Centering on Coronavirus Podcast: Dr. Ashish Jha on COVID-19 Testing
Since the beginning of the COVID-19 outbreak, testing has been a topic at the forefront of our national discussion. Testing is a crucial component of any effective pandemic response plan, but the U.S. has faced some unique testing challenges with respect to both quantity and quality of tests. To learn more about the state of testing in the U.S. and our testing goals moving forward, as well as some effective strategies for meeting those goals, policy analyst Julia Baumel had the opportunity to speak with the director of the Harvard Global Health Institute, Dr. Ashish Jha.
Dr. Jha also practices internal medicine in Boston, and he teaches courses at Harvard Medical School as well as the Harvard T.H. Chan School of Public Health. In the fall, he will take over as the dean of the Brown University School of Public Health. Throughout this pandemic, it’s been his mission to share information about the virus and data-driven pandemic response guidance with a national audience of policymakers and citizens.
Julia Baumel: You’ve been working pretty tirelessly to share the facts about what is truly happening with this virus on every major news network from CNN to Fox, and you also recently testified before the House Select Subcommittee on the Coronavirus Crisis. Before we dive into the details of testing, I wanted to ask you about your experience sharing this important information in a hyper-partisan environment. As a doctor and as a scientist, do you feel confident that your message is getting through to policy makers and the general public?
Ashish Jha: You know, I do. And I find that most people are pretty receptive to information about things like testing, about coronavirus. And obviously a certain proportion of people take what I’m saying and try to put a partisan spin on it. But I have actually found that to be a relatively small number, a small minority.
And actually one of the few times where it’s happened was at the congressional hearing, surprisingly, when Representative Jim Jordan argued that my claim, which is widely accepted in the public health community, that our testing failures really were at the heart of a lot of the problems that the country got into. He called that a partisan attack, which I thought was surprising only because so few people I’ve encountered really think of testing as a particularly partisan issue. So that was a surprise, but I’d say other than that, there haven’t been very many instances like that.
JB: That’s pretty encouraging. So something that you and other public health leaders have worked hard to convey is that we still need more diagnostic testing for COVID-19 to reopen the economy safely. Your latest projection with the Harvard Global Health Institute set a new daily testing goal of 900,000 tests across the nation. Can you talk a little bit about that projection and how those results should be interpreted?
AJ: Sure, absolutely. So, you know, part of the question, when people say, well, how much testing do we need? Part of the way to answer that is to answer what are you using testing for? And why and how does testing help us keep our economy open? And, you know, there are different approaches to answering even that question. And of course, with different approaches, you get different results.
The approach that we took was we basically first try to make our best estimate of how many people are likely to be infected in the country on any given day. How many new infections are there likely to be around this time when states are opening up. And then ask the question, if we took a strategy where we tested everybody with even the mildest of symptoms and traced their contacts and tested those contacts. And first of all, that would be a pretty good strategy for keeping the amount of infections in the community low, assuming you could get everybody to isolate once they’ve been tested and tested positive. And we then did some calculations on, if that was our strategy, how many tests would we need across the country? And the answer was about 900,000 tests per day. And obviously, that varies a lot from state to state, partly based on population, partly also based on how big an outbreak the state has had. So states with large outbreaks need a lot more testing than states that have had relatively small outbreaks.
JB: You mention some variation between the states, and several states are starting to open up even though they haven’t yet built the kind of testing infrastructure you’ve recommended. Are you worried that we’ll see new outbreaks in these areas in the coming weeks and months?
AJ: Yeah, so when I think about the upcoming weeks and months, you know, the way I think about it is that there is no one secret sauce to keeping economies open and keeping states open, and the economies functioning and keeping people safe. I mean, it’s a set of different policies that you can put in. And testing—an adequate number of testing—and then doing tracing and isolation is a cornerstone. But it is not the only thing. And so if a state opens up with a lot of social distancing intact, universal mask wearing, let’s say their testing is somewhat inadequate, but still pretty decent and they had a declining number of cases before they opened, they can potentially get away with it. So the less testing you have, the more of everything else you have to do. I think of testing and tracing and isolation as, you know, kind of degrees of freedom. The more testing and tracing and isolation you can do, the more degrees of freedom it buys you in everything else. So if you open up with less testing, you’re just going to have less margin for error and less ability to do other things.
JB: Georgia is one of these states that started reopening a few weeks ago, and they’ve gotten some criticism for it. But some recent numbers suggest its COVID-19 caseloads might still be declining. What do you make of this? Do you think Georgia might have made the right choice, or is it still too early to tell?
AJ: Yeah, so there’s certainly some folks who think it’s still too early to tell. And these things can take several weeks and, you know, even four to six weeks between the time that you make a decision like this and you really see the effects.
It’s also entirely possible that those of us who worried about it—and it wasn’t just me, you know, Dr. Fauci, Dr. Birx, even the president said that he thought Governor Kemp was going too fast— it might be that there are some things about what has happened in Georgia that we don’t fully understand.
So, for instance, I think most of us believe that transmission outside of your home or outside of buildings happens much less efficiently than the transmission that happens inside. So maybe part of opening up is they’ve been able to maintain social distancing, gotten some people outside, and then there’s another kind of element which we’re still really trying to figure out, which is the effects of humidity and weather and temperature. And it may be that a combination of all of those things has let Georgia go faster than most of us thought it probably could.
You know, at the end of the day, there are two points that I sort of take from this. One is you want to be as data-driven as possible. But data-driven doesn’t mean that you’re going to make every prediction about the future correct. And second is you’ve got to let things play out over longer periods of time. And that’s why I try not to be overly prescriptive about what I think is going to happen over the next couple of weeks, because this virus continues to surprise all of us in some of the things that it does, how long it takes to manifest itself, et cetera. So good news on Georgia. I’m keeping my fingers crossed. But that said, you know, they’re not out of the woods yet.
JB: Yeah, that makes a lot of sense. The FDA recently approved the first antigen test for detecting COVID-19. Can you talk a little bit about the differences between this type of test and the more traditional PCR test, and how the antigen test could help achieve our broader testing goals?
AJ: Yeah. So, again, the big goal of testing, just so your listeners remember, is—the entire strategy of how to fight a pandemic, how to fight any outbreak, is—to keep separate infected people from susceptible people. Because infected people, when they come near contact with susceptible people, will pass the disease on and you’ll get spread of the disease and you’ll get exponential growth.
And so the question is, how do you do that? And obviously, shutting down the economy is one way, which is you just keep everybody at home. But the other way is to do, you know, really rigorous testing across large chunks of the population and try to really identify who is infected. And one of the problems of coronavirus has been there’s a lot of asymptomatic transmission, presymptomatic transmission, and that has meant that our testing strategies have to be much broader. We just have to be testing a lot more people. And the current way we test people is using something called RT-PCR, which is essentially looking for genetic material of the virus.
Antigen testing is something I’ve been very enthusiastic about, which is instead of looking for genetic material, it looks for antigens, or proteins that might be on the surface of the virus. And these tests are generally easier to do. They’re cheaper, they’re much faster. They tend to have certain types of error rates in them. Every test has an error rate. And the other thing is the antigen tests can be scaled up to millions of tests a day much more quickly than PCR can. So a lot of us have been waiting for antigen testing to come online. And the first one that got approved, it isn’t a great test. It’s not a super high quality test. I probably would not use it right now, but there are about 20 other companies that are working on antigen tests, and I’m confident one of them will end up having a much higher quality test and that will make it much more feasible for us to test millions of people a day.
JB: So assuming we do get something more reliable, would the efficiency of the antigen test make it the best bet for businesses that want to gradually start reopening? Say, for example, a medical center where outpatient surgeries are performed that wanted to test all patients for COVID-19 before they had their procedures.
AJ: Yeah, so there are probably two or three different new technologies that are in the works. You could potentially use the old one, the one we’ve been using so far, the RT-PCR. But I think antigen testing, there are some CRISPR-based technologies and tests that are coming down the pike—so there’s a few different opportunities where in ambulatory surgical centers, in hospitals. You can even imagine as air travel starts getting back up and running and people are going to want to feel more comfortable getting on an airplane. You can imagine that airlines could require that people show up a half hour early or an hour earlier than normal and get an antigen test before they board the plane with the idea that if they’re negative and if it’s a highly sensitive test, everybody’s gonna feel much more comfortable flying on that airplane. So I see all sorts of uses for this kind of rapid antigen test once they become higher quality and more reliable.
JB: That’s all pretty encouraging. At this point, what are the major barriers preventing us from scaling up diagnostic testing to where we need it to be, and what are some ways states and the federal government might work together to overcome them?
AJ: Yeah. So, you know, we have been making progress on diagnostic testing again through the RT-PCR. We were, about a month ago, we were at about 150,000 tests a day and now we’re up to about 300, 350. Again, our personal analysis—not my personal analysis, but our group’s analysis—is that that number should be closer to nine hundred thousand. And then there are others like Paul Romer, former Nobel laureate—or economist who has argued for like 30 million tests a day. So 30 million tests a day, you can’t do using PCR, but you could get to nine hundred thousand, theoretically, people being tested a day using PCR. But you’re going to require a very kind of aggressive look at the entire supply chain, ramping up various supplies, ramping up a testing strategy that lets you get to that nine hundred thousand number.
States are doing this on their own right now because the federal government has largely been absent from this entire process. They have a testing coordinator who is helpful on the margins. But what I hear from states is, and even what you I think hear from the administration in some of its policy briefs is a sense that, you know, they want 50 states to all go out and do this on their own. And the problem with that, and this is not really a federalism issue, you know, what’s state versus federal responsibility. This is just a practical issue. The supply chains for these tests are national and international supply chains. And so what you then have is states competing with each other, states competing with hospitals, states competing with other countries. And, you know, my take has been that big states like California might be able to figure it out on their own, but smaller states like South Dakota or Montana, they’re going to have a much harder time playing in an international supply chain, trying to outcompete, let’s say, Germany for some critical supply. It’s not really clear to me that that’s what we want our states to be doing, let alone competing with each other. So I wish the federal government was more willing to take a helpful role here. But states are kind of plugging along and trying to make, you know, solve all these issues and they are making modest progress.
JB: Now I want to shift gears and talk a bit about antibody testing, where there are concerns about accuracy and these tests potentially giving people a false sense of security. At this point, who should consider getting tested for antibodies, and how should these people interpret their results?
AJ: That’s a good question. So antibody testing is getting better. And the problem with antibody testing is not quantity, it’s been much more quality. What happened initially with antibody testing is, because the diagnostic testing for the virus was such a kind of calamitous failure all through February and March, the FDA’s approach on antibody testing was very permissive in March and April where basically almost everybody could get some sort of an emergency use authorization or some way to be able to get it out into the community. And what that has meant is that there are a lot of very poor quality tests. And there are a few very high quality tests, but there are probably about 70 or 80 antibody tests out there, and my sense is probably three or four of them are worth using and the rest are not.
And so every time there’s a new result of an antibody test, my first question is what test did you use and what’s the test characteristics of that test? The ways that I think anybody tests are really useful—and again, we’re going to assume that you’re going to use one of the really high quality tests—is first and foremost, as a public health person, I love it because it really teaches me a lot about how much has the disease spread in a community, how many people have been infected that we didn’t know about. We know that because of our limited testing capacity, a lot more people were infected than were tested positive. And so it really tells us a lot about the disease spread, in which communities, and which types of patients.
But the second is this idea that maybe it allows some people who’ve tested positive to then feel comfortable that they’re immune from reinfection, and maybe that changes what they can do, what kind of business they can do, how quickly they can get back to work. And this is the idea of the immunity passport that a lot of folks have been pushing. And my caution on that is, first of all, you want to make sure you’re using a really high quality test, because even tests that sound pretty good on paper, like a ninety five percent specificity will mean, in a low prevalence population if you apply that test, half or more of the of the positives will be false positives. So you don’t want to be telling people that they have a positive test when there’s a 50 percent chance that it’s a false positive. But the second is, and just as importantly, you want to make sure that even if you have a true positive result—you have antibodies—you want to be sure that those are truly protective. Because if not, you’re going to give people a sense of confidence that they can go out and do things that then can get them infected again.
Now, this is a place where the science really has changed, I’d say even in the last three weeks. Three weeks ago, I would have said not everybody who was infected generates antibodies and antibodies may or may not be protective. The science here in the last few weeks has evolved so that it looks like almost everybody who’s been infected and clears the infection does develop antibodies. And it looks—and we haven’t really nailed it down—but it does look like those antibodies should be protective. But I think it’s probably another couple of months away before we have the science that’s going to be comfortable enough that I would recommend to a patient that they think of themselves as immune, just because they have antibodies. I just think we’re close, but we’re not quite there yet. And of course, the cost of getting it wrong is very high because if somebody thinks they’re immune and they’re not, they might get infected again and get sick.
JB: So like you said, there are still so many unreliable tests on the market, but also a handful of good ones from private labs like Abbott and Roche with specificity rates close to 100%. How might breakthroughs like these change the antibody testing conversation?
AJ: Yeah. So first of all, I do think you really need to be using tests with a specificity greater than 99 percent. If you do the math, you see very quickly why it is. But if you have a low prevalence population and if your specificity is worse than 99 percent, you’re going to just get a ton of false positives. A large proportion of people are going to be false positive. So you don’t want that. But assuming and I do think that, again, there are a few tests now which are high quality enough that they can be used and you can feel confident that if you get a positive result on that test that you really are positive. Those tests, I think, will help a lot, again, in helping us both understand the level of the spread of the disease, but also giving us some better sense of, you know, what’s happening in specific communities. Or if you want to test your employees and know which ones are positive, I think it can be helpful. The key, again, is don’t assume that their positive means that they’re necessarily immune. I think they probably are. But, you know, probably is not probably good enough for people to act differently based on that.
JB: How far can research get us in terms of understanding exactly how immunity manifests? In other words, do we just have to wait and see how it plays out for people who have recovered? Or is there another way to get insight on immunity faster than that?
AJ: Yeah, so there’s a whole big debate about how to quickly understand what level of immunity it is and, you know, obviously if there were no ethical constraints in our lives or just there were no such issues, one thing you could do is you could just try to reinfect people who’ve been previously infected and cleared the infection.
And these challenge studies are actually being proposed. Obviously, you’d never do this without people’s consent. But there is talk of taking people who have recovered and then actually exposing them to the novel coronavirus and seeing if they get reinfected. I have, as you might imagine, a lot of concerns about that. There are real ethical issues. And if you’re wrong, these people might end up getting very sick or even dying. And so any experiment that, you know, exposes people to severe risk has to be thought through very, very carefully with a lot of consent and a lot of discussion.
There are other things that people are proposing, which is, of course, to track high-risk people over time. So let’s say physicians and nurses who’ve recovered. If you track them over time, because they’re being exposed a lot more than the average person, they may be somebody who if after a long enough period of time, none of them have gotten reinfected, you may feel pretty confident that they really have immunity. So there’s a lot of efforts, I think, that will get us much further along. I suspect in the next two, three months we’ll have a lot more information and a lot more confidence about this than where we are today.
JB: That’s pretty exciting. So you’ve mentioned using antibody testing as a public health professional to do some really useful analysis at the population level. But like you said, the false positive rate is going to be pretty high if not enough of the population has already been exposed. At what point in this outbreak do you think that type of analysis will be meaningful, and how should it be used to guide policy decisions?
Yeah, so first of all, it’s always helpful to know how many more people have been infected than you’ve identified. And there are a lot of studies now that have tried to quantify that question. And overall, my synthesis of the literature is, kind of in my head, as I read through all the studies, is that we’re finding that about 10 to 20 times more people have likely been infected than have been identified. So in New York City, if you just look at the straight up infection numbers, it looks like it’s about one and a half percent of the population. But then if you look at the antibody test, it’s about 20 percent of the population. And so that’s helpful because if it’s 10 to 20 X, you can look at that for the country. So for America right now, you know, as of today, May 19th, about half a percent of the population that has been found to be infected. And my best guess then is it’s probably five to 10 percent of the actual American population that’s been infected. Well, that’s useful. It tells us a couple of things. One is that we’re nowhere near herd immunity. Some people talk a lot about, you know, could we reach herd immunity? Yeah, that’s at 60, 70 percent of the population being infected. Right now, we’re probably five to 10 percent of the population, and already ninety thousand Americans have died.
So we don’t want to try to herd immunity strategy. But some cities like New York that are at higher levels, 20, 25 percent, it tells you that, there’s a pretty broad set of people with immunity there. And that’s useful to understand from what the economic implications for New York are. And obviously, the last part is the antibody testing tells us which communities have been really hard hit. And so as we think about the future, we really need to design policies to make sure that, for instance, minority neighborhoods, poor neighborhoods which have really borne the brunt of this disease, are protected much more in the next waves that will come down the pike.
JB: Yeah, absolutely. Before I let you go, is there anything about any of these types of testing that I didn’t cover that you think would be important for the general public to know?
AJ: You know, the thing I would say about testing is it’s obviously gotten a lot of attention. Part of the reason it’s gotten a lot of attention is because our approach to testing has been so abysmal, certainly compared to most other, or many other, high income countries. In some ways, it’s gotten too much attention. Ironically enough, and I realize I’m part of the problem because I’ve been raising it a lot. But the reason I would say it’s gotten too much attention is this should have been done day one. This should have been done early. Testing should not be a major constraint, a well-functioning public health system and a well-functioning government should have ensured that we had plenty of testing because testing is just a step towards disease control. Because then it’s testing, then you’ve got to trace people, you’ve got to isolate them. You’ve got to come up with strategies for how you do supportive isolation so people are willing to take time off of work and be isolated. How do you protect nursing homes? How do you do appropriate social distancing?
There’s a ton of issues that tend not to get enough attention because we haven’t sort of fixed the first order of business of just getting testing going. And so part of it is I look forward to a time—and I do think it’s going to happen—when testing constraints won’t be the single biggest issue that prevents us from really being able to handle the outbreak and that we can then move on to the more interesting and complicated challenges. But testing has been such a debacle, again until very, very recently, that has really made focusing on the entire pandemic much, much harder.
JB: Well, fingers crossed that we’re able to move on from testing to those other big steps pretty soon. And that’s all I’ve got, so I just want to say thank you so much for your time. This has been really informative, I really appreciate it.
AJ: It was my pleasure, and thanks for reaching out and talking to me about all of this.