A study finds smartwatch data combined with blood tests can accurately predict type 2 diabetes risk earlier than traditional methods.

A new study in Nature shows that combining smartwatch data—like heart rate, activity, and sleep patterns—with routine blood tests can accurately predict insulin resistance, a precursor to type 2 diabetes. Using machine learning on data from over 1,100 people, researchers achieved high accuracy in identifying metabolic risk earlier than traditional methods. The approach, which leverages real-world physiological signals, could enable widespread, low-cost screening and early intervention. While promising, broader use requires validation across diverse groups, clear clinical guidelines, and solutions to privacy and access concerns.