What If an App Could Tell You When You’re Getting Sick?

A Stanford geneticist may be onto something. Body data collected by smartwatches and other sensors can tip us off to brewing colds or infections

Michael Snyder-Stanford.jpg
Stanford researcher Michael Snyder led a study on how wearable sensors could help predict illnesses. Steve Fisch/Stanford Medicine

Wouldn’t it be great if you could get an early warning that you’re coming down with something, even before you start to feel sick?

One day, says Michael Snyder, there may be an app for that.

Snyder is a professor of genetics at Stanford and lead author of a recent study published in PLOS Biology suggesting that data gathered from smartwatches and other wearable devices could be used to clue us in on brewing health problems, from a cold to conceivably, diabetes.

For the past two years, he and his research team have been compiling the results from roughly 60 volunteers—himself included—who have been diligently tracking their bodies’ behaviors through wearable biosensors. All those monitors—some people wore half a dozen—collected more than 250,000 measurements a day on everything from heart rate to blood oxygen levels to skin temperature to physical activity to sleep patterns.

The idea was to first establish baseline measurements for each person, then determine how deviations correlated with illnesses or other environmental factors that can affect health. For years, Snyder has been creating health profiles for a group of volunteers through more conventional methods, such as blood and urine testing. Then he took notice of the wave of new devices on the market.

“When wearables came out, we were very intrigued and wanted to see what we could learn from them,” he says. “There is a huge advantage to working with wearables. They do continuous measurements and that’s a real plus.”

It allows tracking to occur over a long timeline and not just sporadically, when any number of random factors could affect the results. And that gave the researchers a clearer picture of how a person’s body could reflect an illness before he or she was aware of it.  

“The overall impetus,” Snyder notes, “was to define what it means for people to be healthy on an individual level, and then when disease occurs, understand what’s happening when it’s occurring.”

Catching Lyme in time

As it turned out, Snyder’s participation in the research paid off for him personally. While on a flight overseas, he began to feel a bit light-headed, so he checked his devices. His blood oxygen level had dropped during the flight, which wasn’t unusual, but it had fallen much lower than normal.

“And it didn’t come all the way back when we landed. My heart rate was elevated as well,” he says. “I measure myself often so it was pretty obvious that something wasn’t quite right.”

Before leaving on the trip, Snyder had spent some time in western Massachusetts helping his brother put up a fence. He didn’t notice if he had been bitten by a tick, nor did he see any indication of it. But the readings on his devices suggested something was going on in his body. Maybe it was Lyme disease, he thought.

Snyder was right, and able to get a prescription for the antibiotic doxycycline, which quickly took care of the bacterial infection.

He concedes that most people aren’t so in touch with their body’s signals that they can make that kind of diagnosis. But that’s the point, he says, of exploring the potential of wearable devices as a warning system. And not just for festering colds or infections, but also possibly for chronic diseases, like diabetes.

The researchers found that based on blood tests, a dozen of the people in the study had insulin resistance, a precursor to Type 2 diabetes. The scientists then designed an algorithm that combined participants’ data tracking their daily steps, daytime heart rate and the difference between their daytime and nighttime heart rates. From that, they were able to identify similar deviations in those who were insulin-resistant.

Snyder says more research needs to be done to prove those types of correlations, but the ultimate goal is to create an app that will be able to alert people when their body’s numbers are going out of whack.

The case for health sensors

Snyder disputes the notion that having constant access to this kind of body data will turn us all into hypochondriacs or that it will boost patient demands for more lab tests.

“I don’t think it would be any more than the invention of the oral thermometer led to more people going to the doctor,” he says. “If you’re a bit of a hypochondriac, you could set the alert threshold higher. But I think it could alert you to something going on and cause you to manage things better.

“I think it would be particularly valuable for parents with their kids or people who are caring for the elderly,” he adds. “In the future, I think you’ll just be able to look at your smartphone and see that your kid’s temperature has been running high the last three hours. No school for them.”

Snyder acknowledges resistance from some parts of the medical community regarding the potential risk of people relying too heavily on sensors to self-diagnose. But he sees a day when physicians will have easy access to their patients’ digital data, and that that range of information, instead of a single heart rate and blood pressure check in the doctor’s office, will make it easier to make an accurate diagnosis.

“More information has got to be more valuable than less information when it comes to managing your health,” says Snyder. “[Otherwise] That’s like driving around in your car with no gauges on it. Who would do that?”

Get the latest stories in your inbox every weekday.