Researchers led by B. D. Ergin fitted blinded Dexcom G7 continuous glucose monitors (CGM) to 32 older hospital inpatients with type 2 diabetes and cognitive impairment, according to a feasibility study posted on medRxiv on February 9, 2026. All participants were naive to CGM and wore the sensors for up to 10 days; the study asked feasibility, acceptability and tolerability questions at the point the sensor was removed, rather than measuring whether the device changed blood sugar control or clinical outcomes. Mean participant age was 78, and the trial ran as an open, single-arm design, meaning there was no comparison group and no blinding of the research team to who received the sensor.
The question the study set out to answer is narrower than it might sound: not whether CGM helps people with dementia, but whether the technology can even be worn, tolerated and read on a ward where patients may be confused, agitated or unable to consent to a routine finger-prick check. That question follows an earlier body of observational work already reported by this newspaper linking glucose monitor use to fewer deaths and shorter hospital stays in people with dementia, findings that could not show cause and effect. This new study does not test whether CGM improves outcomes at all; it tests only whether the sensors stay on, whether staff can use the readings, and whether patients and carers find the process acceptable.
Because the design used blinded sensors, meaning glucose readings were not shown to clinical staff or used to guide treatment during the study, the trial cannot say anything about whether real-time CGM data would change insulin dosing, hypoglycemia rates or hospital stays in this population. It is also a single-arm study of 32 people at one design stage, run before any larger comparative trial, and the researchers themselves frame it as feasibility work meant to inform whether a bigger study is worth doing, not as evidence the technology should be adopted on wards now.