Smartwatch data can predict blood test results, study reports

Stanford researchers found that data from smartwatches can flag early signs of some health conditions and predict the results of simple blood tests.

- By Hanae Armitage

Michael Snyder and his team have shown that smartwatches can predict dehydration, anemia and illnesses such as the flu or a cold. 
nimito/Shutterstock.com

A smartwatch can signal physiological changes, such as a change in red blood cell count, as well as early signs of dehydration, anemia and illness, according to a new study led by researchers at Stanford Medicine.

The study is among the first to show that smartwatch data correlates with laboratory test results.

Scientists from the lab of Michael Snyder, PhD, professor and chair of genetics, tracked data from smartwatches, blood tests and other tests conducted in a doctor’s office in a small group of study participants. They were curious whether smartwatch readouts, such as heart rate and physical activity, could show physiological changes that are typically revealed through clinical measurements, including blood tests.

“We wanted to know if smartwatch-derived signals could predict the same values without any invasive intervention,” said Lukasz Kidzinski, PhD, a research associate in bioengineering.

The answer was a resounding yes. Changes in heart rate, for instance, often predicted changes in red blood cell count and hemoglobin, the molecule that carries oxygen. The team also found that smartwatch data gave more consistent heart-rate readouts than those taken at a doctor’s office. Consistent heart-rate measurements are believed to be the most accurate ones, Snyder said.

The results are another win for wearables — such as smartwatches, continuous glucose monitors and smart rings for your finger — which keep people constantly informed about their health.

A paper detailing the study was published online May 24 in Nature Medicine. Kidzinski is the lead author. Snyder, the Stanford W. Ascherman, MD, FACS, Professor in Genetics; Trevor Hastie, PhD, the John A. Overdeck Professor and a professor of statistics and of biomedical data science; and Jessilyn Dunn, PhD, a former postdoctoral scholar who is now an assistant professor at Duke University, share senior authorship.

Skin-deep data

For the study, which ran more than three years, 54 volunteers donned a smartwatch that tracked heart rate, step count, skin temperature and electrodermal activity, a measure of electrical activity in the skin that acts as a proxy for sweat production. Throughout the study, researchers also used traditional means, such as a heart rate monitor and blood tests, to track heart rate, red and white blood cell count, blood oxygen levels, glucose levels, and other health indicators. The researchers found that smartwatch readouts correlated with health parameters measured at a doctor’s office.

Michael Snyder

Certain combinations of smartwatch-derived data were able to provide insights into specific clinical blood tests. For instance, the researchers saw that participants with a rise in skin temperature and a decrease in movement generally also had a higher immune cell count — an indication of an illness such as a cold or the flu.

“It makes sense because if someone is getting sick, they may spike a fever, and they’re likely more tired and less active,” Dunn said.

They also found that low electrodermal output — or less sweat — was associated with dehydration, and that alterations in heart rate could be used to predict changes in blood oxygenation. They noted, however, that the smartwatch data could not accurately predict values such as exact red blood cell count, but it could flag early signs that something is amiss, such as a low level of red blood cells, which can be a sign of anemia.

Signals such as heart rate, body temperature, physical activity and oxygen levels are usually not enough to diagnose a disease, but the measurements are still useful in assessing overall health, or for monitoring recovery after surgery, Snyder said. He stressed that smartwatches are not a replacement for a doctor.

“I think this is just the beginning,” he said. “Devices are becoming much more sensitive and with many more capabilities. As the technology continues to advance, people will be better equipped to understand what’s going on with their own health in real time, just through their wearable devices.”

Other Stanford co-authors are former undergraduate researcher Ryan Runge; research engineer Jennifer Hicks, PhD; Sophia Miryam Rose, MD, instructor of genetics; former postdoctoral scholar Xiao Li, PhD; genetics lecturer Amir Bahmani, PhD; and Scott Delp, PhD, professor of bioengineering and of mechanical engineering.

Researchers from Duke University also contributed to the study.

The study was funded by the National Institutes of Health (grants 1U54DE02378901, UL1TR001085, NIH U54 EB020405 and K08 ES028825).

About Stanford Medicine

Stanford Medicine is an integrated academic health system comprising the Stanford School of Medicine and adult and pediatric health care delivery systems. Together, they harness the full potential of biomedicine through collaborative research, education and clinical care for patients. For more information, please visit med.stanford.edu.

2023 ISSUE 3

Exploring ways AI is applied to health care