Primary Care–Initiated Continuous Glucose Monitoring in Adults With Insulin-Treated Diabetes
Primary care prescribing of continuous glucose monitors was associated with better blood sugar control and fewer acute care visits.
*Retrospective cohort study; Level 2 (OCEBM).
Citation
Milosavljevic J, Rasquin Leon L, Rikin S, Schechter C, Hodgson S, Agarwal S. Primary Care–Initiated Continuous Glucose Monitoring in Adults With Insulin-Treated Diabetes. JAMA Network Open. 2026;9(7):e2621713. doi:10.1001/jamanetworkopen.2026.21713
Background
Continuous glucose monitors are recommended for people with diabetes treated with insulin, but most diabetes care occurs in primary care. This study examined whether starting these devices in primary care improves outcomes in a diverse safety-net population.
Patients
8502 adults with insulin-treated diabetes and primary care visits in a Bronx health system. Excluded: uninsured patients, prior monitor use within 2 years, and first monitor prescribed outside primary care.
Intervention
First continuous glucose monitor prescription by a primary care clinician.
Control
No monitor initiation during follow-up.
Outcome
Hemoglobin A1c level, all-cause hospitalizations, and emergency department visits.
Follow-up Period
Up to 3 years.
Results
| Outcome | Key result |
|---|---|
| Hemoglobin A1c at 12 months | Greater reduction with monitor: between-group difference −0.49 percentage points (95% CI, −0.62 to −0.35) |
| Hemoglobin A1c at 24 months | Between-group difference −0.54 percentage points (95% CI, −0.72 to −0.35) |
| Recurrent hospitalizations | 13% lower risk; hazard ratio 0.87 (95% CI, 0.77 to 0.98) |
| Recurrent emergency department visits | 18% lower risk; hazard ratio 0.82 (95% CI, 0.74 to 0.91) |
Results were based on prescription, not confirmed use, similar to an intention-to-treat estimate. A 0.5 percentage-point hemoglobin A1c reduction is commonly considered clinically meaningful. The control was usual care without a monitor, not current guideline-preferred care for many insulin-treated patients.
Limitations
Observational design limits causal certainty. Monitor use was not confirmed. Events outside the health system may have been missed. Findings may not apply to uninsured patients or those with little primary care access.
Funding
ADA and NIH; author had Dexcom and Abbott grants.
Clinical Application
Clinicians should consider initiating glucose monitors in primary care for insulin-treated diabetes, while ensuring training, access, and follow-up.
Discussion
Sign in to join the discussion.
In this cohort study, primary care–initiated CGM was associated with greater 12-month HbA1c reduction (between-group difference, −0.49 percentage points) and fewer acute care events; given the nonrandomized design and baseline differences, how valid and applicable are these findings for changing primary care diabetes practice? In this cohort study, CGM initiation was associated with a hazard ratio of 0.82 for recurrent ED visits. Which interpretation is most appropriate?