Read on Twitter
Super excited to announce our COVID wearables paper that came out online in @natBME today! We show that consumer #Wearables can be used to detect #COVID19 at a #presymptomatic stage. #Covid19research #EarlyDetection @SnyderShot @StanfordMed
@aametwally1
https://www.nature.com/articles/s41551-020-00640-6
1/
@aametwally1
https://www.nature.com/articles/s41551-020-00640-6
1/
Thanks to all co-authors and collaborators, @Stanford University IRB, Privacy Office, and to all the participants for providing their data. We enrolled 5,262 (now > 6000!) participants, but such is the nature of human research - could only get complete data from 32 of them!!
2/
2/
We developed two "offline" algorithms to identify HR & steps anomalies & used them to detect COVID-19 in 26/32 (81%) of individuals. We detect COVID-19 a median of 4 days before symptoms start. We also ran these algorithms on non-COVID-19 infections & detected 14/15 of those.
3/
3/
Median detection was 2d ahead of symptom onset for non-COVID-19 illnesses. We also report alterations in sleep duration pre-vs. post-detection. Finally, we developed a real-time 'online' detection algorithm + two-tier alarming system that can detect HR anomalies in real time..
4/
4/
...Initial warning alarm is yellow, turning red after 24h of persistent elevation. 63% of COVID-19 cases have an alarm on/before symptoms & 4 cases 1 day after. Some cases had minor non-COVID HR elevations that triggered the alarms, including very intriguing "holiday bumps".
5/
5/
We compare alarms in 32 COVID-19 cases vs. 73 healthy individuals. Alarms are more frequent after COVID-19 illness than before. They persist longer in individuals with COVID-19 compared to healthy folks and alarm signal (peak height) is also higher.
6/
6/
We plan to enroll more people to validate our findings with an increased sample size + clinical testing for COVID-19 & other respiratory pathogens. Using our custom study app, MyPHD, we will send real-time alerts & notifications to study participants about HR elevations and..
7/
7/
...ask for feedback - could it be stress, exercise, menstrual cycle, meds, travel, or anything else? The goal is to develop a system that's encourages individuals to be more cognizant of their own health without scaring people or adding unnecessarily to the healthcare burden.
8/
8/
The scope of this is huge! Imagine doing this at mass scale for:
(1) individual consumers (patients) and physicians to track/monitor health
(2) augmenting employee testing programs for workforce safety, &
(3) community public health surveillance to identify emerging hotspots
9/
(1) individual consumers (patients) and physicians to track/monitor health
(2) augmenting employee testing programs for workforce safety, &
(3) community public health surveillance to identify emerging hotspots
9/
Employed at mass scale, we could use wearables to identify potential clusters of outbreaks *before* they get out of hand. Lots of validation work still to be done, but the promise is there. Come join us in our effort to fight COVID-19 at:
https://innovations.stanford.edu/wearables
10/
https://innovations.stanford.edu/wearables
10/
Also need to think about equity of access for underserved communities, who has access to devices vs. who needs them the most (HCWs, other workers), identifiability, privacy & security & tech governance.The @wef is doing a fantastic job at starting some of these conversations.
11/
11/
