Kent Peterson, senior editor, has also produced award-winning work in television and radio.
Is there nothing our electronic gadgets can't do?
Researchers used data from the Apple Watch's heart-rate monitor to look for signs of diabetes, and it worked - an impressive 85 percent of the time.
But how can that be? What does your heart rate have to do with diabetes?
Answers come in a joint study from the University of California, San Francisco and Cardiogram, the creator of an app to help people make better use of heart-rate data from smart watches and other devices.
The heart-pancreas connection
"Your heart is connected with your pancreas via the autonomic nervous system," explained Cardiogram co-founder Johnson Hsieh in a news release. "As people develop the early stages of diabetes, their pattern of heart-rate variability shifts."
To detect the specific heart patterns that previous studies have linked to diabetes, researchers turned to the world of artificial intelligence.
Learning by example
In computer systems, a neural network uses "deep learning" to analyze an enormous amount of training data in ways that enable it to learn to make decisions about other data. It's much the way humans make decisions based on knowledge we have acquired previously.
Cardiogram's deep neural network, DeepHeart, was trained to distinguish between people with and without diabetes by comparing heart data examples from more than 14,000 Apple Watch users. Then, DeepHeart was put to the test. It used its new knowledge to analyze data from a separate group of people. Eighty-five percent of the time, DeepHeart could tell who had diabetes and who didn't.
The Apple Watch isn't the only device that can be used this way. Cardiogram data from Android Wear watches were also involved in the study. And Hsieh says the results show that data from ordinary heart-rate sensors in FitBits and other common devices should work too.
Data vs. doctor
One of the problems with diabetes is that many people who have it don't know it. Could a neural network linked to your smart watch detect and diagnose the blood sugar disease with a few taps of your fingertip? For now, the answer is no. There's a lot more work to be done to improve and prove the technology.
A neural network such as DeepHeart should never take the place of a physician. Only a qualified healthcare provider can make a diagnosis. But if this evolving technology says you may have diabetes, that's a good reason to see your doctor and find out for sure.
Do you use a smart watch, fitness tracker, or other device to improve your health? Tell us what you think of it by adding a comment below.
Image courtesy of Cardiogram