Continuous glucose monitors are genuinely useful for some patients, a JAMA Internal Medicine review concludes, and the wellness market has outrun the evidence.

Continuous glucose monitors were built for people whose lives depend on knowing their blood sugar. A sensor worn on the arm reports a reading every few minutes to a phone, which for someone with type 1 diabetes replaced a regime of finger-pricks with something close to continuous information.

They are now sold to people who do not have diabetes, as wellness devices, some of them available without a prescription. A review in JAMA Internal Medicine from the University of Gothenburg examines what evidence supports that, and the answer is none.

The review was prepared at the journal’s request and published with an accompanying Patient Page. Minna Johansson, associate professor at Sahlgrenska Academy and a general practitioner at Herrestads Primary Care Center, is among the researchers behind it.

What the devices do well

The review is not a dismissal, and the distinction it draws is specific.

“This technology is tremendously valuable for certain groups of patients,” Johansson said.

For people with diabetes, continuous monitoring makes treatment adjustment easier and reduces repeated finger-stick testing, and it is particularly valuable for those at high risk of hypoglycaemia — where a reading that arrives before symptoms do can prevent an emergency. On average, use is associated with a modest reduction in HbA1c, the measure of longer-term glucose control.

What the review did not find was evidence that the technology prevents disease in people with type 2 diabetes. Easier management and lower HbA1c are real benefits; demonstrated prevention of the complications that make diabetes dangerous is a different claim, and it has not been established.

And for everyone else

For people without diabetes, the researchers found no evidence that continuous glucose monitoring improves health or prevents disease.

That absence is the review’s central point, and it identifies two ways the devices may do harm rather than nothing.

The first is interpretive. Normal blood glucose fluctuates — after meals, with exercise, with sleep and stress — and a healthy person watching those movements on a phone has no reference for which are meaningful. The review warns that normal fluctuations can be overinterpreted, potentially leading to unnecessary dietary restrictions or anxiety.

“Continuous glucose monitors generate enormous amounts of data that are difficult to interpret,” Johansson said.

This is a known failure mode of self-monitoring technology in general. A number that varies for ordinary physiological reasons invites a search for the cause of each variation, and the available response is usually to eliminate foods. A person building a restrictive diet around spikes that have no clinical meaning has been made worse off by the information.

The second harm is systemic. The review notes that healthy people are already seeking medical care to have their readings interpreted — which puts a primary care physician in the position of explaining a normal result from a device the patient bought, using time that was allocated to people who are unwell.

The marketing

Johansson’s view is that the companies selling the technology should be clearer about what is and is not known, and specifically that for people without diabetes there is no evidence of health benefit. She said those marketing it “should be much more transparent”.

The commercial logic is straightforward and worth stating. A device that monitors a chronic condition has a market the size of the patient population; the same device sold as a wellness product has a market the size of everyone interested in optimising themselves, which is far larger. The incentive to extend from one to the other does not depend on evidence appearing.

What a narrative review can and cannot settle

The method needs stating, because it determines the strength of the conclusion.

This is a narrative review rather than a systematic one — commissioned by the journal, surveying and interpreting the literature rather than applying a pre-specified search and inclusion protocol. That makes it a reasoned expert assessment rather than a reproducible accounting of all available evidence, and a reader cannot check which studies were considered or why.

The finding should also be read precisely. No evidence of benefit is not evidence of no benefit. Continuous monitoring in healthy people is recent enough that the long-term outcome trials which would settle the question have not been run, and absence of data is the expected state of affairs for a technology that became affordable to consumers only in the past few years.

The release reports no statistics at all — no effect size for the HbA1c reduction, no count of the studies reviewed, no quantification of the overinterpretation problem. For a review, the second of those is the number a reader would most want.

The useful summary is that the burden of proof sits where the review places it. A medical device being sold for a non-medical purpose should be able to show it helps, and the people best placed to say whether one of these devices would help a particular person remain that person’s doctor rather than its manufacturer.

The paper was published on 3 August 2026. The release names only Johansson among the authors, and states neither funding sources nor conflicts of interest.

---

Notes

1. “Continuous Glucose Monitoring in Type 2 Diabetes and Beyond.” JAMA Internal Medicine, 2026. DOI: 10.1001/jamainternmed.2026.2772.
2. University of Gothenburg. “No evidence that continuous glucose monitors improve health in people without diabetes.” EurekAlert!, 4 August 2026. https://www.eurekalert.org/news-releases/1138726.

Photo by JÉSHOOTS on Pexels.com