Coronavirus (COVID-19): Press Conference with Michael Mina, 01/22/21

You’re listening to a press conference from the Harvard School of Public Health with Michael Mina, assistant professor of epidemiology. This call was recorded at 12:00 p.m. Eastern Time on Friday, January 22nd.


MODERATOR: Dr. Mina, do you have any opening comments you like to make?

MICHAEL MINA: Yeah, well, I just want to implore anyone who’s willing to write about it. There has been continued incredible lack of understanding of rapid testing. Now, Biden is clearly making these rapid antigen test programs part of his agenda in terms of how to combat this virus. I think it’s an exceptionally good idea, obviously, I guess, because I am the biggest advocate of it. But one thing we are contending with is a continued, very unfortunate series of research papers continue to come out where unfortunately the scientists just aren’t understanding exactly how to evaluate these tests, and the media is jumping on to these and repeating the conclusions of the science. And this is usually good. You know, the media is not in a position to necessarily question the scientists creating the research. But in this case, given what is about to happen, I’m extremely concerned that scientists will continue to, unfortunately, evaluate this wrong. And it’s really easy to evaluate them wrong because the FDA is essentially suggesting the wrong way to evaluate these tests. And so it’s hard to blame the scientists who aren’t experts. But the point is, every time there’s a new paper that comes out and from the CDC or from the Lancet or whatever, that says that these tests are very low sensitivity, which I’m sure that anyone has followed at all, has read these concerns. But the studies are not correct. They’re just not correct. They continue to evaluate these tests based on the wrong metrics. They’re finding that these tests are low sensitivity to detect people after they are infectious. And this test right here, which Biden is going to make a big part of the agenda to tackle this virus, is intended to detect infectious people, not post infectious people.

And so I want to use this moment to ask that if you’re about to write a report on these rapid tests and you see a new study from the CDC or anywhere else that says that the tests are faltering, especially in asymptomatic people, that they’re not working. Well, please recognize that it’s very, very, very likely we’ve seen it over and over and over again, that the tests are actually working extremely well. But the scientists are making the wrong assessment of them because most people are not experts. Most people who are studying these are just simply not experts. It’s a physician somewhere who got their hands on a thousand tests and wanted to use them. The reason this is so important right now is with the U.K. government, this exact issue happened, and some papers came out suggesting that these tests were only catching 30 percent of infected people and therefore, the U.K. government was putting their citizens in harm’s way by using these rapid tests. Turns out in those exact same studies that show that these were 30 percent sensitive, they were over 95 percent sensitive to catch infectious people, which is all that matters for public health test. But it led the U.K. government to put their testing program on hold. And frankly, it’s been leading our government or at least state governments to really back away from what the utilization of one of the most important and potentially effective public health tools we have at our disposal. So if you come across a paper that says that these tests are 30 percent sensitive, or 50 percent, I hope that you will recall this and either not write about it or write the headline that is not these tests are 50 percent sensitive, but maybe a headline that says scientists have to learn how to evaluate the tools that they’re using. You know, talk to me, I’ll give you headlines and sound bites. But what I don’t want to see is for Biden to push out a new plan and for the public to completely lose trust in it because the media is listening to the scientists. And unfortunately, the scientists in this case have just been wrong on more occasions than not. So I’ll stop there and I can take questions.

MODERATOR: OK, thank you, Dr. Mina. First question, she would like to know, what exactly do you expect the Biden administration to do now in terms of rapid tests? And have they invited you into consult or oversee them?

MICHAEL MINA: I would say that I’ve been consulting or sort of informally advising the Biden transition team and now the current team, so I haven’t been formally invited anywhere or anything like that. But it’s very, very clear to me that the administration, if you’ll look at the executive orders from Biden, for example, one of the first things that he did was he made an executive order to increase the accessibility and production of nitrocellulose. Nitrocellulose without going into any detail is one of the main ingredients in this little paper strip test, it just laminates the paper, essentially. And so that is a clear signal that this is going to be a major direction that the Biden administration is going to go. How exactly they get rolled out is not fully understood at the moment. But I believe that there will probably be a focus on schools. There will probably be a focus on sort of essential businesses. And I hope that there is also a focus on getting these tests into homes. And so it’s very clear that this is going to be a major focus. And I think in part, that’s probably because of some of the work that we’ve done over the year to educate. But we also have a CDC director who is extremely favorable to this approach of frequent mass testing. In fact, Rochelle Walensky is a colleague of mine and is now directing the CDC. She was also one of the first people to really write some scientific papers on the use of frequent testing as a means to control this fire. So I think knowing that we have a CDC director who herself has been the senior author on numerous research reports showing the need and utility of frequent testing, coupled with the new executive orders, I think that there’s going to be a massive increase in this type of availability of tests, hopefully in people’s homes. And importantly, it’s going to be in the context of so many other crucial tools that the administration is going to continue to emphasize and continue to utilize like vaccines and improving accessibility and speed that we get people vaccinated. I’m encouraged because on day one, we’ve just been seeing science-based policy. We’ve been seeing policy that is based on the best science, that’s based on real realistic expectations, that’s ambitious and also that’s based on equity. I think that if there’s a story about the coronavirus, thus far, it’s been in equities and Biden clearly understands that. And in everything that he’s doing, including these rapid tests, which can bring equity to the testing landscape and to the control the virus, I think that that’s a major underline for common thread that he’s really been pushing on.

MODERATOR: Great, thank you. Next question.

Q: Hi, thanks for taking questions. I wonder if really quick you could talk a little bit about that CDC study on the BinaxNOW, because it did seem, you know, they did say it was it was less sensitive, but they also sort of gave a second bite at the apple, they said when the virus was actually cultured, it performed better. It will be great if you could talk about particularly what it means to detect the virus in culture.

MICHAEL MINA: Yes, this is a great question. So in that exact study, for example, they found that amongst asymptomatic, that these rapid tests only detected 38 percent of asymptomatic people compared to RNA positivity. In all of our simulations, in all of our experience, if you look at the whole landscape of all of the people who are PCR positive today, only 20 percent to 30 percent of them will actually be infectious at any moment in time, and it’s because people remain positive for weeks after they’re finished with isolation, we know this isn’t a hard question that should no longer be in the FDA’s policy in terms of how these tests are evaluated. In fact, it’s CDC policy not to test a second time after somebody leaves isolation because we know that people will stay positive on RNA. And so what happens is that it just swamps out the signal of the benefit of these tests. More people than not turn up positive on RNA. And I don’t want to call them false positives, but they’re late positives. They’re no longer infectious. So in that study, what they did was they, of course, ran into the same trap as everyone else. They said the test is only catching thirty eight percent of people. That is close to the optimal. You could have a one hundred percent sensitive contagiousness test that would only detect thirty eight percent of RNA positive people on a PCR. And when they in fact, when they did look at the culture positivity in that which is generally, I think of being culture positive, should be the lowest bar to consider somebody at risk for really being transmissible. If you can’t get the virus to grow in a petri dish, it’s very unlikely that virus is going to infect somebody else. And so in that very same study that said it was 30 percent sensitive, it showed something on the order of 94 percent sensitive for culture positive virus amongst the tests that were sort of negative on it. So it’s just extremely, extremely important to recognize that scientists continue to make the same problem. Reporters continue to write report on the science. And in this case, I’m asking reporters to become a little bit of scientists and recognize things like culture positivity. Did they look at the CT values and try to understand sort of what level of those PCR results were actually relevant to public health and we’re actually of concern? We just did a study, for example, indirectly to evaluate this particular question. We knew the ground truth of people’s previous exposure histories because we’ve been testing them for a long time. And then we introduced rapid antigen tests across the board in a college. What we found was that once people turned positive, the test becomes quite literally on the second day of positivity, the test caught one hundred percent of people who are positive. But if we added in the people who remained RNA positive from a long time ago, some weeks ago, the sensitivity wasn’t listed as one hundred percent, but forty eight percent. But luckily, in this case, we knew the ground truth, we knew those people had already gone through their isolation period. So I don’t know if that answers your question.

Q: No, that definitely helps. So just to understand, you don’t view cultured virus as a perfect predictor of infectiousness or transmissibility.

MICHAEL MINA: No, I think it should be considered a low bar, because what I think is really hard for people to recognize is that this virus, from the moment that the PCR turns positive, if you are able to monitor somebody continuously, when it goes positive, when it goes negative, the epidemiological window when people are most infectious is very short. It’s like three to four days or so when people are really likely to transmit during that window of time. People have literally tens or hundreds of billions of viral particles, if not trillions in their nose. This test will detect down to about one hundred thousand culture. We’ll do something around that. So if you’re able to culture a virus on a petri dish, it might show up in an hour that it’s positive. You might be able to start seeing a signal in that petri dish in an hour and it might start showing a signal in four days. And so it’s four days. It generally means that it was an exceedingly low number of viral particles that went into it eventually will turn positive. But I think of culture positivity as really the lowest bar to just say that somebody is maybe possibly still infectious, and it should only go up from there.

MODERATOR: Really quickly, I put in a link to a New York Times article that discusses coronavirus infection and when you’re infectious, it has a nice little viral load chart as well.

MICHAEL MINA: And I just want to mention one thing on that, because if people link to that and take a look at it, it’s extremely important to recognize that when we draw these charts and I’m as guilty as anyone else, we draw it on a log scale. And so just to be very clear, the difference between if you just take the top 10 percent of this, you know, in terms of vertical space here, that might be a difference of like 10 billion viral particles versus the bottom 10 percent of the area on this chart is like one hundred viral particles. So it’s a log scale. It’s really hard to appreciate, but if you put that on a linear scale, something that our minds can really understand, the peak viral titers on that same graph, if you put it on a linear scale, will be extremely narrow. It’s not this broad kind of curve like that. 95 percent of that curve is essentially really, really low viral load compared to the peak. So I can show people later through an email or something.

MODERATOR: I think if you want to send that to me, I can send it to everybody on the call. So next question.

Q: Hi, thanks for taking my question. So my question has to do deal with some Facebook posts, some online messages that have been saying that there was new information released by the World Health Organization yesterday sort of changing the PCR protocol and that the World Health Organization admitted that PCR testing at high cycle rates produces false positives. And these articles are saying that they’ve inflated the number of COVID patients. So do you know anything about any change and how would you explain what this actually is?

MICHAEL MINA: So the change that took place was really just an encouragement for individual labs to understand better their CT values. It wasn’t suggesting that all these PCR tests are false positive. And it’s not suggesting anything of that sort. What it is suggesting is that cycle thresholds and PCR positivity can mean different things. And it was pretty loose guidance and just sort of some updated language that the cycle thresholds in one lab aren’t necessarily equal to another lab. And it’s worthwhile understanding better and keeping track of sort of what your cycle thresholds are. But there was no suggestion that it was that this is a story that kind of came unfortunately originated in part from me and The New York Times, where I was intending to suggest that PCR CT values don’t always implicate somebody as being currently infectious, which is what I was just talking about. That more time than not is spent being post infectious. But that doesn’t mean they’re false positive. They’re still actually detecting the SARS-CoV-2 RNA even at high cycle values. What it does mean, and I think we do need to contend with this. And I don’t want to fuel the flames of the of the conspiracy theorists, but we do absolutely need to contend with this issue that PCR stays positive for a long time. So if you’re asymptomatic and you just happen to go get a PCR test through surveillance, it’s more likely than not that if it’s a very low viral load, very high CT value, that you probably already were infectious and you maybe don’t need to be isolated in your recent contacts yesterday. You probably don’t need to be quarantined. But it’s very hard to know that you need to do repeated testing. You need to take that person and test them again or test them with an antigen test. That would be the other way to approach it. So that’s what the WHO is. They were just saying, CT values, you have to be discerning about how to evaluate them. And if you’re going to start using the actual evaluating the PCR as part of your decision making to really know what it is you’re doing. And they weren’t trying to say that they were false positives, though.

Q: Thank you so much.

MODERATOR: All right, next question.

Q: I know this is a little bit off topic and maybe beyond the purview of this call, but just curious what you think about the likelihood that the Biden plan to have FEMA involved in having one hundred federal mega sites, how well that will help roll things out?

MICHAEL MINA: Well, to be honest, I didn’t get a chance to read about that and what the actual plan is for the FEMA sites. What I can say is that I think we need all hands-on deck. I’ve been saying for a long time that we need a World War Time effort here. We need the whole country to rally around. We need all the resources we can get to tackle this virus. We need to do it efficiently, too. And I think having some sites set up that could allow efficiency gains, whether it’s the distribution of tests, whether it’s through distribution of vaccines, I do think that will likely be a benefit versus asking everyone to reinvent the wheel on their own, which has been very difficult for many.

Q: Thank you.

MODERATOR: Do we have time for one more question?

MICHAEL MINA: I think I bought myself another five minutes.

MODERATOR: OK, great. You can go ahead.

Q: Yes, thanks for doing this. I just wanted to get your input on a PCR test that the FDA put out a safety warning against a couple of weeks ago from Curative. Just wanted to see if you had any insights into why that may have been, considering that officials have encouraged the use of other tests that weren’t officially authorized for asymptomatic patients.

MICHAEL MINA: Yeah, you know, this is a real problem with the FDA at the moment. I have to stop criticizing the FDA, I guess. But it’s very unclear, it’s really hard to know how the laboratories are doing. These labs get an EUA and then, you know, is it just curative or is it every PCR testing laboratory? I know for a fact that things go wrong in a PCR lab. And this is one of the reasons I feel very strongly that we should not try to scale up PCR any more than we have, because when you try to scale PCR, you’re playing with fire. It’s an extremely easy technology to get wrong, it’s easy to get false negatives, it’s easy to get false positives. If used appropriately, and all the steps are put in place and designed and everyone’s following every rule perfectly, then the tests can be amazing and amazingly sensitive. But when you’re scaling up testing for one of the highest complexity tests, we could do, which is manual PCR, where if you drop one specimen that has already been amplified with PCR and it splashes, you could contaminate your entire lab because the PCR can detect one molecule and one specimen post amplification makes literally trillions of trillions of molecules. So if it crystallizes and goes airborne, not the virus, just the molecules, then all of a sudden you can get whole plates for a week that are false positive, you could misstep and lead to a lot of false negatives.

And so I don’t know the details of what happened at Curative, what I can say is I guarantee it’s not specific to Curative. I guarantee you that this is happening at labs everywhere. The problem is the labs don’t get evaluated on any regular basis. And when they do, they usually know that they’re getting evaluated. So they put on their best, best face. And this is a real problem, I think. I think we need the PCR testing. So I don’t want to deride it. I think we need high quality PCR testing labs that aren’t rushing to get hundreds of thousands of samples done unless they’re good at it. The Broad Institute, for example, the reason I started the effort in partnership with the Broad to do testing is because I knew that if there’s any lab in the world that could do this in a high-quality automated fashion, it was them. And we see this all across the country. I can’t speak directly to the Curative experience because I haven’t been able to get more information for now. But it’s another reason why I think we should not focus so heavily on PCR moving forward, it’s laborious. It’s at risk for massive problems. The lab could just completely miss a whole plate. And there’s five hundred samples gone with the wrong results. It’s difficult. Versus things like these rapid antigen tests, they’re distributed, you’re not going to run into the same kind of issues.

MODERATOR: All right, you have two minutes left. Do you want to take one more question?


Q: Thanks for taking my question. I don’t want to beat a dead horse here, but the CDC study, I just want to go back to again, just in terms of the science versus the lack of science, it seems like, you know, this was a study from CDC which you are critical of, and you’re also critical of the way the FDA communicates regarding antigen tests. And I just want to, I guess, go back to some of the suggestions. First of all, FDA has flagged both false positives and false negatives as they relate to these points of care antigen tests. Is that not the case? I mean, you just laid out a case for why PCR tests are flawed in that regard. But is that not the case with antigen test as well?

MICHAEL MINA: Well, some antigen tests are much better than others. And so I’m not saying that the FDA should not be critical, the CDC should not be critical about this test, but they need to be evaluated appropriately. So, for example, the BD, like some of the earlier antigen tests that came out, they do show a lot of false positives, but technology has progressed, and people keep putting them in the same boat. It’s like saying nobody is saying that the quality coming out of some small PCR lab over here is the same as that’s coming out of the Broad Institute. These are the same. Just because they look the same doesn’t mean they are the same test. And so we have seen all different qualities from a rapid antigen test in particular, can be of all different qualities because of the nitrocellulose used in all different things. And so they need to be evaluated appropriately. But I’ll give two examples. That CDC study was evaluating the BinaxNOW. And that, they found almost no false positives. I think it was like ninety-nine point nine, eight percent or something or I forget, exact number. Very, very small number of false positives. We just completed a trial of the ANOVA test on campuses. We did three thousand tests with zero false positives so far and one hundred percent sensitivity to detect people by two days into their PCR positive status or one day after they would normally get their results. So that’s the appropriate way to evaluate these, is to know exactly when people are in the course of their infection. So, you know, and to actually know some of them will have false positives. And I know that we’ve seen that with quite a few of them, in fact. And if they’re not high quality, they’re going to just like PCR. It’s changing the thresholds of things. Antigen tests can also be dialed in, even though they look so simply that the amount of ingredients you put on each strip, you can have a test that you’re really trying to go for sensitivity, especially if you’re using maybe lower grade ingredients. But that might come at the expense of specificity. But the Abbott BinaxNOW has been exceedingly specific and when evaluated appropriately, has been amazingly sensitive to detect the viable virus. And the ANOVA test has been essentially the same in our experience. And so I can speak to those two because I’ve evaluated them the most closely so far. But I do think that your point is very well taken. It’s extremely important to keep false positives in mind. But in general, for example, in our PCR testing program at the college campuses where we just finished this test, we found zero false positives with this and we found three false positives with PCR. So PCR is not infallible. It’s just that we just usually assume that PCR means that you’re positive. It was still one in a thousand. So it’s low rates. But we did find false positives with PCR as well. And you have to retest people, you have to know the previous history and you have to know their future to really be able to tell somebody that they were a true false positive on PCR.

Q: Quick follow up, if I can, just on that issue of FDA guidance, they still say that any negative results from an antigen test, I don’t think they say should, but may need to be confirmed with a molecular diagnostic. Would you recommend against that?

MICHAEL MINA: Absolutely, I think the FDA is making a horrible mistake here. They are just wrong. They haven’t kept up with the science. They are holding some of these tests up that could be made in the tens of millions every single day today. They are not authorizing the EUAs because they’re looking at these bad media reports coming from the U.K. instead of the application sitting in front of them. And they’re getting scared and saying, hey, why are things like these reports coming around, just like the CDC report with the BinaxNOW saying that it was 30 percent sensitive or something. It’s not 30 percent sensitive. It’s like 95 percent. You just have to know what you’re looking at. And the FDA, by continuing to require these tests to be compared against any time PCR, especially in asymptomatic, is just not keeping up with the science. They are making a grave mistake. And by suggesting that negative results on a rapid test have to be confirmed on the PCR, it completely depletes the potential use of a rapid test. If you have to get them confirmed PCR, that will overwhelm the PCR laboratories. We could never do it. And so as long as that recommendation remains in any way, we’re not going to see these tests. We’re going to keep seeing people get confused about them and all different things. So I think that in general, the decision here, the way that the FDA is approaching rapid antigen tests has been just completely off base. And I hope that the Biden administration hears me saying this. Problem number one to solve is to infuse an updated view of the viral dynamics. It’s a very simple story. The kinetics in terms of the clearance rate of real virus versus the RNA left over from the virus are just different. And if we don’t appreciate that, then we’re going to keep confusing scientists. I think the FDA doesn’t realize how much power they have here and how much their recommendations for how to evaluate these tests is playing into what scientists and the media and then policymakers think. So I would say that their decision making is quite poor unless you have a really good reason to expect that somebody might be positive or if the stakes are extremely high, then of course, no test is perfect. Get a second test. And the nice thing about these is you can have multiple tests. You can confirm them right after another due to at a time, three at a time, if the stakes are really high. But it doesn’t need to be a PCR.

Q: Thanks for taking my questions.

This concludes the January 22nd press conference.

View more press conferences