isCGM reduces acute diabetes hospitalizations

Publicly funded scanned glucose monitoring was associated with fewer acute diabetes hospitalizations in insulin-treated type 2 diabetes.
*Population-based quasi-experimental cohort; Level 3 (OCEBM).

Citation

Rodríguez de Vera Gómez P, Mayoral E, Rodríguez Jiménez B, Gomez-Peralta F, Umpierrez GE, Martínez-Brocca MA. Diabetes Care. 2026;49(8):1442-1450. doi:10.2337/dc26-0547

Background

Continuous glucose monitoring can improve blood sugar control, but large-scale effects on hospital use and costs in real-world type 2 diabetes care are less certain.

Patients

15,413 adults in Andalusia, Spain, with type 2 diabetes using multiple daily insulin injections. Excluded: pregnancy, no public authorization for the device, or insufficient before-and-after follow-up.

Intervention

Publicly funded intermittently scanned continuous glucose monitoring using Abbott FreeStyle Libre.

Control

Patients’ own prior period using conventional finger-stick glucose monitoring.

Outcome

Hospitalizations for acute diabetes complications and cardiovascular events, length of stay, hospital costs, and hemoglobin A1c.

Follow-up Period

Mean 22.5 months before and 19.1 months after device start.

Results

OutcomeBeforeAfterEffect
Hemoglobin A1c8.09%7.65%Mean difference -0.44 percentage points
Acute diabetes hospitalizations74.6 per 10,000 person-years27.5 per 10,000 person-yearsRate ratio 0.37; about 212 person-years treated to prevent 1 admission yearly
Severe high-blood-sugar crises41.8 per 10,000 person-years14.2 per 10,000 person-yearsRate ratio 0.34
Simple high blood sugar admissions26.7 per 10,000 person-years9.3 per 10,000 person-yearsRate ratio 0.35
Median hospital stay, all studied admissions4 days3 days1 day shorter
Total inpatient costs$9.59 million$5.82 million$3.77 million lower

Severe low-blood-sugar admissions and overall cardiovascular hospitalization rates did not significantly change in simple before-after comparisons. Interrupted time-series modeling suggested cardiovascular admissions were lower than expected from prior trends. Results used within-person before-after comparisons. The control reflected the then-current usual monitoring approach.

Limitations

No randomized or external control group; residual confounding and selection bias are possible. The rollout prioritized more complex, vulnerable patients, so results may not apply to lower-risk or non-insulin-treated patients. Emergency-only visits and outpatient costs were not captured.

Funding

Spanish Diabetes Society grant; device-maker conflicts may bias interpretation.

Clinical Application

Consider scanned glucose monitoring for high-risk insulin-treated type 2 diabetes patients to reduce acute admissions; evidence supports but does not prove causality.