Apple Watch and other wearables can help in early detection of diabetes with over 85% accuracy, claims a study from the University of California, San Francisco (UCSF) and Cardiogram, a health startup. The researchers used just heart beat data, which they collected using Cardiogram’s DeepHeart neural network, to detect diabetes.
DeepHeart can analyze the beat-to-beat heart rate variability to look for signs related to diabetes, like slower heart rate recovery after exercise, or elevated resting heart rates. The pancreas is linked with the heart through the nervous system, and the changes are reflected in the heart rate when diabetes just starts creeping in, explains Cardiogram.
To track the heart beats, the researchers did not use any high-end devices. Rather, they just tracked the data of wearables such as Apple Watches, Fitbits and Android Wear devices that come integrated with the heart rate sensors.
Prentice Capital Benefits From “15 Million New Robinhood Accounts”
Michael Zimmerman's Prentice Capital returned 3% in June, taking its year-to-date performance to 18.6% net of fees and expenses, according to the firm's June investor update, a copy of which ValueWalk has been able to review. Prentice Capital Benefits From "15 Million New Robinhood Accounts" Prentice employs a low net equity long/short strategy with a Read More
In a statement to Engadget, Cardiogram co-founder Brandon Ballinger said, “It’s the hardest thing I’ve ever worked on, but the most rewarding.”
The startup has done several such studies in the past like training the neural network – DeepHeart – to identify the signs of strokes among approximately 6000 users.
For the latest study, however, where the machine also attempted to spot high cholesterol, sleep apnea, as well as, hypertension the startup had to deal with a larger pool of 14,011 users. Cardiogram notes that this is the first large-scale study to reflect that a normal heart rate sensor such as Apple Watch and Fitbit can help in catching the signs of diabetes.
Both Cardiogram and UCSF have been working over the past year or so to test the efficiency of the wrist-worn heart rate sensor in the detection of serious health conditions. Previously, Cardiogram used Apple Watch and its AI-based algorithm to detect abnormal heart rhythms with 97% accuracy, hypertension with 82% accuracy and sleep apnea with 90% accuracy.
Going forward, Cardiogram is planning to incorporate DeepHeart directly into the Cardiogram app to help in detection of the early signs of disease. Cardiogram publishes the majority of the findings in clinical journals or abstracts. The latest study will be published as well after it is presented at the AAAI 2018 conference this week, notes TechCrunch.
Diabetes is one of the most gripping issues in the U.S. with over 100 million adults in the country living with either pre-diabetes or diabetes condition. Also, more than 1 in 4 such cases are never diagnosed, according to the CDC (Centers for Disease Control and Prevention). Something that makes the patient reluctant in identifying the problem is the pain of the process. The patient has to prick themselves after each meal to measure the right amount of insulin and maintain the balance.
If all goes well, then Apple Watch and other wearables could be the painless solution to diagnosing diabetes. There has been no dearth of heroic acts of Apple Watch, where it saved people from heart attacks and pulmonary embolisms. Apple has been shipping record numbers of watches and most of the people are buying it for health and fitness benefits.