What your wearable knows

Changes in usual patterns can tip you off to illness

Winter 2017
Michael Snyder collected nearly 2 billion measurements from 60 people. Data from his own biosensors helped him realize he had Lyme disease.

Soon enough, Snyder developed a fever. Two weeks prior, he had been helping his brother build a fence in rural Massachusetts. He was worried he might have been bitten by a tick and contracted Lyme disease, so he convinced a Norwegian doctor to write him a prescription for the antibiotic doxycycline. Blood tests subsequently revealed that he had indeed been infected with Lyme disease.

What happened to Snyder is an example of the type of early warning of disease that wearable sensors can supply, according to a recent digital health study in his lab. Using continuous data from biosensors plus periodic lab tests, the research team collected nearly 2 billion measurements from 60 people, including weight, heart rate, oxygen in the blood, skin temperature, sleep, activity, calories expended, acceleration, and exposure to gamma rays and X-rays. They first established baseline ranges for each participant; by monitoring deviations from those baselines, they were later able to detect the onset of infection, inflammation and even insulin resistance.  

The results of the study, published in January in PLOS Biology, raise the possibility of identifying inflammatory disease in individuals who may not even know they are getting sick. For example, in one instance, higher-than-normal readings of heart rate and skin temperature correlated with increased levels of C-reactive protein in blood tests. C-​reactive protein is an immune-system marker for inflammation and often indicates infection, autoimmune diseases, developing cardiovascular disease or even cancer.

Snyder’s own data revealed four separate bouts of illness and inflammation, including the Lyme disease infection. “Wearables helped make that initial diagnosis,” he says of the Lyme disease.

Wearables could also help identify people with insulin resistance, a precursor to Type 2 diabetes. The research team designed and tested an algorithm combining participants’ daily steps, daytime heart rate, and the difference between daytime and nighttime heart rate. The algorithm could identify individuals in the study who seemed to be developing insulin resistance.

Today’s wearables — smart watches and fitness monitors — primarily measure activity, but Snyder says they could easily be adjusted to more directly track measures of health. Every person’s wearables could carry normal baselines for dozens of measures, and automatic data analysis could combine patterns of outlier data to flag the onset of ill health and provide opportunities for intervention, prevention or cure.

“We have more sensors on our cars than we have on human beings,” Snyder says. In the future, he predicts, that situation will be reversed.

Author headshot

Jennie Dusheck

Jennie Dusheck is a science writer for the medical school's Office of Communication & Public Affairs. Email her at dusheck@stanford.edu.

Email the author