Apple Watch Detects Diabetes with 85% Accuracy

The Apple watch is not just a hi-tech fashion statement but is now able to detect diabetes with 85% accuracy, according to part of a larger DeepHeart study with Cardiogram and UCSF.  It used data from Apple Watch users and detected 462 of them as having diabetes using the watch’s built-in heart rate sensor and artificial intelligence-based algorithms.  In previous studies it was found that heart rate variability could predict incident diabetes and hypertension, thus opening up the possibility of using the watch to to detect a diabetic patient.  Other conditions were detectable as well, please see below:

• Hypertension – 82%
• Diabetes – 85%
• Sleep Apnea – 90%
• Abnormal Heart Rhythm – 97%

With 1 in 4 adults living with undiagnosed prediabetes or diabetes, it is becoming a growing problem in the U.S. Contributing to the growing problem is mainly the pain involved with checking blood glucose levels and keep their insulin balanced.  Early detection is one of the best ways to get diabetes under control before things get more complicated.  While other efforts have been made to create glucose sensing hardware, this is the first to use ordinary heart sensors paired with A.I. based algorithms to identify diabetes with no additional hardware.

More research is being done to increase the amount of diseases detectable through heart sensors, although caution is given that those in the study were already known to have diabetes.  Anyone who believes they are diabetic should see their doctor and not rely on the watch for diagnosis.  The conclusions are promising and we will see how far these methods will improve detection of diseases with simple readily-available tools used intelligently and efficiently.

Read More.