Researchers at Stanford University are using artificial intelligence (AI) to Digging deeper into diabetes diagnosis — and the results could mean better, more accessible care.
We commonly understand diabetes as either type 1 or type 2. However, in recent years, scientists have I discovered Significant mutations or subtypes within type 2, which accounts for 95% of diagnoses, reveal a person’s risk for related diseases such as kidney, heart, and liver problems.
Also: Two OTC Continuous Glucose Monitors Win Awards at CES – Try It Now
“Understanding the physiological functions behind [diabetes] “Metabolic testing performed in a research setting is needed, but the tests are laborious, expensive, and impractical for clinical use,” explained Tracy McLaughlin, M.D., professor of endocrinology at Stanford University.
Researchers developed an algorithm that uses data collected by blood sugar monitors to identify three of the four most common subtypes of type 2 diabetes.
When compared to clinical data, the algorithm “predicted metabolic subtypes, such as insulin resistance and beta-cell deficiency, with higher accuracy than traditional metabolic tests” (approximately 90% of the time).
Knowing patient subtypes can influence treatment efficacy. Physicians can develop individualized care plans and focus resources on each patient, reducing costs. Additionally, the study applies AI to data that is already easily collected by human blood sugar monitors. This means that the algorithm does not require a large or complex clinical environment to work.
“This is important because some drugs work better than others, depending on the type of patient,” McLaughlin says. “Our goal was to find a more accessible and on-demand way for people to understand and improve their health.”
Related Article: CES 2025: 22 Most Impressive Products You Don’t Want to Miss
Researchers believe the algorithm could make health information available at home to people who lack access to medical infrastructure due to geography, poverty or other factors.
Given that almost 13% of the U.S. population has been diagnosed with diabetes, these subtle differences can make a huge difference in treatment options and effectiveness. Especially if AI can glean better insights from data collected by wearables that patients already own and need.
Also: 3 ways AI is revolutionizing how healthcare organizations serve patients. Can an LLM like ChatGPT help?
Following CES 2025, where two commercially available blood glucose monitors were named winners in the digital health category, this study marks another step towards accessible health technology.
