Coronavirus in Context: Can Wearables Diagnose COVID Before Symptoms?

[MUSIC PLAYING]

Hi, everyone. You're watching Coronavirus in Context. I'm Dr. John Whyte, chief medical officer at WebMD.

Do you wear a Fitbit? Are you interested in knowing whether or not you might have COVID? Well, today I'm joined by Dr. John Moore. He is the chief medical officer at Fitbit. John, thanks for joining me.

Thank you for having me, John. Great pleasure.

Now, I want to talk about this recent study that you've done, the Fitbit COVID-19 study. Now, I want to go over it, but what struck me is it says nearly 50% of COVID-19 cases-- you were able detect one day before onset of symptoms. And as we talk about the importance of asymptomatic, presymptomatic spread, this is big news. Tell us about the study.

Sure, absolutely. When-- when COVID-19 came on the scene, we immediately started thinking, well, what can we do to make a difference? And it was clear that monitoring temperature, monitoring symptoms, monitoring exposure to COVID-19, all really important pieces of the puzzle. Obviously, testing an important piece of the puzzle. But all of them suffer from the same limitation, which is prior to development of symptoms, you don't have a temperature, you don't have symptoms.

And the sensitivity of any testing, it-- not sure exactly where it is, but it's pretty well understood that it's very low prior to development of symptoms. And given work that we have done in the past with Scripps Translational Research Institute in the detection of flu-like illness, we had some good evidence to suggest that we might be able to-- to find signals earlier than symptom onset in the Fitbit data.

So we reached out to our community and we said, do you want to be citizen scientists? Do you want to help the cause and see if we can figure out a way to make a difference? Because we believe that transmissibility is really high prior to symptoms, so if we can detect it early, we might have a chance.

And-- and our Fitbit users came to the aid and over 100,000 people signed up for the study, agreed to share their data with us for the sake of the science. And over a thousand people had positive COVID test results and thousands of people had negative test results. And we had a whole spectrum of data there to help the cause.

And it also talks about that, uh, shortness of breath and vomiting were the symptoms most likely to predict that someone would be hospitalized. How did you include vomiting? Uh, 'cause last time I checked, I don't think Fitbit tells us [LAUGHS] that I vomited. So was there other associated aspects of the study as well?

Yeah, absolutely. So with this study, not only did we ask participants to share their Fitbit data, but we asked them, did you develop symptoms and when? When did those symptoms start and when did they end? And what symptoms did you have? Did you have headache, vomiting, sore throat, cough? All the standard list of symptoms, including loss of taste and smell, et cetera.

And we also asked, were you tested? What kind of test did you have? When did you have that test? And what was the result of that test? Also we asked, if you did develop COVID and it was confirmed, did you not need any treatment? Did you stay at home and-- and just, uh, take care of yourself?

Did you get treatment from the medical establishment? Did you need to be hospitalized? Did you need to be in the ICU? Did you need to be on a respirator-- or on a ventilator? So we-- we tried to get some proxy measures about the severity of disease as well. So then we could match up symptoms and other reported data with the outcome.

Now, the article's in pre-pub, correct?

Correct.

Now, how generalizable is it to the overall population? Some people might suggest that people who are using the Fitbit to begin with are more interested in health, are more aware of what's going on with their body. Wh-- what's your response to this?

Sure. Um, over the past decade, really, our demographic of Fitbit users has expanded to the point that it pretty well mirrors the population of the United States in terms of different demographics. Particularly age has been a big consideration in the past. People have asked, well, is Fitbit mainly younger, healthier people? Right now, we-- we pretty much mirror the demographics of the United States.

Now, there-- there is the chance of some selection bias of individuals who select into this study. Um, and I don't have a crisp answer for you on that in terms of did we see any major selection bias there. But from a COVID perspective, I don't expect that-- that parameters about people's likelihood to be fit affected their likelihood to join this particular study. So, uh, I think it's gonna be like our general Fitbit population that it reflects the general demographics of the United States.

Were there any differences that you know of based on age, gender, race, ethnicity?

No, we haven't-- we haven't explored that in depth yet. Um, the thing that we're excited about is we have quite large numbers, right? Over 100,000 people, uh, but a-- a thousand people who tested positive. In order to get into the multivariate analysis to look at subgroups, we-- we need to continue to collect more data. Uh, it's something we're very interested in, but we need more data to dive into those specifics.

So what's the future? What's the role of trackers, sensors in managing this pandemic as well as, you know, other health issues?

Sure. In terms of COVID-19, we think it's a piece of the puzzle, right? It's not something that stands on its own. The things recommended by the CDC and other public health authorities around doing symptom screening, doing, um, screening for exposure, temperature screening, all still important piece of the puzzle. But we know their limitations.

So this we feel can complement that. It might allow us to, one, make more cost-effective and operationally feasible usage of testing in a responsible way, because now we have some early cues. It might help people to make more responsible and safe decisions about their exposure to other people in periods even before they're tested and even before they develop symptoms. And it might help us to-- to, uh, create a-- an experience along the way where we get more and more confidence.

Because it's not infallible, right? As you said, about 50% of people we will catch a day before. But on the day of symptoms, it's a bit more. On the day after that, it's a bit more. So we have this opportunity to be with people on that journey and not just detect things early, but also inspire them into action.

OK.

The-- the data out there isn't clear, but a lot of people in my estimation, on day zero, on day one, they're scared. They're not sure. Uh, do I have a cough that might be COVID? Or is there-- am I just thinking about it? So having extra signals there I think can help inspire people into responsible action. Staying-- staying, um, physically distanced--

Mhm.

--wearing a mask, which they should be wearing all the time when they're out in public. But avoiding congregating completely during that period. So a lot of different opportunities to help the cause, I think, by complementing the traditional mechanism.

What specific [INAUDIBLE] advice do you have to current users? Is there something in particular they should be looking for? Is it, you know, a change in, you know, heart rate? Is it-- how are they going to interpret this in their day-to-day life that might make them say they should call their doctor earlier than they might have? Or they might go get a test earlier than they might have? Because as you said, if we can catch this earlier, we really can work on decreasing transmission.

Right. So at-- at this stage, we want to be wary about giving our users any-- uh, any particular directives on that front. Because we want to seek regulatory support on this first, before we go out to our consumer population and say, hey, use these metrics.

Mhm.

Uh, I'll take a step back from there and say, from a research perspective, what we're seeing is that resting heart rate typically increases. Um, it increases whether or not people have a temperature, which is nice. It increases the specificity there-- or the sensitivity there a bit. Uh, we also see changes in heart rate variability and breathing rate, which are two metrics that we don't currently show to users in the Fitbit app, but we're exploring what can we do next. And like I said, we'll be seeking the advice and the support of regulators around the world to figure out how we can make a difference on this.

OK. Now, I have to ask you-- I'm sure viewers are wondering-- what is this mural--

[LAUGHTER]

--about behind you? I-- is that a lobster? [LAUGHING] That's-- [INAUDIBLE]. Wh-- what's going on behind you?

Yeah, so that-- uh, that's-- the big mural is the Fitbit logo.

Mhm.

Um, prior to COVID, I worked predominantly from home, although I-- I travel quite a bit as well. So I-- I-- it's not just a Zoom background there. It's a-- a Fitbit logo up on my wall. My mom helped me hang it up on there.

OK.

And-- uh-- and the lobster there is a tribute to the fact that I live in Maine. And a big part of the-- the culture here is the lobster--

Where--

--industry.

Where could viewers go to learn more information, either about the study or, you know, about what Fitbit is doing in terms of sensors and trackers and, uh, in managing health?

Yeah, absolutely. Go to Fitbit dot-- dot com, and we have information there about the study. Um, if you're interested in participating in this, you can still share data and participate in-- in order to advance the cause. Like I said, we'd love to have more of a var-- a larger set of data so that we can look at changes in-- in age, race, gender, et cetera, so we learn more and more about this.

And yeah, please go and learn more and participate if possible. We think this is just the beginning, right? COVID is-- is one use case, but can we detect flu-like illness and other flu-like illnesses that might arise? Can we in the future detect things like depression and diabetes and hypertension early?

And then, not just that, right? That's just about detection and getting data. Can we inspire people into action and give them daily support to be effective in improving their activity, their nutrition, their sleep, and their mindfulness? Because that's what's going to combat diabetes and hypertension. But, as we know, it also puts you in a better health state so that if COVID-19 or the flu or other conditions come along, your reaction to it will be more favorable. So--

Well, hopefully we can check in with you along this journey, uh, to see the progress that we're making.

Absolutely.

And thank you for watching Coronavirus in Context.

[CHORD]