Unlike snapshot doctor visits, wearables get full baseline for precision medicine.

There’s been plenty of hope and buzz in the recent years that wearable technology will upend healthcare as we know it. Commercial devices strapped to our persons will—makers promise—empower individuals to monitor and control their own health, plus they'll help guide the care people receive by medical professionals.

But so far, there’s been a rather noticeable gap between the data we collect on our little devices and better health and healthcare for most. Does your doctor really care about or know what to do with your Fitbit data? Is your Apple Watch making you healthier?

Often, when people use wearables, “they get all excited for about three months, and then they stop looking at them,” precision medicine expert Michael Snyder of Stanford told Ars.

Snyder and colleagues think they can change that, though—making wearables finally work for our health in the process.

Using current, commercially available wearables with simple sensors—such as those that measure skin temperature, heart rate, and blood oxygen levels—Snyder and his team translated basic physiological data into a personalized baseline of what “healthy” looks like for each individual wearer. Then they correlated deviations from that baseline with different disease states. Their method successfully detected oncoming infections and could predict progression toward diabetes in study volunteers.

The findings, published Thursday in PLOS Biology, suggest that wearables could provide highly individualized and sensitive monitoring of a wearer’s health—something of value to doctors that they currently just can’t get otherwise, Snyder argues.

“When you’re healthy, you probably go to the doctor once a year, or every two years,” Snyder explains. And you get a battery of tests in those infrequent visits—your weight, temperature, blood pressure, etc. Those aren’t necessarily enough to know what’s normal for you, Snyder said. Maybe you get nervous in doctors’ offices or you’re just having an "off" day. A wearable that monitors you all the time, on the other hand, could tell.

Snyder likens his vision for wearables to sensors in your car’s engine—and the little check engine light that comes on when something is out of sorts.

This isn’t a wholly new idea, of course. But the sticking point is being able to take data that we can measure with wearables—accurately, at least—and translating them to helpful medical information. To do this, the team recruited around 60 people, including Snyder himself, to wear a variety of wearables. Snyder is a monitoring enthusiast, and he even wore seven wearables at one point (he insists this wouldn’t be necessary for other consumers).

Your 'check human light' is on...

The researchers collected more than 250,000 data points from the volunteers’ wearables over the course of a few years. With that data, they could come up with physiological patterns for each user, accounting for daily cycles, sleep habits, and activities.

When they paired those patterns with results from the volunteers’ medical tests, they could create correlations between basic physiology and disease states. For instance, patients with skin temperatures and heart rates a bit above their norms were linked to having higher levels of something called C-reactive proteins in their blood, which is a general sign of inflammation. Low-grade inflammation can be an early sign of infection or disease.

During the study, this actually played out for Snyder. On a trip abroad, he noticed his wearable data was off baseline after a long flight. He had an elevated heart rate and skin temperature, plus low blood oxygen levels. When he went to the doctor, he learned he had Lyme disease—a bacterial infection that can cause serious damage if left untreated.

From 12 participants whose medical tests revealed insulin resistance, a precursor to diabetes, the team came up with an algorithm that could predict insulin resistance just based on data from wearables—specifically, their daily steps, relatively high daytime heart rates, and the relationship between their daytime and nighttime heart rates.

Snyder and his team are already working to follow up on the studies, collecting more data, developing their own apps, and even talking with wearable designers. They’ll need a lot more data and tweaking before this gets to consumers. And of course, if they’re going to make any medical claims, they’ll need approval from the Food and Drug Administration, which can take years.

Still, Snyder is hopeful that round-the-clock health monitoring wearables and apps are not too far off. “In general, doctors are very resistant to new technologies because they want to make sure they’re validated and work,” he said. Plus, when a lot of wearables first came out, they weren’t that accurate. “But they’re quite accurate now,” Snyder said, and today there’s a lot of wearable-derived data to study and experiment with.

He envisions that not only healthy adults can keep track of their own health, but that parents can monitor their kids' health and caregivers can track their elderly patients.

One of the big criticisms Snyder has heard is that such wearable data could drive people to be hypochondriacs or request medical tests they don’t need with every little data blip. But Snyder argues it may do the opposite: future apps that generate the human equivalent of a check engine light could be set to high thresholds, making false alarms less likely. And being informed about normal fluctuations for you, specifically, could put some at ease.

“The nice thing about all this,” Snyder says, “is that it’s all done on a personal level—you’re measuring individual people’s data, and you can find changes for that particular person rather than the population as a whole. And that is what personalized medicine, precision medicine is all about.”