estimated the percent of patients with DKA events and increase in glycated hemoglobin (A1c) of >0.3% between visits in the 180 days following E2.
Results:
We identified a total of 519 patients that disengaged from pump or CGM. Compared to patients who remains engaged with diabetes devices, we observed a higher DKA rate (5.3% vs 3.4 %, p=0.01).
| Device | category | E1 | E2 | Obs | N | A1c ± STDEV | ∆A1c > 0.3 | DKA |
| CGM | Non-engagement | No | No | 4526 | 1320 | 9.2 ± 1.9 | 40% | 8.2% |
| New engagement | No | Yes | 955 | 806 | 8.5 ± 1.7 | 35% | 3.3% | |
| Disengagement | Yes | No | 410 | 364 | 9.0 ± 1.7 | 30% | 4.1% | |
| Continued engagement | Yes | Yes | 3735 | 1075 | 8.0 ± 1.3 | 32% | 2.4% | |
| PUMP | Non-engagement | No | No | 3118 | 938 | 8.9 ± 2.0 | 42% | 7.5% |
| New engagement | No | Yes | 385 | 373 | 8.3 ± 1.7 | 38% | 4.7% | |
| Disengagement | Yes | No | 213 | 206 | 9.1 ± 2.0 | 38% | 8.5% | |
| Continued engagement | Yes | Yes | 3754 | 1184 | 8.5 ± 1.5 | 34% | 3.4% |
Conclusions:
CGM and PUMP engagement support 2 of the 7 pillars of diabetes self-management. This study highlights that disengagement from these devices associates with increased prevalence of rising A1c and DKA in the near term. Future studies identifying predictors of disengagement may offer novel intervention opportunities to forecast device disengagement and to develop interventions to prevent it.
Key Words:
Glycated Hemoglobin A, Retrospective Studies, Insulin
