How AI could supercharge your glucose monitor – and catch other health issues

Researchers at Stanford have been using artificial intelligence (AI) to dive deeper into diabetes diagnosis — and the results could mean better, more accessible care. 

We commonly understand diabetes as being either Type 1 or Type 2. But in recent years, scientists have discovered important variations, or subtypes, within Type 2 — which makes up 95% of diagnoses — that shed light on the risk of contracting related conditions, like kidney, heart, or liver issues. 

Also: Two OTC continuous glucose monitors won awards at CES – and you can try them now

“Understanding the physiology behind [diabetes] requires metabolic tests done in a research setting, but the tests are cumbersome and expensive and not practical for use in the clinic,” explained Tracey McLaughlin, MD, an endocrinology professor at Stanford.

Using data collected by glucose monitors, researchers developed an algorithm identifying three of the four most common subtypes of Type 2 diabetes. 

Compared to clinical data, the algorithm “predicted metabolic subtypes, such as insulin resistance and beta-cell deficiency, with greater accuracy than the traditional metabolic tests” — roughly 90% of the time. 

Knowing a patient’s subtype can impact treatment efficacy. Doctors can develop personalized medicine plans and better focus resources from patient to patient, reducing costs. Plus, the study applies AI to data already being easily collected by a person’s glucose monitor, meaning the algorithm doesn’t require a larger or more complicated clinical setting to work.  

“This matters, because depending on what type you have, some drugs will work better than others,” said McLaughlin. “Our goal was to find a more accessible, on-demand way for people to understand and improve their health.” 

Also: CES 2025: The 22 most impressive products you don’t want to miss

Researchers believe the algorithm will make health information more available at home for those who may not have access to healthcare infrastructure due to geography, poverty, or other factors. 

Considering almost 13% of the US population has been diagnosed with diabetes, these nuances could make a big difference in treatment options and efficacy — especially if AI can gather better insights from data collected by a wearable that patients often already have and need. 

Also: 3 ways AI is revolutionizing how health organizations serve patients. Can LLMs like ChatGPT help?

Following CES 2025, where two over-the-counter glucose monitors were named Honorees in Digital Health, the study marks another step forward toward accessible health tech.