Coronavirus in Context: Dr. Eric Topol Wants 100 Million Smart Watches Detecting COVID-19

[MUSIC PLAYING]

You're watching Coronavirus in Context. I'm Dr. John Whyte, Chief Medical Officer at WebMD. Everyone is talking about testing as a solution to get back to work and reopen the country. But is testing all it's cracked up to be? What are the challenges? To answer these questions is my friend and guest, Dr. Eric Topol. He's Executive Vice President of Scripps Research and the editor-in-chief of Medscape. Dr. Topol, thanks for joining me.

Great to be with you, John.

So everyone is saying that we need to do more testing, whether it's diagnostic testing or antibody testing. And we can break it down. But what are your thoughts about the role of testing? You've been somewhat vocal, uh, about it.

Well, John, I think it's clear we do need, uh, a big jump, exponential jump in testing. And I was part of the, uh, Rockefeller Foundation report that was just issued yesterday where we actually called for scaling up from 1 million to 5 and to 30 million tests a week.

[INTERPOSING VOICES]

--3 million by June, which is not far off.

Right. Well, we-- we have barely, uh, done that since starting, 1 million. You know, so we have-- we're so far behind. And this is a big problem. So the-- even the 30 million a week is a compromise because you have 330 million people, and you-- you can't just think that you're going do that once, because a lot of people, of course, are going to come out with a negative test, but they could subsequently become positive, but-- whether it's for the virus or for the--

Yeah.

--antibody. So we have a problem with not being able to reach such a large number of people, and also on a frequent basis. The other thing we have, John, is that the tests are not as good as we'd like for accuracy. We have a false negative problem for the virus. We have a largely a false positive problem for the antibody. But there's false everything, for both these tests.

How bad-- how bad is it, do you think? Is it based on the prevalence of the disease in the population, in terms of its positive predictive value and negative predictive value, those people who test positive who are truly positive, et cetera? What-- what do you think the number is? Some people are saying it's 70%. Is it that low?

You're bringing up a really important point for the sero positive antibody test. Because the prevalence is so low--

Mhm.

--probably, you know, 2% or something like that, or-- or lower, their chance of getting false positives are-- are higher. Uh, for the-- for the virus, um, RNA test, the false negative rate could be as high as 25% to 30%.

Wow.

The other key point on the testing is that there's so many different tests. You know, today we've learned, uh, from Yale of a saliva test would be much more palatable than a nasopharyngeal swab.

Yeah.

That looks much better, in fact.

Mhm.

But obviously, we need more data. And there's so many-- there's over 90 antibody tests, and there-- a lot of these are getting emergency use authorization from the FDA with limited data. And that doesn't help. So the-- the key point there is we have too many people spread out all over the country, we have to test them frequently, and we have shakiness in the tests. Now, with that background--

Mhm?

--we need another plan. We can't rely on testing. So in the Rockefeller report, um, there were three big initiatives. One is testing. The second is tracing, contact tracing. And we don't have the labor force to do that. We need to have that. The third is digital surveillance.

Mhm.

So the idea here is to use smartwatch, for example, that we have a big successful app called Detect. And that--

Tell us about that.

Yeah.

How's that working?

We launched that a-- a few weeks ago. We have, uh, you know, many thousands people on it, and we're collecting resting heart rate. Now, resting heart rate goes up before fever.

Mhm.

So we had already published, uh, back in the Lancet Digital Health Journal--

Mhm.

--earlier this year that we could predict the flu before the CDC, uh, and-- and at least as accurately.

Mhm.

So the point being is that resting heart rate is a really valuable marker. And that's something that over 100 million Americans have a smartwatch or wrist heart rate detector. So if we got a large number of those, uh, to use it, and to be part of the app, of Detect, which is a-- you know, an opt-in story, that all different types of smartwatches. Doesn't have to be, uh, Fitbit or Apple or any kind. Um, we could really get our arms around this. And that's just one tool, John. Uh, you could also go for body temperature.

Mhm. But--

[INTERPOSING VOICES]

Yeah.

--scale. Yeah. And you mentioned contact tracing. There's talk about hiring 100,000, uh, workers to do contact tracing. Is-- is that realistic? You've been a big proponent of tech. Uh, and how has--

Yeah.

--tech played into this? I mean, it's not one or the other. But where do you think we're going to get the biggest bang?

Well, the digital contact tracing is attractive, that it uses a smartphone chirps and be able to alert people that they've been in contact. Um, the only problem is, um, Singapore has the most experience so far. And even with the, uh, instruction of the leadership of the country, there's relatively limited use, and it doesn't-- it isn't clear, it isn't validated that it's as good as human contact tracing. We want to see that work well [INAUDIBLE] supplement. The key point here, uh, uh, John is that we got this thing for the next year and a half or two years, and we need methods that don't rely on slowness-- Mhm.

--and don't rely only on the human capital, uh, labor force, because that takes a while to develop. It's very expensive. But I have to say that, you know, the Google, Apple program for digital contact tracing, which is supposed to start, uh, in May--

Right.

--um, we have to get that validated. We have to do [INAUDIBLE] studies to show that it comes up with accuracy as-- as good or better than human contact tracing, which is the gold standard.

But don't we also have to get 60%, 70% plus to opt in? That's hard to do, for-- for any type of [INAUDIBLE].

Oh, right, so you-- yes, you need a large proportion, uh, of people willing to do it. It's not something you can mandate. Uh, there are obviously legitimate concerns about privacy, especially when you bring in the tech titans on something like this. But you know, one thing I think, John, that's worth emphasizing, um, if it's validated, the chance of being able to get back to work in a pre COVID-19 world, people might trade off some temporary privacy compromise in order for everyone to try to restore the way life used to be.

So these are the sorts of things we have to think about. But a tracing story, we're-- we're not prepared for that at all, as you've pointed out, John. So testing, we're not prepared, tracing, we're not. The only thing we're prepared for is we've got 100 million people out there with a smartwatch.

Right.

And all they have to do is get on the app and start to donate their data, and we can likely-- I mean, we have to prove it, but we can likely pinpoint where clusters of people with resting heart rate elevation is starting to appear before an outbreak has gotten to any, uh, significant extent.

Dr. Topol, what are other countries doing better? You've been tweeting a bit about what's happening in Germany. What are the differences?

Yeah, it's actually pretty remarkable, John. Uh, in Germany, they read, uh, our paper in Lancet, uh, about the-- the ability to predict flu from resting heart rate from a smartwatch. So they initiated a program, and within, uh, a matter of days, they now have over 300,000 German citizens who have donated their heart rate data continuously. And-- and they're extremely, uh, happy with this ability at essentially no cost to have digital surveillance throughout Germany.

Yeah.

And they're going to keep building the number of people on it. So this is, I think, a very attractive, alluring way that we can get our arms around it. Uh, and if it happens immediately, uh, in the US, we're fortunate. Over 100 million people have a smartwatch or a wristband that gets heart rate data. So it could be quite valuable for us.

Can we add a pulse ox to that so we get some sense of oxygenation? Is there value there? Some of them are, you know, in terms of where we are in technology, it's amazing. Well, you're right. Good point, John. So that some of the newer Fitbits and other watches do have a oxygen saturation capability. Um, and some have validation. Uh, the-- the problem is that, if you start having drops in your oxygen saturation, you're much more likely to be symptomatic, and have already, uh, had, um, some medical contact. You know, we're trying to pick things up really early.

But, uh, you're bringing up another point, which is using digital surveillance to keep people out of the hospital. And the-- and we need to be thinking about that, too, because a hospital is not a place to go unless you really are-- are pretty darn sick. But the problem with that, just to be clear, we don't know, uh, we don't have algorithms to say it's safe to stay at home, uh, because sometimes people suddenly, as Medscape has covered, there can be sudden demise. So we've got to also make sure we have that nailed, as well.

OK. Eric, where do you feel we are on antibody testing today?

Yeah.

You referenced it a little earlier, challenge, as you pointed out, their emergency use authorization, not a true approval for most tests. There's really only about three or so that have truly been approved. They're lower standards of accuracy. So are-- are we there on antibody testing, or should we not get too excited about it right now, and these certificates of immunity some people are talking about?

Actually, that goes back to Germany with the immunity certificate. But that is flawed, to some extent. And first point is no one has a good serology test that's been fully validated. The problem is there's four other coronaviruses that can cause common colds, and some people have high titers of antibodies, IgG, to those coronaviruses that cross-react to the coronavirus of interest, COVID-19. So we have false positive problems. It's a delicate assay. Um, there's over 90 of them that-- none of them have had, you know, a validation at scale.

Mhm.

And so the other issues are let's say you have a test that's 100% accurate, which doesn't exist yet.

Yeah.

Then you have, OK, how long does that protection last? Oh, by the way, if you have antibodies, could you still spread? Oh, by the way, could you-- might you reactivate, or potentially even re-infection? Although that hasn't been clearly demonstrated. So even if we had assurance that the test was fully accurate, we still have lots of unknowns about it. And we have to fess up about it. So just because you say, oh, you have a positive antibody test, well, it's not just 01. Some people have low levels of IgG against COVID-19, some have very high levels. So you know, it-- this is, uh, full of holes right now.

And you're always very practical. So let's be practical, if we can, for, you know, our last minute. So Rockefeller report, you talk about we need to do more testing and better testing, whether it's antibody testing, diagnostic testing, whatever. We need to do better contact tracing, whether it's hiring people or using tech.

And then we need to have these digital tools, which in some ways is surveillance, and also help with diagnosis. We can't do all three well. And you know, there's some element of resource. So I-- I want to kind of get your thoughts on where do we need to be spending more of our efforts? Not that other two aren't important, but if you had to focus more on one, where would it be? Would it be kind of on these digital tools? Because that's [INAUDIBLE].

Yeah, I think they've been grossly underemphasized. And the reason I say that is, in the Rockefeller panel of 20 some experts, none of them were particularly interested in the digital side. They were just arguing more about how many tests. But frankly, you know, we recognize that it would take six months to get to 30 million tests a week.

We can do the digital surveillance tomorrow if we had, you know, uh, a national support. And it's cheap relative to-- heart rate by-- by smartwatches is quite accurate. And again, it isn't at the individual level. We're looking for clusters of people where their heart rate is changing resting. So I am actually a proponent of that. Uh, but you know, I don't want to dismiss that the testing at scale, frequently.

Sure.

And the-- you know, both antibody and virus, uh, diagnostic, and the contact tracing. These are all important. But the critical issue, John, is we got so far behind this pandemic in the US, and now we're playing catch-up. And there's no great strategy. All the things we're talking about are-- are really, uh, difficult. The easiest one is, hey, you've got a smartwatch? If you don't have one, we'll send you one. Um, but let's get that heart rate data. It works quickly. It's passive. There's nobody's sticking a swab up your nose. You know, it's-- it's pretty-- pretty straightforward.

OK And it's a good point. And we can learn more about Detect at-- at the site. So thank you, Dr. Topol.

Hey, thank you. It's great to talk with you again, John.

And thank you for watching Coronavirus in Context. I'm Dr. John Whyte.