An artificial intelligence (AI) network, DeepHeart, developed by Cardiogram, can use data from wearable devices to detect the early signs of diabetes, a study has found.
A joint study between Cardiogram and the University of California, San Francisco, was able to detect diabetes with 85% accuracy in 14,000 wearable users, solely based on individuals’ heart rate and step count from devices. Developers, however, caution that this tool was not designed to diagnose, but to provide a potential screen for pre-diabetic patients; still, the implications of this would shake the current makeup of the diabetic space.
According to GlobalData, over 32 million people in the US are diagnosed with diabetes; a further 70 million people are thought to be living with prediabetes. With 88% of pre-diabetic individuals unaware of their condition, and one in four cases of diabetes going undiagnosed, DeepHeart could provide an easy and affordable means of diagnosing a condition with ever-increasing societal importance. In 2017, diabetes alone cost the US economy $245bn, according to the American Diabetes Society.
An opportunity for tech giants
The combination of AI with wearable devices is another win for tech giants Apple and Google, as they continue to explore the realm of digital health. Apple’s intentions to become a key player in the healthcare space are crystal clear, with 2017 seeing multiple health improvements and updates to its Apple Watch, a new patent for the potential use of iPhone as a health sensor, the launch of its first sponsored clinical trial, and multiple acquisitions. The company is therefore aware of the commercial rewards the healthcare space could reap.
This is not the first time Cardiogram has trained AI to identify disease; the start-up company has previously used its deep learning technology to detect atrial fibrillation, hypertension and sleep apnea.
Cardiogram’s progress in this field signifies the continued penetration of wearables into the healthcare industry, and the prolonged benefit to patients and the state. If patients and physicians are able to exploit wearables to better diagnose, manage and treat disease at earlier stages, these technologies could completely alter the way diseases are currently managed.
The power of wearables will be further realized as healthcare systems increasingly move towards value-based care and pricing models. As such, if diseases are able to be detected and managed in a more streamlined way using developments and data from wearables, the value of these devices will become fundamental to healthcare providers and payers.