Coronavirus (COVID-19): Press Conference with Michael Mina, 05/22/20

You’re listening to a press conference from the Harvard T.H. Chan School of Public Health with Michael Mina, assistant professor of epidemiology. This call was recorded at 11:00 a.m. Eastern Time on Friday, May 22.

Previous press conferences are linked at the bottom of this transcript.


MODERATOR: All right. So let’s get going. Are there any questions out there?

Q: Hello. Good morning. Oh, yes. Yeah. It’s evening, good morning. So, yeah, I have the opportunity that we are not many today, so I can ask you the very long question and the question is so long that it would take a long time for me to share the question and for you to answer. So actually, I wanted to sum up basically everything you said during your amazing, amazing conferences so far, to ask you if it was really inevitable that we would wake up in the middle of the pandemic unprepared as we were. Or do you think that if all the pieces were in place, like a particular commitment, funding, research, clinical facilities, market opportunities, then we would be able before ahead of the pandemic to have all the system in place, meaning having testing systems, viral tests, serological tests and everything in place to face this pandemic in a in a better way? Or there was no way actually, even if we did everything we could have done, there was no way actually to to contain this pandemic more than we are able to do now?

MICHAEL MINA: I think there’s plenty of ways that we could have at least mitigated the effects of the pandemic. For example, some of us had called years ago in a paper in The Lancet for a global immunological observatory and essentially a system, a network of laboratories that monitor the globe in a way that’s very similar to how we monitor the weather today. And that would be a way that had we had a system like that set up, a true 21st century global network that that looks for for infectious diseases and epidemics, I think that that would be something that’s very possible with current technologies and also should be done, and probably would have been crucial to for both really discovery and detection of where this virus was spreading, how quickly it spread to Europe and the United States, for example. And it would have given us a much earlier window into what was happening with the virus.

And of course, testing for this specific virus would not have necessarily been possible immediately. But having the testing, having that infrastructure in place would have gone a very long way to to rolling out testing and being able to survey and know where the infection was hitting very early on rather than what we saw, which was essentially a rapid but still slow re-use of existing infrastructure in order to sort of hodgepodge together a testing network. And of course, now we’re seeing if we don’t make the change now to our infrastructure, and that’s not just testing, that’s also using 21st century technologies like mobile phones and like why do we still have people calling people to contact trace when we have everyone walking around with the G.P.S. unit. If that could somehow be leveraged in a safe private way, I think that that would be absolutely crucial to controlling this.

So I think there’s a lot of use of technology that could have been, that can still be used to control this virus and could have been in place before. The problem is there’s just no funding for it. I remember years ago. So I develop in my normal code pre-COVID life, I develop diagnostics, very high throughput diagnostics that can look for antibodies of any virus that could infect people. And I look for all of those hundreds of thousands of different antibodies in any drop of blood. And I recall going to a venture fund to try to get funding to really develop this into a public health platform years ago. And the response I got was we definitely don’t fund public health projects, you know, there’s just no money in it.

And I think that idea permeated throughout our society globally, that public health has been just completely underfunded, particularly in the United States, but elsewhere as well. So I think that we have the ability to – if we can have a video, if I can be if you can see me right now and you can actually see where I’m sitting, a picture of me in real time, then I’m pretty sure we can figure out how to monitor for infectious diseases. And so I think we just have to put their minds to it and we have to figure out some way to commercialize it, too, if that’s the way our world works to make these things viable, solutions to get companies like Facebook and Google. Instead of creating apps that will get people frozen yogurt quicker, we should be creating apps that will help save lives. And I think COVID will probably initialize some of those types of new endeavors, because now we’re seeing the massive economic and public health destruction of this virus. But before hand, I think there was just no real will or or interest in funding these programs.

Q: I have a follow up question on the detection part. So there were some studies, joint studies made by researchers in the US and virologists in one laboratory in China that used the viruses, so viruses produced artificially in the lab, to show that the SARS family of viruses actually could mutate in a way that it could actually infect humans directly. The one laboratory virologist, they did the study the bats to try to trace this kind of SARS diseases to animals. So what else, actually, the different countries like the US and Western countries, they could have done to, you know, in terms of microbiology, research, and to track the coronavirus in a way that we would have been able to detect them ahead of the pandemic. Since I’m not a scientist, I’m trying to figure out how you can track these kind of viruses all the time.

MICHAEL MINA: So I think – the best way to track the pathogens, I think, and to see if they are going to jump in to humans, or if they have jumped into humans, is to use the immune responses of people. So people, we all carry around inside of us these amazing pathogen detection systems that are more advanced than anything we develop, that we artificially develop and that detection system that exists within all of us are our own immune systems. We know how to read the immune system now. And so if we can utilize new technologies to read out the immune system and think of it like a hard drive that we just have to decrypt, then we can do that and we can do that across millions of people a day or month or whatever it might be, then all of a sudden we have a really useful tool to track viruses and see when they have jumped into humans, for example.

Currently, the way that it goes is quite different. Usually we look for viruses themselves and so we have to identify the actual individual who currently has a virus inside of them and then sequence that virus and pull out the virus, try to amplify it and sequence it. That’s very difficult. And it’s also difficult to find that individual who happens to be sick at the moment. But if instead we focus on looking at antibodies in the immune response as a mechanism for that, then it opens up a whole new slew of options for us where we can – we know how to read antibodies, for example. We are learning how to read B and T cells, and we can use those as our as our sentinel surveillance system, and just start collecting samples on a routine basis from people around the around the globe, in particular in places where viruses are likely to emerge from zoonotic transmission events. And I think that that’s actually a very smart and good way to go. And it’s something that we are thinking about setting up currently as well.

Q: So based on what you said that the you said the Center for Disease Control in the US, for example, together with the for Center for Disease Control in China, they should have collaborated and put in a system which actually would have tracked the antibodies development in people in China, received these antibodies, actually provided the evidence of the new coronavirus? That’s correct?

MICHAEL MINA: That’s right. And you can you can look for antibodies that don’t even – so it’s hard initially to find the antibodies that are for a novel virus. But we know that antibodies create cross reactive responses, for example, with other coronaviruses, so we could be tracking antibodies that are not so specific for the novel one, but that are more generic for lots of coronaviruses. And we would see a spike in those, but we would at the same time, so we’d see a spike in antibodies against a generic coronavirus antibody platform. And then we could dive into it if we do see that spike. We could say, okay, which coronavirus is it? And if we recognize that it’s not any of the ones that we know of, then we would be able to infer that it’s a novel one and that we need to look very closely and sequence it. So it would be an easier and more efficient way to monitor for these things rather than waiting for the first 10 people to come into the hospital and die before we actually recognize that there’s a novel virus killing people.

Q: Thank you. 


MODERATOR: Next question.

Q: Thank you. Thank you so much for answering and by the way, I think Room Rater would give your your background high marks.

MICHAEL MINA: What is that?

Q: Room Rater. They’re going around and reading all the rooms of everybody doing online video presentations. It’s a fun Twitter account. Anyway, my question has to deal with a particular statistic that’s being put out down here in Florida. And I am trying to understand generally the best way to communicate positivity rates, right. Positivity rates in a normal world would be the number of people tested versus the number of people who tested positive on a normal day right. Here in the state of Florida, the governor is communicating a positivity rate that includes the number of people tested and retested on that day, right. So not just the people tested for the first time, but the people getting tested for second, third or fourth time, divided by the total number of new cases.

And we’ve figured out the actual rate of positivity of people just getting tested for the first time on that day. Does that make sense, that distinction? And I’m trying to figure out the sort of epidemiological reasoning or the reasoning behind showing one and showing the other, you know, because I don’t want to go out and say, oh, the governor is tamping down the positivity rates or anything like that, because I’m sure he has a reason to do that. But obviously, the positivity rate that we figured out is much higher than the one that he’s showing. I kind of want to learn a little bit more about the rationale maybe behind showing one or the other.

MICHAEL MINA: So when you’re saying multiple tests, you mean like somebody cuts positive today, is getting tested tomorrow and is negative in this test of the next day and is negative?

Q:  Exactly right. And perhaps doctors are going and getting retested six, seven times, and that’s included in his positivity rate that he’s showing. So it’s significantly lower.

MICHAEL MINA: Yeah. So the problem would be, I think that it’s an acceptable thing if, let’s just pretend for a moment that it’s a representative sample, which it’s not. But let’s pretend it is. If you’re looking at people who are negative and have not yet been positive, then that’s an acceptable approach because each negative person, if they’re getting tested again and they remain negatives are actually just representing the people in the population.

But if they’re including people who have already been positives and then having those people be tested multiple times again in the future and including those, then you are artificially – at least if you’re enriching for those people who have been tested positive already and then testing negative and reporting those out, you will artificially decrease the apparent risk to a susceptible person if you’re including people who are now resistant. So I don’t think that there’s a good epidemiological reason to do it. I can think that pragmatically thinking like a laboratory director in a COVID testing laboratory, it is very difficult for the results that get pushed to the public health authorities to know if the data is coming from the exact same person. And if they have already been tested negative or positive in the past.

Now they do have the data capacity to do it for sure. Because when we report to the state laboratory, we are reporting identifiable information so that contact tracing can progress. So I suppose actually it would not be that difficult to filter out those results for people who have already been positive and not incorporating their future negative results into the data. So, I can’t answer the question. I don’t think there’s an epidemiologically sound reason to do it. I think it might be a more pragmatic reason to do it.

Q: That’s fascinating. And so then I’m also presented with the unenviable task of making this distinction easy to understand. And I was wondering if you could think of an analogy or perhaps check me on an analogy that I’m planning on using, which is to show that it’s something like if a basketball player has a statistic for three point shots and you’re sort of like filtering in the free throws. You think, is that analogous in your thinking here?

MICHAEL MINA: I don’t know when free throws happen in basketball.

Q: Well, I don’t know basketball that much either. I’m just trying to connect.

MICHAEL MINA: But the analogy, it sounds like it’s getting through what I would think, although I would say adding in the free throws would do the opposite. They would enrich the counts, whereas what’s really happening is if you are, you’re essentially adding to your documented rate of positive tests. People who are not eligible and if you will, to be positive because they have already been positive and they have recovered. So I would say that it’s what is it like? It’s. It’s. I’m trying to think what it might be like it. I mean, to me, you’re essentially inflating or deflating the positives right by incorporating ineligible people. There is a good analogy. I’ll try. Come to me. Well, nobody else has their hand up right now, so I’ll just keep thinking about it.

MODERATOR: I was going to say to testing men for pregnancy, because -.

MICHAEL MINA: That’s a great idea.

MODERATOR: Because you’re always going with a pregnancy test analogy, so there you go.

MICHAEL MINA: It is very much like throwing in or asking, you know, what are the pregnancy rates and adding in females who are on – is it is it right to add and females who are on contraception reliably into sort of the probability of getting pregnant? But adding men is just a much more clear example. Exactly. And I think that’s a good example.

Q: Oh, that’s perfect. Perfect.

MICHAEL MINA: I like that. Yeah, it does. You know, I don’t know what the mindset is there, but I do think it would deflate the positivity rate. Do you know how many people are actually being, what proportion of the people tested and being reported fit that bill? And how how do you know that they’re in there. Is being reported? How many times people have been.

Q: So it took me four days and two sleepless nights to figure out the distinction between these two things, because we had been tracking the total amount of new tests reported by the state and graphing that. And we saw a divergence between our positivity rate graph and the positivity rate graph of the state. And I was tasked with explaining that divergence. And the reason why it diverts is because they’re graphing people who’ve been retested, but they have another database where they show you that number without the retests. So we know the difference between all of their new tests on one day and all of the retests on that day.

So, for example, on the 19th of May, they had a huge increase in testing. Seventy seven thousand tests came back that day from the laboratories, which is great. And you’d think the governor would be crowing about it. But only fifty five thousand of those tests were new tests. And they’ve had days where they’ve actually had more retests than actual tests. The reason why we started looking into this is because on the 15th there was a big positivity spike. We went from 4 percent positive to 11 percent positive state wide because there were only thirteen hundred new tests and six thousand retests on that day.

MICHAEL MINA: Mm hmm. And I guess the other part is that the retests could come back positive quite frequently too. So if they’re including everything, then it’s hard to say what exactly the bias will be. If you’re testing somebody today and then you test them three more days, that’s a retest. But it might very well be positive. And then if you test them again four more days later, it might be negative. However, the practice, at least in the health care setting of requiring two negatives before a person can go back to work, for example, would likely bias it towards having more testing of negatives than positives amongst those retest. So it does run the risk of throwing the the bias towards a picture of fewer cases, of a lower rate. But it’s a tricky question in terms of what it would actually do to the results. Because we know that a lot of people are getting retested pretty frequently after they’ve been tested once. So it could actually be inflating.

Q: I have to get somebody on the line from their office of epidemiology or the governor’s office, and they’ve just been giving me the cold shoulder for three days down here, so, you know. Well, thank you, that answered my question.


MODERATOR: And I’m going to ask a question that has come up a couple of times in my inbox in the last couple of days, so there have been reports of states, Georgia, Virginia and others that have been putting all the antibody or serological testing in the same group as the virologic testing and then giving it as one data point for the state. Could you explain why that’s problematic to put both those types of tests in the same category? Does that make sense?

MICHAEL MINA:  Well, the question makes sense. The fact that you’re asking the question or have to ask the question is beyond me. It doesn’t make sense. It doesn’t make sense at all to do that. And I really worry that it obscures our ability as epidemiologists and as public health people around the globe to be able to understand the best way to do public health and to control this virus. For example, I mean, they just have to be separated because the antibodies represent a cumulative case count across a population of cases that may have happened last week or last year, for example. And virus represents what’s happening today.

And that’s a big limitation of putting the case counts of the virus, but it represents what is happening today alone. So I think that if they’re adding the two together, then they have to start considering it as a cumulative case count rather than a daily, but then you’re double counting. So, for example, if somebody, I mean if you can assume that everyone only gets tested once, which know is is a different question, though, is just being asked, but if you can assume that everyone just gets tested once, then you could start assuming that they can be grouped together, but it has to be interpreted as a cumulative case count of all the people who have been infected throughout the entirety of the epidemic until today or up through today.

So if that’s the goal, then that’s fine. Actually, that’s an appropriate thing to do. But if the goal is not that, the goal is to get a daily counter of the new cases happening each day, then an antibody is just totally the wrong way to go. I think that in general, though, I would say antibody testing is the right way to go from a public health perspective overall. But the databases just need to be separate. There’s just no question in my mind. They look for different things. It’s like it’s like combining, I don’t know, daily smoking deaths with daily melanoma deaths or something. They’re just two totally different, two different things, I would say. So I don’t know why they’re doing that, to be honest. I think that they should be always kept separate. They’re coming from two different labs. So that’s all.

MODERATOR: Thank you. Are there any questions out there?

Q: Yes. And I’m taking advantage today. I’m really happy about it. So, can you see a link in the chatbox? I shared it with you. I just found this link. I don’t know if you can see it in the chatbox.

MODERATOR: I’ve got it. Just a second and I’ll share it with everybody.

Q: So, Dr. Mina, you know this initiative called the Global Health Initiative?

MICHAEL MINA: Sorry, I didn’t understand what the question was.

Q: The Global Health Security Initiative. Yes. So it was an international initiative launched in 2014 in the wake of the outbreak of Ebola. Actually, it was launched by President Obama, I think, in the U.S. And it was to actually establish a cooperation framework between countries in order to improve the preparedness of the other national systems that had pandemics and these initially consisted of different working packages. For every working package, there was a number of countries that took the lead or they were cooperating. So there was one specific working package that I think relates to what you said answering to my question.

So there was a working package called the Laboratory Cooperation. It was led by the US and one of the contributing countries was China. And so this package actually says that the aim is checking the new pathogens. I don’t know if the methodology they set up corresponds exactly to what you told me or it’s something different. I just wanted to know if, uh. I mean, if you if you can look at it in the comment and see if it’s what you what do you in mind or something else? Can you do this?

MICHAEL MINA: I think – I actually hadn’t seen this, which is really surprising, but there’s a million things that are started all time. So this is actually very much in line with our idea in general. Whether it’s actually come to fruition yet is is a whole different thing, and I don’t think it’s really – my very quick understanding or read of this is I don’t think that this has really come to fruition in any practical way. And I do think, however, that it needs to be happening.

And this is very much in line with what I was just mentioning a moment ago about our idea of the global – there should either be a global pathogen observatory of global health, you know, something along these lines. I think is absolutely essential and we should be doing it. And so it’s good to see that at least the plan has been put out there besides, you know, being in just in the academic papers. But I think, you know, it’s one thing for them to put this out there as something to do, but how it’s really being used and whether it’s being used for epidemics and novel epidemics is a different question and I think it’s probably most useful at this very moment in time for things like HIV and measles and other viruses. But I’m not sure that it’s set up for rapid deployment of of observations for new and novel outbreaks. And that’s something that I think should be put in place.

Q: OK. Because the document also mentions novel pathogens but I think, as you said, that it was a nice project on paper but I don’t think there’s ever been implemented seriously concerning new pathogens.

MICHAEL MINA: Yeah. No, And looking at it, I mean, this is from – I see a link from 2012. So I don’t think it’s come very far. Certainly there are networks, though. We have an example of this. The influenza network, the way that we monitor flu strains every year, which was, of course, utilized for COVID. It is a good system and one that can be very useful and deployed because it exists already. And so I think that, you know, the co-option of co-opting that system to use it for collecting COVID strains was smart. And those are the type of things that we should be doing much more robustly and much more globally.

Q: OK. Thank you.

MODERATOR: Next question.

Q: Hi. The Lancet pushed out a paper today, phase one results from CanSino’s vaccine trials using their Ad5 vaccine. You know, they see some good things, but they also see a dampening of the response in people with preexisting immunity to the vector. And they see quite a bit of preexisting immunity to the vector. What does that tell us about the prospects for this vaccine?

MICHAEL MINA: So I apologize, I thought that this was going to be about a different Lancet paper, the hydroxy one that just came out.

Q: So many.

MICHAEL MINA: Yeah. I clicked on the wrong link. So I believe that – so, I can’t speak directly to that paper until I read it, but I would be happy to talk to you later about it if you want. I mean, this is the worry. This is definitely one of the concerns about using vectored vaccines for which people might already have preexisting immunity.

It’s also the concern about vaccines in general in terms of what we call antigenic seniority. If you create a response, that might skew it, so if you already have seen a virus or have some preexisting immunity to it and then you try to vaccinate against something that’s very similar, you run the risk of having your immune response get skewed and kind of picking up primarily the thing that you’re already immune to or that you’ve already seen and kind of not focusing so much on the new aspect, which in this case would be the coronavirus proteins that were placed onto the adenovirus vector. And that can lead to an imbalance where you want to be prioritizing your new approach, the immune response against the protein of interest for coronavirus but you end up picking up and running with the pieces of the virus that are forming the vector that’s carrying the coronavirus protein. So that can certainly, that is sort of in the same framework as antigenic seniority.

And it’s one of the reasons why we think that various people have different susceptibility to flu over their lifetime and why some vaccines against flu don’t seem to work so well. And the same applies here. And so I think if it turns out to be really a big issue, then I think it does mean that we have to reevaluate the utility of some of these vaccines or we have to reevaluate the utility of some of these technologies and think, is there some way to create a safe and effective vaccine from the ground up without having to put it on a vector that many people might have already been exposed to in some way or another. Does that make sense?

Q: Yeah, it does. Thanks. But if you had a chance to take a look at the paper, I’d appreciate your thoughts if that’s doable. Thanks.


MODERATOR: Another follow up question?

Q: Yeah. Yeah. I promise you, it’s the last one and I will shut up forever. So I need to kind of follow up on the question that was asked by the other journalist. So it’s a question I’ve also asked the researchers. So, I would like to know what is your thought about of the vaccines that try to target the common elements of all coronaviruses in such a way that actually they can be useful against all kinds of coronaviruses. And some companies I was speaking with said that they had this kind of an antibody that they developed in the wake of the SARS-1, 2002/2003, but they never ended up being able to use it because there are no more infected people. So I was wondering whether this kind of approach is something realistic or actually it’s impossible to have the vaccine which is good against all kind of viruses related to the SARS family.

MICHAEL MINA: I think what you are asking about is very much in line with the idea of a universal flu vaccine. And I think it’s possible. I would never say it’s not possible. I would say that we don’t know how to build one right now. The way that we have historically done things like this, we can look at pneumococcus as an example and streptococcal pneumonia and we can ask how has the pneumococcal world dealt with this question? And the way that they have done this is to just to add more serotypes to the vaccine, rather than looking for a single piece that will cover all of them and that might be the way to go for coronaviruses, in the event that this does not necessarily mutate rapidly.

It won’t work for flu because flu mutates quickly. With this virus, with coronaviruses, they might not mutate quite so much. So you can envision if you’re making vaccines, you may as well just put multiple strains in instead of one. And that could be the way forward for that. If you’re trying to build a vaccine that’s going to preemptively protect against any novel viruses that may emerge, that may help, although we would have trouble knowing it. If everyone’s vaccinated, then they would never really have the chance to jump into the humans in the first place and spread. But short of that, I think I’d be harder to do that. But if we could find some conserved regions of the coronavirus that actually are sufficiently immuno-dominant, immunogenic to create good protective immune responses that it’s theoretically feasible.

Q: If more advancements were done in the past, this kind of universal coronavirus vaccine, we could have had a vaccine ahead of the pandemic or is that optimistic?

MICHAEL MINA: Well, I think it would be asking a lot to to assume that we would be able to have done that. I think it’s – we don’t have any good universal vaccine. I think that it would be asking a lot to try to to have a expected or that a universal coronavirus vaccine wouldn’t have existed. There’s just, despite SARS and MERS, it was not considered such a serious global threat just because it hasn’t been a serious global threat. Of course now, in hindsight, we can go back and say we shouldn’t have really pushed for serious billions of dollars to go into vaccines against coronavirus and try to create as a universal coronavirus vaccine back in 2007 or whenever it might have been. But I think we probably will attempt to be creating those in the future. I can’t imagine that people are not going to be working on them. So was it likely and feasible and should it have been produced before this particular pandemic? I think that’s an unlikely scenario. And now an impossible one. But in the future, I think we’ll try to push for one.

Q: OK. Thank you.

MODERATOR: Next question.

Q: Thank you very much for doing this. I’m going to I want to ask you a question about testing. And it’s very local. In the city of Cambridge, now there are free tests being offered to any resident, and yet there is not a huge demand for people to have these tests. There are not people lined up to get tested. And I’m wondering, I know that that, you know, you try to measure the when you have enough tests by looking at the positivity rate. But is is there a problem with this that even when you open it up, that people, especially those who don’t weren’t sick, do not want to be tested?

MICHAEL MINA: It’s very difficult to get – I didn’t realize that anyone could be tested. Do you know what the – this is virus testing, I presume.

Q: Right. This is a PCR test.

MICHAEL MINA: Yeah. So I think that this is a difficult aspect of this virus and any virus. I mean, in reality, I don’t advocate for everyone to be tested for the virus. Testing for a virus isn’t an efficient public health effort. I’d say it’s a very inefficient public health effort. It’s needs to be done in the midst of an outbreak to know who is infected, when and if it can be rolled out very quickly and rapidly and across many, many people, great. But when you’re not sick, then there is little utility in testing for the virus.

In particular, if you’re not sick and the prevalence is very low. If you know that people around you are sick with this virus, then it makes sense to go and get tested for it. But in Cambridge right now, the prevalence is not particularly high. And so I’m not surprised to hear that people are not taking a good call to get tested or the offer. And I don’t think it’s a problem either.

I think that actually the better way to, the better program is always serology, I think. If you’re in terms of ongoing surveillance and monitoring of the population, you want to monitor who has had the virus so that when it does reemerge, you know, who is maybe at more risk than others, what the risk of an emergence is in the population that you’re looking at. So I believe that I would be interested to know what people would do if the offer was on the tables for antibody testing and people could know if they were infected at the height of this epidemic, or outbreak. I believe also that as things open up and if we do see increased cases, at that point, people might again become concerned that they might be exposed.

But in general, I think that if you know that you haven’t been around people who are sick, then why go out of your way to get tested? If you feel that, you know, if an individual feels that they had very low likelihood of becoming exposed and it’s still a transient virus. So even if you were exposed to months ago, then there’s no good reason to come back now and get tested. Whereas with antibodies, it’s a whole different thing. If you were exposed two years ago, obviously not for this virus, and you want to know if you were, then the antibody test will tell you that. And so I think that they are two different things. But I would say that it doesn’t surprise me that appetite to be tested for this virus has decreased as prevalence has gone down.

Q: Thank you.

MODERATOR: I’m going to piggy back on that one real quick. So as things are opening up and people are moving around more, there’s been more and more talk about a second wave coming. How do you see testing playing into monitoring for a second wave and how should that be rolled out? When should that start? And what are your thoughts on that?

MICHAEL MINA: I think it should start now. It should never have stopped. But I do think that it should shift. I think that one of the most powerful tools we have in public health are antibodies, it’s why I study antibodies and a whole field called serological surveillance. We rarely talk about biological surveillance as an appropriate way to monitor populations for pathogens. It’s just so expensive and so unwieldy because you have to find the person at the moment that they are infected in order to find the virus.

And so if you’re monitoring – what I think Massachusetts should do and I think the country should do, is not set up these baseline testing things like Massachusetts has done for the nursing homes, for example, but we should set up really well designed longitudinal surveys where we’re monitoring individuals over time, maybe once a month, take twenty thousand people across the state – and frankly, this is something we’re doing now or we’re trying to get set up anyway – but you take thousands of people and you monitor them every month and you ask them to put a drop of blood from a finger prick, for example. These are generally practically painless. And you have them send in a finger prick drop of blood and you can use that to know if they’re seropositive or seronegative. And you do that every month with a sufficient number of well representative people in the population. And you can easily – well, not necessarily easy, but you can you can efficiently monitor for the reemergence of the pathogen.

Because it’s just so much more powerful to do antibody testing, because it’s because the signal lingers with somebody potentially for life. So if they go from negative to negative and then they turn positive a few days after their last test, if it was viral testing once a month, then you’d completely miss it. You test negative here. They would get an infection a few days later. And by the time you go back and test again, you would have missed the infection in that person. Antibodies, on the other hand, you test and then they get infected and then you test another month later and you’ll see that they seroconverted. They became positive. And so that’s a very powerful tool to use because you can really space out the number of tests that you do so that you’re not trying to capture every individual infection, but you’re capturing the evidence of the infection. And as long as you’re measuring longitudinally in the same person, you can use that to understand, you know, you’ll have a good window of when the infection occurred. And if you do that effort along enough people, then you can actually identify new outbreaks emerging in the population.

So I think that that’s what we should be doing now in the state and across the country. And you get a lot of additional information. So, for example, surveillance so far has been so incredibly focused on the viral testing. But if we look at what’s happening in the nursing homes, and a few of us just had an op-ed in The Boston Globe on Monday about this. But the nursing homes, the governor, not just in Massachusetts, but many places are mandating what they call baseline testing in a nursing home and baseline biological testing. So if you go into that nursing home today and you just do one baseline test of it, you might find that twelve percent of people are positives. And so then from there on out, people say, oh, that nursing home is twelve percent positive, but that’s not true.

Without serology, you don’t know if that 12 percent is actually representing the down slope of a big epidemic that just happened and maybe eighty percent have actually been infected, but only twelve percent today remain infected, or if everyone’s your negative in that group in that nursing home, then that’s 12 percent could be a strong warning that they’re about to experience a massive outbreak. And maybe next week, 40 percent will be positives. And so serology in conjunction with that would go a long way to understand if the outbreak has already gone through or if the outbreak is about to take off. So if antibody testing at scale across populations allows you essentially to risk stratify communities and to know where you should allocate resources, for example, where you should do enhanced surveillance, where you can let up on surveillance.

So it’s a way to essentially complement PCR testing, which PCR testing and viral testing should be very acute. It should be done acutely in clear windows of ongoing transmission. And antibody testing is something that just kind of goes on throughout the population over time, even in periods when outbreaks are not happening as a true surveillance system to monitor for when they do pick up. So that’s where I think that things should really head from a testing perspective.

MODERATOR: OK, great. Thank you. Do you have any final thoughts before we go?

MICHAEL MINA: I find it funny that I just got the – oh, I think I’m probably quoted in that actually. Getting an antibody test for the coronavirus – here’s what it won’t tell you. It’s what popped up on my phone from NPR. So I think I don’t have anything else right now. I think there’s a lot that antibodies won’t tell you. But I think it’s a tremendously good way to monitor the population for new outbreaks. And I think it should become a cornerstone of the US’s surveillance system for both this outbreak of this virus, but also novel viruses as well. And I hope that the U.S. will get behind serological testing because it’s just much more efficient.

This concludes the May 22 press conference.

Sarah Fortune, the John LaPorte Given Professor of Immunology and Infectious Diseases and chair of the Department of Immunology and Infectious Diseases (May 21, 2020)

Howard Koh, Harvey V. Fineberg Professor of the Practice of Public Health Leadership (May 20, 2020)

Joseph Allen, assistant professor of exposure assessment science (May 19, 2020)