Current research suggests that a person infected with COVID-19 can infect others as early as three days before they begin to experience symptoms. In fact, the most contagious period likely occurs during the 48 hours before they start to feel unwell. A new digital platform that works with an Oura Ring health tracker can now predict a diagnosis of COVID-19 three days before symptom onset.
The West Virginia University Rockefeller Neuroscience Institute (RNI) created the new platform in collaboration with Oura Health. The Oura Ring is a sleek, discrete wearable that looks like a piece of jewelry. It combines advanced sensors, an accelerometer, and a gyroscope to track heart rate, HRV, steps, body temperature, and sleep. Oura Health partnered with the University of California San Francisco for research into symptom prediction at the start of the Coronavirus crisis, which Health Tech Insider covered in March.
The new platform created by RNI involves a mobile app that allowed participants in an initial study to track stress, anxiety, memory, and other metrics. This information is combined with data from the Oura Ring, including activity patterns, body temperature, respiratory rate, and more. AI algorithms then analyze the comprehensive data for patterns that suggest early symptom emergence.
A study that included 600 healthcare professionals and first responders showed that RNI’s platform forecasted the onset of COVID-19 symptoms three days in advance with more than 90% accuracy. RNI is partnering with Thomas Jefferson University, Vanderbilt University, and other institutions for a second trial phase, which will also incorporate location information that could help identify outbreaks and assist in contact tracing.
Promising results like these suggest that consumer wearables could impact public health in the pandemic era. The platform could lessen presymptomatic spread among healthcare and essential workers, first responders, and communities. And individuals willing to invest in an Oura Ring could potentially use the RNI app to prevent spread among household members by allowing them to self-quarantine in advance if infected.
I am suspicious of a a body temperature reading taken on one’s finger so I did a simple experiment. I taped a 1 cm square x 2mm thick platinum resistance thermometer element (Pt-100) to my left index finger near the base, as shown in the Oura advertisements. The readout was a Transmation Model 1074 Universal RTD Calibrator, battery operated and hence portable. All measurements are in deg. F. I chose not to use a thermocouple because of difficulties maintaining an accurate reference.
1. At a room temperature of 66.0, the RTD showed my oral temp.to be 98.0 Pressing the element in my armpit I got 95.5
2. I then taped the sensor securely to the base of my left index finger on the palm side with a 3/4 in. wide Velcro strap. The reading was 82.4.
3. Stepping outdoors (no jacket) where the ambient temperature was 54deg. I got readings between 75 and 78, the lower readings when a light breeze was blowing.
4. Returning indoors, I reproduced the earlier readings.
The readings on my finger therefore had a range from 75.0 to 82.4, a variation of 7.4 deg. depending on the momentary environment, while my internal body temp. remained constant to within <1 deg.
Therefore while some other date taken by the ring may be useful for diagnosis, I remain extremely skeptical of the Oura's temperature measurements.
Good points, Jonathan, and there’s nothing quite like making an empirical case! On the other hand, I infer that you are enough of a scientist to understand the perils of anecdotal evidence, even when supported by numbers.
The fact is that I don’t know how Oura measures body temperature. I don’t know what the thermal effects of the ring housing might be. I don’t know how much they massage the sensor data using algorithms based on thousands or millions of correlated data points. Based on what I know about the role of data processing for other wearable sensor platforms, there’s some deep math magic going on in many of them. I have learned not to underestimate the ability to extract a useful signal from what appears to be random noise.
As a result, I’d encourage you to send your message above to the researchers who relied on the ring for biometric measurements and see what their view is on the problem you present. I expect that they will have a far-more-informed understanding than I have.
All the best,
Alfred Poor
Editor, Health Tech Insider