Smart wearable device warns of diabetes with 85% accuracy

Apple Watch - Photo: REUTERS

According to technology news site CNET, research results from the Cardiogram app developer and UCSF University show that heart rate sensors on smart wearable devices such as Apple Watch, Android Wear, Garmin or Fitbit can detect diabetes with up to 85% accuracy.

The study was conducted on 14,000 users of Apple Watch and Android Wear smartwatches. Health sensor data was used to train a deep neural network called DeepHeart to determine whether someone had diabetes or not.

DeepHeart has previously been shown to accurately "diagnose" conditions such as atrial fibrillation, high blood pressure and sleep apnea.

In a 2015 study called the Framingham Heart Study, the authors showed that abnormal heart rhythm signals, such as an abnormally fast heartbeat while resting, were a marker for the risk of developing diabetes over the next 12 years.

The basis of this study provided information on 200 million different heart rates as a reference database as well as heart rate counting methods.

According to the US Centers for Disease Control and Prevention, more than 100 million adults in the US now have diabetes or prediabetes.

One-quarter of people with diabetes are undiagnosed and more than 88% of those with prediabetes are unaware of their condition.

Research from Cardiogram suggests that heart rate sensors on smart wearables, coupled with artificial intelligence algorithms, could become essential tools in early diabetes detection.