Researchers at Stanford University have developed an artificial intelligence (AI) algorithm that improves the diagnosis of diabetes, potentially leading to better and more accessible care.
Traditionally, diabetes is classified as Type 1 or Type 2, with Type 2 accounting for 95% of cases. Recent studies have identified important subtypes within Type 2 diabetes that can influence the risk of complications such as kidney and heart issues.
Tracey McLaughlin, MD, an endocrinology professor at Stanford, stated, "Understanding the physiology behind [diabetes] requires metabolic tests done in a research setting, but the tests are cumbersome and expensive." The new algorithm utilizes data from glucose monitors to identify three of the four most common metabolic subtypes of Type 2 diabetes with about 90% accuracy, surpassing traditional testing methods.
Identifying a patient's diabetes subtype can significantly affect treatment effectiveness, enabling healthcare providers to tailor personalized treatment plans and allocate resources more efficiently. This approach also reduces the need for complex clinical settings, as it leverages data from glucose monitors that patients often already use.
McLaughlin emphasized the importance of this development: "This matters, because depending on what type you have, some drugs will work better than others." The research aims to provide more accessible health information for individuals lacking adequate healthcare infrastructure.
With nearly 13% of the U.S. population diagnosed with diabetes, these advancements could greatly enhance treatment options and outcomes. This study follows the recognition of two over-the-counter glucose monitors at CES 2025, marking a significant step toward more accessible health technology.