- Design
- Single-centre prospective protocol evaluation
- Population
- 54 general medicine and renally compromised inpatients, 62 admissions
- Primary outcome
- CGM accuracy vs point-of-care glucose; fingersticks avoided
- Effect
- MARD 11.9% overall, 10.0% after protocol exclusions; 53% of POC tests replaced
A single-centre evaluation in the Journal of Hospital Medicine (September 2026) extended a hybrid continuous glucose monitoring (CGM) protocol — CGM for treatment decisions, with fingerstick checks when the protocol requires them — to 54 general medicine and renally compromised inpatients across 62 admissions.
Across 732 paired readings, mean absolute relative difference (MARD) between CGM and point-of-care glucose was 11.9%; after the protocol removed 258 pairs flagged for confirmation, it was 10.0%. CGM replaced 53% of fingerstick tests, short of the 75% target. There were 24 hypoglycaemic events below 70 mg/dL.
This is a small uncontrolled evaluation without a comparison group, so it cannot show whether CGM changed hypoglycaemia or glycaemic control. What it shows is feasibility: in non-critically-ill ward patients including those with kidney impairment, CGM was accurate enough to guide most insulin doses when paired with confirmation rules.
- CGM can guide ward insulin dosing in non-critically ill patients when a protocol defines when to confirm with a fingerstick
- Always confirm hypoglycaemic or rapidly changing readings with a fingerstick
- Expect CGM to replace about half, not all, of point-of-care tests
- Kidney impairment did not make CGM unusable in this series
Why it matters
It extends inpatient CGM from selected infectious admissions to the renal and general medical patients who make up the ward.
Don't overread it
With no comparison group, it shows accuracy and feasibility, not better glucose control or fewer hypoglycaemic events.
The statistics, in plain English
MARD is the average percentage gap between CGM and fingerstick readings; below about 10-12% is generally treated as acceptable for dosing decisions. The protocol reached 10.0% only after excluding readings it flagged, which is how it would be used in practice.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for internal medicine, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free