Fever monitoring has developed something of a bad reputation under COVID-19. While having a fever is one of COVID-19’s telltale symptoms, temperature checks capture only a moment in time. Unless someone is stricken with fever, they tell us very little about a person’s state of health. But a new report suggests that body temperature can play a far more useful role in understanding health—we’re just using it wrong.
Researchers at the University of California, San Francisco, and the University of California, San Diego, have shown that constant temperature surveillance could be a promising method for detecting and predicting the onset of fever in COVID-19. In a 50-person feasibility study, researchers used the Oura Ring, a finger-worn sleep tracker, to monitor temperatures of participating healthcare workers and adult volunteers.
Everyone has a slightly different baseline body temperature. If that baseline starts to tick upward over a series of days, it may indicate illness. However, this study was not designed to definitively show that the Oura Ring—or any other temperature-tracking wearable—can predict COVID-19. This was merely a feasibility study, showing whether it’s possible to use this tech to conduct further studies into fever-related illness detection. Still, it represents a first step in understanding whether temperature monitoring can ultimately be used to detect or predict illness.
The study is part of an ongoing research project between UCSD and UCSF called TemPredict that hopes to understand the role body temperature, heart rate, and other physiological signals can play in predicting the occurrence of COVID-19 symptoms. The concept under investigation is whether data from wearables like the Oura Ring could be used to help public health officials navigate outbreaks like COVID-19 or even a regular flu season.
“For example, if you’re a public health official, you could have a weather map and just know a bunch of people are starting to get a fever in Seattle and [you] should go look at that,” says Benjamin Smarr, assistant professor of bioengineering and data Science at UCSD and lead author on the paper. “What this paper shows is that not only is that technically feasible, but that actually the physiology data seemed to do a much better job than the people of reporting when it is they seem like they are getting sick.”
A slow process
Tracking an outbreak like COVID-19 or any incident in public health can be slow moving, Smarr says. Public health officials have to wait until people are sick enough to go to a hospital or a doctor’s office and have their symptoms correlated to a particular disease. Then that data has to be reconciled with public health officials at the county level before a summary statistic is released.
“If people are willing to act as tracer particles, we could know within a day, if not in real time, where sickness events are happening,” Smarr says. Furthermore, he notes, public health officials could more quickly connect illness to either a specific disease outbreak or environmental effects such as weather, pollution, or accidents.
To use wearable-collected data would require a digital public health infrastructure that currently does not exist. The U.S healthcare system is frequently called a patchwork, but that would suggest that it’s in any way sewn together. In reality, each healthcare operator is its own carefully constructed garment. A person can amass a collection of different healthcare services, piecing them together much the way one does an outfit. This structure allows individuals to choose and shed services and systems based on what they think they need. This choice forces health systems to compete and, in theory, provide better care. In practice, it creates complications for public health efforts, as revealed by COVID-19.
While there is no public digital infrastructure currently that would allow for the kind of real-time health weather mapping that Smarr is envisioning, smart thermometer maker Kinsa has a program that gives a glimpse of what such a public health program could look like. The company works with schools that support students from low-income families to navigate flu season. Kinsa supplies families with smart thermometers and an app that records temperature data and symptoms and lets them text school nurses. School nurses get a platform where they can see anonymized temperature data for each grade. The system allows these schools and families to coordinate and reduce flu spread in schools.
Smarr says the difference between using a smart thermometer and a ring that constantly monitors body temperature is that a thermometer relies on an individual to experience symptoms and seek out a temperature measurement. It also captures only a single moment in time. The Oura Ring, by contrast, records broader and more nuanced temperature fluctuations over time, potentially giving public health authorities an ability to see an outbreak coming before it strikes.
The U.S. military is testing a version of this system. The Defense Innovation Unit, in collaboration with the Defense Threat Reduction Agency, has paired a Garmin watch, the Oura Ring, and an analytics platform designed by Philips Healthcare to spot illness among troops before symptoms arrive. So far, it seems to be able to see illness coming two days beforehand. The program started with 400 participants in June but is expanding to 5,000, according to Defense One. Meanwhile, West Virginia University Rockefeller Neuroscience Institute is running a study that uses Oura to predict illness in healthcare workers.
To create a comparable system for the greater public, health officials would have to contend with privacy concerns. “I think there should absolutely be a public authority or even just a public record of the kind of physiological events that are happening,” Smarr says. “How you secure that and scrub individual information away from that is an interesting challenge.”
For now, using temperature as a beacon for public health status is merely in the research phase. Researchers at UCSF and UCSD have given the Oura Ring to 3,400 healthcare workers in various regions of the U.S., while Oura has enrolled more than 60,000 of its users. The results of the observational study are expected sometime next year.
Recognize your brand’s excellence by applying to this year’s Brands That Matter Awards before the early-rate deadline, May 3.
