Previous: Antibody Tests, Lockdowns, and Why Isn't This Working? Coronavirus Q&A 5-2-2020
Next: Lessons from Singapore and Reopening The United States



View count:20,855
Last sync:2024-05-18 09:00
Testing for coronavirus has been one of the most contentious aspects of the pandemic response in the United States. This week we're talking to Nobel Prize-winning economist Paul Romer, who has developed a plan to roll out Coronavirus testing on a truly massive scale. We talk to him about what testing on this scale would look like, and what it would mean for the way we live now.

Related HCT episodes:
When Can We Get Back to Normal?

Be sure to check out our podcast!

Other Healthcare Triage Links:
1. Support the channel on Patreon:
2. Check out our Facebook page:
3. We still have merchandise available at
4. Aaron's book "The Bad Food Bible: How and Why to Eat Sinfully" is available wherever books are sold, such as Amazon:

John Green -- Executive Producer
Stan Muller -- Director, Producer
Aaron Carroll -- Writer
Mark Olsen – Art Director
Tiffany Doherty -- Writer and Script Editor
Meredith Danko – Social Media

#healthcaretriage #Coronavirus #Covid19

 (00:00) to (02:00)

CARROLL: The United States is averaging about, what, a million coronavirus tests a week? And experts know we have to do a lot more than that. But how much more?

Nobel Prize-winning economist Paul Romer is our guest this week on Healthcare Triage, and he's gonna talk to us about perhaps the most ambitious plan yet, but it's one that might get Americans back to work. That's the topic of this week's Healthcare Triage.

Americans are worried about the economy and how much the shutdown is costing us, and they're worried that we might have to go back into another shutdown in the future. Paul Romer thinks the way around that is a huge test and isolation strategy far beyond what most people have proposed.

He's got his critics, but he's proposing that we ramp up to do more than 150 million tests a week, basically testing every single American every two weeks. Sure there are critics of this plan, but we asked him to help us go through it so we can learn more about it.

ROMER: I'm trying to provide a roadmap which is more focused. I'm trying to just say there's a critical thing we've got to do, which is to test people and then isolate the people who are positives.

CARROLL: But we should be clear, I think the testing that you're asking for differs dramatically from the testing that pretty much every other plan is asking for.

ROMER: That's right. So it's testing at a much more extensive scale. I feel like in many of these plans, even though they all say testing is an important part of this, in truth, I don't think they take the testing very seriously. Because if they did, they would start talking about testing at a scale that would really make a difference.

CARROLL: So to be clear, Paul is saying that we should test every American every two weeks, and while that sounds crazy to some, some experts are saying it's not enough-- that we should be testing every American weekly.

ROMER: I talked about the 14 days. Every two weeks. This would be enough with a certain false negative rate and a certain false positive rate. There's a paper out there where people are saying, "Oh no, Romer's made a mistake, he's wrong, and you can't do it with every 14 days." I've been meaning to get back to these guys.

There's actually a very kind of subtle mathematical issue that we need to resolve. (2:01)

 (02:00) to (04:00)

ROMER: It took me a while, and this is like computer code and equations. I'm now confident that I had the gist of it right, but there's still a technical discussion where we need to resolve will every 14 days do it or not?

CARROLL: Honestly, once you've made it to where you're testing everybody every two weeks, is every week more of a lift? Regardless, it's clear that to make this plan work, we'd need a massive amount of testing. Way more than we've currently got going.

So what exactly might that look like?

ROMER: That turns out to actually be an interesting question, and there's some mathematical modeling I've been doing and going back and forth with people. You can take 1/14 of the population that you pick at random and just do a random fourteenth of the population every day.

Another would be to divide the population up into slices, so you go through the first slice, second slice, everybody gets tested every 14 days. There's a little bit of efficiency associated with that test every 14 days on a cycle approach, in terms of-- you would get a lower value for this reproduction number, R, if we do it that way rather than just testing at random.

CARROLL: We should be clear here that we're talking about PCR tests, which tells you who's currently infected. That's different than serological testing, which looks for antibodies and tells us about who's been infected in the past.

Both of these tests have their uses, but serological testing takes more time to turn positive and isn't really helpful when we're trying to rapidly figure out who has the disease now and who might have been infected by them.

ROMER: The antibody tests are almost useless for the purpose of getting R below 1 because you don't get to 50% seroconversion until like 12 days after the first symptoms. So those tests are positive way after the horse has left the barn on infecting colleagues.

They help on surveillance purposes, they're not gonna help you filter and screen the population.

CARROLL: Of course, the biggest question on everyone's mind is, "If we can barely pull off like 150 thousand tests a day, how are we possibly going to ramp up to more than 150 million tests a week? (4:09)

 (04:00) to (06:00)

CARROLL: Moreover, how are we gonna have enough equipment to do that? Swabs, reagent, all the other stuff that might go with it? How do you move all of this stuff around?

ROMER: I think the basic answer is nobody's taking testing seriously right now. So if you don't take it seriously, you know, you don't have very much of it. We need to take it seriously.

You're close to, really, a hundred billion a year to do the kind of testing that-- If you could do it for $10 a test, you're basically at a hundred billion a year in annual expenditures. And my attitude is this is just something we start doing on a permanent basis and it just becomes a fact of life.

As you're well aware, our whole medical reimbursement system suffers from some peak(?~4:48) problems, many of which are just intrinsic to the economics of insurance.

So when I talk about going to 25 million tests a day, I'm thinking about an entirely separate system of tests, a different type of test than the clinical tests that an insurance company might reimburse a health provider for administering. This is a completely different system of tests, a different set of suppliers.

Think of a regulatory regime where somebody like the state of New York says, "We need a bunch of tests. Do those tests for us. We want this false negative rate, we want this false positive rate. You figure out the rest. You wanna swap out reagents, wanna do saliva, you wanna do swabs, you figure all that out. We just care about the false negative and false positive rate."

When you run into a bottleneck, "Oh we can't get the swabs!" You know, these labs, this is what they do all the time. "Okay, we don't need swabs, we'll just switch to saliva." This is what Rutgers has already kind of figured out to do.

We need that kind of responsiveness from this university lab side to scale up and work their way around any of these bottlenecks.

CARROLL: Okay, so Paul's asking for a loosening of regulations and an increase in the number of laboratories to get that done. (6:04)

 (06:00) to (08:00)

CARROLL: But even if we could, other problems remain.

Some other problems could be noncompliance, or even worse, wrong test results. What do we do if tests are falsely negative or falsely positive?

Falsely negative is a big concern 'cause it means people are walking around out there and still infectious. Falsely positive means we're gonna be isolating a ton of people who don't need it, especially when we're testing 20 million people a day.

Luckily, Paul's thought about this, and he's done some pretty sophisticated mathematical modelling, and he walked us through what that might mean.

ROMER: So this is where you need to do the math carefully, and the modelling I've done, we've allowed for things like a certain false negative rate, like, you know, 20-30%, even, false negatives. We also allow for noncompliance. It could be that only 90% of people who should be getting tested agree to get tested.

You can still get to this critical threshold of the reproduction rate for the virus, R, this reproduction number, R. You can get that below 1 with a reasonable number of false negatives, like 30%, and 80% to 90% compliance.

But the way you do that is by retesting more frequently. If you think about it, if you've got a 30% chance of a false negative today, you do another test tomorrow, the odds of two false negatives turn out to be about 1%.

If you're starting from a position where everybody is in quarantine anyway, then false positives are not really a problem because you're already being treated like a false positive. So false positives are not really the concern. The false negatives are.

This whole thing works, basically, when you catch somebody who's infectious.

CARROLL: Okay, so let's say we get through some of the logistics. Get these tests made. We even figure out how to get all the testing equipment and everything else.

What about the logistics? How do we get 20 million tests done to know who's got them and then bring them all together and send them where they need to be? (8:01)

 (08:00) to (10:00)

Manpower, physical space, where do we store them? Where does the data go? How do we keep track of it? It's a huge undertaking, and luckily, Paul walked us through that too.

ROMER: You know, imagine a world where we just bought the sodas, you know, flavored soda water at a drug store where the druggist made it by hand. We'd come up to the counter, we'd pay a nickle or something, you know like long ago. And then you ask yourself "could we actually scale this up so we were producing 350 million sodas a day?" 

Well, you know, lo and behold, we did it. And there's a bunch of logistics for managing this, but this country makes 350 million sodas a day. People supply those because we pay about 45-50 billion a year for sodas.


ROMER: So the first threshold, I've been saying, is like we should- if we pay as much for tests as we pay for sodas, we can get, you know, many millions a day.

CARROLL: And Paul's had his share of critics about this plan, and he doesn't take that lightly. He's thought pretty hard about what others say. But he talked to us about the alternatives if we don't do this massive level of testing.

ROMER: I've taken very seriously what the other policy options are, and not just in the "say the right things that people wanna hear", but things that could really work, and we're gonna stick to them and do them. 

I think there's only two other options besides the "test and isolate". One is what people have sometimes dismissively referred to as "herd immunity". What it means is you just let the virus proceed through the whole population. Maybe you slow it down so you don't overwhelm your hospital system.

But eventually you get to the point where say, you know, like 60 percent, or probably more. But 60 percent of the population's infected, and then the virus dies out cuz it has trouble finding more people to infect. What this virus, I think right now, a rough estimate of the fatality rate, the infection fatality rate is about 0.5 percent. So infecting about 60 percent of the U.S. population would be maybe about a million extra deaths. 

 (10:00) to (12:00)


ROMER: So this a really horrific, you know, possibility to contemplate. And now let's take the other one. We know that with lock down, we can get to R less than 1. That means that however many infections we got, they're starting the exponential decline. So you're trying to protect a large population that's majority susceptibles from an outbreak of this exponential growth of the virus. 

And this takes constant effort. If you go down this path of suppression and, you know, declining numbers, you have to be ready to stick with that forever. And you know, maybe in 18 months/ 24 months maybe we get a vaccine, maybe we don't. If lockdown's the only way to do this and it's costing us 500 billion, you know, 20 percent of GDP lost every single month, and we might have to do this for, you know, years on end. That's a really horrific alternative as well. 

So it's either you kill a million people or you have a catastrophe that's worse than the great depression. And it doesn't just make us poor. That may, like, destroy our democracy. I mean, think about the political and social upheaval that we had during the great depression. If you'd asked me, uh, two months ago "Is it expensive? Is it hard to get to 25 million tests a day?" I would've said "Oh man, that's gonna cost you a fortune, and that's a lot of work." 

But now when I think about, okay kill a million people or, you know, lose 20 percent of GDP and risk the destruction of democracy and our entire social life. Boy, if those are my other choices, 25 million tests a day sounds like a walk in the park. 

CARROLL: Walk in the park or not, a lot of people aren't gonna like the economic implications here. But Paul isn't proposing this to make us feel better. He's doing it because he thinks it's objectively better to the alternative.

When it comes to academic experts and policy makers: cold, hard facts are going to be a much better alternative than things that make us feel good but really won't help. 

ROMER: Don't be in this mode of saying things that will make people feel warm and fuzzy about lock down and so-forth.

 (12:00) to (14:00)

You know, like let's just be cold and hard and objective here. And I think if I can get my colleagues to take these two paths seriously, I think they're gonna come around and see "Yeah we get it. In absolute terms, what Romer's talking about is kinda hard, but in relative terms, it's a piece of cake. And we just gotta do this because if we take the other alternative seriously, they're just horrific."

CARROLL: So beyond the criticism of not testing enough, we wanted to know what kind of feedback Paul was getting from other epidemiologists and experts acknowledging that he's an economist and this really isn't his area of expertise first and foremost. 

ROMER: There is a little bit of, kind of, grubbling 'cause I haven't produced a paper yet to justify what's going on on this issue of the 14 days versus the 7 days. The only thing I think I bring as an economist to this, other than a willingness to just like do the math, is they were living in this world where nobody payed any attention to them. And they were saying, you know, "this is really important. We gotta spend you know maybe 100 thousand dollars, 200, quarter of a million- It would really help if we did that."

And then Romer comes waltzing in and says "Oh, listen. You know, we gotta just spend 100 billion and it's simple. And let's just go." I don't blame them for being a little bit miffed. But at the end of the day, we all wanna just solve this crisis so, you know, we'll sort that out. 

CARROLL: So there you have it. Detail on Paul Romer's plans for how we might reopen America and get back to business. And if it seems impossible, he'd remind you that lots of things in America seemed impossible before we accomplished them. And if it seems expensive, he'd remind you that we spend way more money to monitor and avoid other threats to America all of the time.

Now I don't think he's wrong. The current shutdown like trillions of dollars and the idea of spending billions, if not hundreds of billions of dollars, to avoid this happening again seems like a reasonable investment. 

Hey if you found this episode useful, you might find these other episodes useful where we also answer your questions about Covid 19.

 (14:00) to (14:30)

And if you can, like and subscribe down below, really helps. And also support the show in anyway you can through It's harder than normal to keep this going in the pandemic, and again as much as you can, we would really appreciate your support. It helps to us keep this coming out.

We'd especially like to thank our research associates Joe Sevits, James Glasgow, and Joshua Gister, and of course our Surgeon Admiral, Sam.