Smartwatch data can be used to detect and even predict disease, researchers say. This is good news for people betting on the potential of smartwatches to improve health beyond encouraging us to move more. If it becomes common to use this data to predict illness, we might be able to take better care of ourselves and save a lot of visits to the doctor.
Most people wear smartwatches continuously, so it generates an enormous number of metrics that can paint a fuller picture of health data than one doctor’s visit each year. For a study published today in the journal PLOS Biology Stanford University researchers gave 43 participant a Basis B1 or Basis Peak smartwatch and then analyzed nearly a year of their data. These measurements taken included heart rate, skin temperature, and sleep data. Then they analyzed the data and found that out-of-the-ordinary measurements, especially for heart rate, correlated with things like the common cold and even, in one case, Lyme disease.
In addition to the 43 participants, the researchers collected much more detailed measurements from one participant, called Participant #1, for nearly two years. (He sometimes wore up to seven wearable gadgets.) The researchers looked at his data and picked out the four dates where the measurements were out of the ordinary, with the heart rate and skin temperature especially being elevated. (Fitness trackers aren’t that great at measuring heart rate, and the Basis Peak didn’t measure up well, but the researchers said they compared the measurements to ones done in their lab at Stanford. And anyway, they were concerned with the fact that the measurement was unusual compared to normal more than if the measurement was actually accurate.)
During one period where the measurements were abnormal, he had developed Lyme disease. During the other periods, he had a fever or the common cold. All of these correlate with inflammation, so it’s likely that the data is picking up directly on signs of inflammation.
Then, the researchers picked three other people who had been ill during the period that they used Basis watches. For them, too, their heart rate and skin temperatures were elevated during periods of sickness. Using this data, they wore a software program to actually predict which measurements suggest that someone was about to become ill.
In a separate experiment, the team measured the insulin resistance of 20 people, and found that there was a relationship between insulin resistance, body mass index, and heart rate, the last of which was tacked by a smartwatch. Diabetes, then, is another disease that could potentially be predicted by data from wearables.
The authors hope that it’ll become more common to use data from wearables to detect illness. It is possible that this will lead to false alarms, they note, and that there are potential problems with patient privacy. But if this becomes widespread, it could be a convenient diagnosis tool, especially for people who are low-income or live in rural areas or have a hard time getting to the doctor for a routine check-up.