Researchers have found that using Cardiogram‘s deep neural network technology, data collected from ordinary wearables could help flag early signs of diabetes.
In one 14,000-person study, the company’s DeepHeart technology was able to distinguish between people with and without diabetes at 85% accuracy.
Investigators from Cardiogram and the University of California San Francisco presented the study’s findings today at the AAAI Conference on Artificial Intelligence in New Orleans.
In the U.S., more than 100 million adults are living with prediabetes or diabetes, according to Cardiogram co-founder Brandon Ballinger, and one in every four people with diabetes is undiagnosed. Ballinger touted the large-scale study as the first to demonstrate that ordinary heart-rate sensors could be used to ID early warning signs of diabetes when paired up with an artificial intelligence-based algorithm.
To train the DeepHeart network to detect diabetes, researchers recruited 14,001 Cardiogram users and collected health sensor data from people with and without diabetes, hypertension, sleep apnea, atrial fibrillation and high cholesterol.
“Your heart is connected with your pancreas via the autonomic nervous system,” co-founder Johnson Hsieh explained in prepared remarks.
“As people develop the early stages of diabetes, their pattern of heart rate variability shifts. In 2015, the Framingham Heart Study showed that high resting heart rate and low heart rate variability predicts who will develop diabetes over a 12-year period. In 2005, the ARIC study showed that heart rate variability declines faster in diabetics than non-diabetics over a 9-year period.”
Cardiogram’s app is compatible with all versions of the Apple Watch and any Android Wear watch with a heart rate sensor, according to the company. This year, Cardiogram plans to integrate its DeepHeart network directly into its app.
