Apple Watch And Android Wear Smartwatches Can Detect Early Signs Of Diabetes According To UCSF Study

Many wearable such as the Apple Watch, Android Wear smartwatches, and FitBit bands come with built-in heart rate sensors. While these are useful for gauging the impact of cardio-based workouts, it turns out they can also detect early signs of diabetes with a high level of accuracy, according to a new study by health startup Cardiogram and the University of California San Francisco (UCSF).

The study examined more than 14,000 Apple Watch and Android Wear device owners over the course of several weeks. During that time, Cardiogram collected 200 million heart rate and step count measurements, and used that data to train and validate a deep neural network called DeepHeart that can distinguish between people with and without diabetes. According to the study, DeepHeart's analysis was accurate 85 percent of the time.


"Typical deep learning algorithms are data-hungry, requiring millions of labeled examples, but in medicine, each label represents a human life at risk—for example, a person who recently suffered a heart attack or experienced an abnormal heart rhythm. To solve this challenge, researchers applied two semi-supervised deep learning techniques ('unsupervised sequence pretraining' and 'weakly-supervised heuristic pretraining') which made use of both labeled and unlabeled heart rate data to improve accuracy. The final deep neural network contained 564,227 neural network weights and both convolutional and recurrent layers," Cardiogram co-founders Johnson Hseih and Brandon Ballinger stated in an interview with Upbeat.

DeepHeart had already been validated to accurately detect hypertension (chronic high blood pressure), sleep apnea, and atrial fibrillation in previous studies presented at the American Heart Association's annual scientific sessions in November 2017, and Heart Rhythm Society in May 2017.

As for detecting diabetes using heart monitoring sensors and step count measurements, Cardiogram explains that the heart is connected to the pancreas via the autonomic nervous system. As people develop the early stages of diabetes, their pattern of heart rate variability shifts. In addition, a 2015 Framingham Heart Study showed that high resting heart rate and low heart rate variability predicts who will develop diabetes over a 12-year period.

More than 100 million adults in the US live with prediabetes or diabetes, according to the CDC. A quarter of those are undiagnosed, and 88.4 percent of people with prediabetes don't realize they have it. There have been attempts by Apple and others to add glucose-sensing hardware to wearables to detect diabetes, but so far that has not materialized. This is the first large-scale study that shows a regular heart rate sensor can be used in conjunction with an AI-based algorithm to identify early signs of diabetes.