Coronavirus (COVID-19): Press Conference with Michael Mina, 04/28/20

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 11:30 am Eastern Time on Tuesday, April 28.

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


MICHAEL MINA: I am an assistant professor of epidemiology and I am also an immunologist and work on development of serological diagnostics. I’m also a pathologist where I oversee molecular biology diagnostics. So, the questions that I can answer span from epidemiology to testing related items. So, I’m happy to take any questions.

MODERATOR: Great. The first question. Go ahead.

Q: Thank you, Dr. Mina, for doing this. This is my third or fourth with you, so I might be drilling down a little bit compared to some of the previous ones. Without necessarily getting specific, I’ll leave that up to you, I see that epidemiologists are increasingly criticizing and, correct me if I phrase this wrong, but criticizing some scientists perhaps with different skill sets, who are weighing in on COVID-19, and I’m not speaking specifically about studies, but about what appears to be maybe ‘armchair epidemiologists’, if that’s something that you’ve thought of, and just curious your thoughts on what’s happening here and if this is a problem, or if this is something, you know, you face throughout your career. If it’s kind of brand new, given the circumstances.

 MICHAEL MINA: Well, it’s certainly in overdrive right now. Infectious disease epidemiology is a pretty niche field usually and that doesn’t really matter too much, except it’s niche because it’s actually the models that we use, and the actual mathematics that underlie a lot of the epidemiological models are not always intuitive. And they’re not always – they’re much more nuanced than might meet the eye when you look at the result of a model for example.

And so, we see in this epidemic, and it started – it was particularly bad early on with economists, and not even economists, I shouldn’t say that, but with a lot of analysts who work in Wall Street, for example, were building their own epidemic models to understand what was going on. They were getting growth rates that were out of control, for example. We’ve seen a lot of what you referred to as ‘armchair epidemiologists,’ you know, trying to wade in which is great. I think having a diversity of expertise in this field is extraordinarily helpful. However, the science has been – you know, there’s a reason why epidemic modeling is how it is, and a lot of the trials and problems that have been faced in the past have been worked out. But if you’re not really comfortable or familiar with the fundamental nature of how a lot of epidemics can and should be modeled with regard to spread and immunity and things like that, then you can be off by quite large margins and we see that this has actually become – out of the academic sphere and into the very, very practical sphere, this has become a huge issue

For example, hospitals –  a lot of hospitals will have their own epidemic modeling efforts to understand what was going to be, what were the expectations for ICU beds and testing and things like that that were going to be necessary across the country. So, a lot of hospitals were just wildly out of line with what ended up happening and some of that’s because of the actual epidemic models by infectious disease modelers were off. But then what you really saw a lot of were armchair epidemiologists jumping in, or I should say armchair infectious disease epidemiologist jumping in and building just exponential growth models that were, you know, suggesting to hospital leadership that they were going to need, you know, twenty times more beds than they had and none of that ended up materializing for the most part, across the country.

And so I think that there’s a real danger in that. And unfortunately, I blame us because perhaps we haven’t really explained well enough why epidemic modeling, what are the nuances of it, and maybe we’ve created an environment where it seems more straightforward than it necessarily is. And sometimes it really is straightforward, but sometimes it’s really not, especially when we’re trying to get into the nuances of true control measures and supply chain questions and things like that. It really takes pretty specific modeling efforts that are grounded in robust methods.

And so, I think that this has been a problem and it goes beyond epidemiology modeling. We see that all of a sudden overnight, you know, most investigators in any medical school have become infectious disease investigators trying to do serological surveillance or development of models, you know, and any number of things, and you run into issues when you don’t sort of ground your efforts in – all of the problems that have been worked out in the past, you know, are kind of materializing again, I see. And I think that it’s a lot of sort of redundancy or use of resources where they’re not necessarily needed because you have a lot of people who aren’t so familiar with the landscape of how to do these things, trying to reinvent wheels when there’s better methods already developed.

Q: Thank you. Appreciate that.

MODERATOR: Next question. Now that at home serological testing is starting to be offered to consumers, what would be the caveats that you would most want people to know about and in what cases, if any, would you even recommend that someone order a direct to consumer test?

 MICHAEL MINA: Was that specific for serology did you say?

MODERATOR: Yes, serological tests.

MICHAEL MINA: Yeah, so the serological tests, I think, I don’t recommend them at the moment. On the one hand, we don’t you know, especially for individual level testing,  there’s no – so, we can deal with some false positives and false negatives at a population level because we can account for them statistically. But at an individual level, if you go and buy one of these point of care tests and have it delivered to your home for example, and sent to you, then if it’s wrong, you won’t have any idea and you won’t be able to account for the error rate. You can guess if it’s right or not. But you won’t really know, and they’re not exactly reliable at the moment. So I don’t recommend to anyone at the moment to use these point of care tests for personal use.

We are developing much better quantitative tests that will be more accurate and more reasonable to help somebody to understand what their immunological levels are. So, I think that, despite the flood of the market of these point of care tests, they are generally not great. And things like temperature and how they’ve been stored and packaged and shipped can really greatly affect these things, and when they’re just being shipped out, the proteins can unfold that are in them and things like that. And so there’s actually a lot of biological reasons they can be damaged and not work appropriately by the time they reached a household.

So, until some of those issues are worked out and you have really credible sources of these tests and they’ve become a little bit more stable, I think that people shouldn’t consider them as anything more than a toy. You know that might sound trite, but that’s kind of where I come from with that question.

MODERATOR: Thank you. Next question.

Q: Hi, Michael, just following up on that last question. It’s a question about serology testing and let’s talk about the high-quality quantitative kind of serology testing that you’ve been involved with. I’m very interested, and this is a Massachusetts specific question, on how those tests are being used for surveillance testing. Can you talk about what the prospects are for that kind of testing using the antibody testing, the serology testing in Massachusetts or, more broadly, that you might be involved in, or that your colleagues might be involved in, that could help tell us more about how widespread the infection has become?

MICHAEL MINA: Sure. So despite not really fully understanding how to interpret the tests in terms of what sort of protection having an antibody response provides, one thing that we can know how to interpret from the quantitative antibody test is that we can generally infer that somebody has had the infection if they do have the antibodies. And so that does allow us to be able to go out into the community and survey individuals, specifically if we have quantitative tests, we can get two different pieces of information. On the one hand we can get how many people have seen the infection overall, at least as of about 10 or 12 days prior to the collection. And beyond that we can also try to understand from the ratios and from the numbers of the two different antibodies from tests for IGM and IGG, we can actually start to infer how long ago they were infected and whether they were infected within the last couple of weeks or the last couple of months or, you know, further back before that. At this point, of course, it doesn’t go much further back.

So these are being deployed in surveillance studies. There are a number of big community surveillance studies that are ongoing or about to get started. Some of them by myself and some of my colleagues, some of them by colleagues at University of Massachusetts, BU. So a number of these studies are getting going. Some of them focus on health care workers, while others focus on sending out kits for sample collection that then can be mailed back into a laboratory to be processed. And so that’s actually what we are working on, is how to develop these kits, so that people can do a finger prick at home, send that back in, and then we can process what we call a dry blood spot for antibodies and that actually works very, very well.

I think that this type of testing is going to be increasingly important in understanding the epidemic. For example, if we go out and we do surveys across all the nursing homes in the state, which is which an initiative that’s ongoing right now, one of the problems with that is we might go in and measure how many people are currently infected with a virus at any given point in time. So, between now and the middle of May, maybe we will be able to test for virus every or most nursing home patients, but what we won’t understand is, if a nursing home, for example, has 10% virus positivity, does that mean – is that because the outbreak is just getting going? And maybe in the next few weeks you’ll have 30 or 40% positivity? Or is the outbreak already passed and you’re actually on the downslope?

And that’s where viral testing lacks, because it doesn’t give you a long history and where antibody testing ,when coupled with that, can really provide a lot of information on what has happened. So if you if you go into nursing home and 10% are positive for virus, but 50% are positive with antibodies, then you know that that 10% that’s positive environment, you can still act on on co hoarding and quarantine them. But you also know that, in fact, of the large fraction of the nursing home has already been infected as an example.

So I think that the antibody tests and particularly quantitative ones will become increasingly useful and important towards opening the economy, understanding where people have not – where the virus hasn’t really hit yet in terms of different communities and understanding where it actually already has hit and the actions that we can take based on that type of information is different and and then the viral testing. So I think it will go hand in hand.

Q: Thank you. 

MODERATOR: Great. Next question.

 Q: Hi, thanks for doing this. How do you – with these drive thru test sites being offered by big box retailers, they often ask patients to self-swab, and I was wondering if you could say how that could compare to a test someone would receive from their medical provider and, depending on your familiarity with these retailers, can you say if one retailer is offering a more reliable test than others?

MICHAEL MINA: So, we are still trying to work out just how well self-swabbing works. There is some evidence that it does not work as well. Might be, might have 15% loss in sensitivity, for example. We’re trying to work on, you know, trying to understand maybe saliva works better than self-swabbing and maybe even saliva could work better than nasal pharyngeal swab. But at the moment, we really don’t know for sure which is going to be the best. There’s, of course, been some movement to approve some of the self-swabs, but in the current climate, most things are getting approved or at least passed through the FDA favorably or at least not checked yet.

And so I think we have to be a little cautious when we interpret things that the FDA is approving at this point in time. Sometimes it’s not with the same rigor that it may be should be and that we’re used to expecting from the FDA. So we have to take it with a little bit more with a grain of salt than usual.

And in terms of which tests they’re using, in general, the big box retailers are using good – when they’re doing the viral testing, they’re generally using, if they’re doing it right there at the location, it’s probably the Abbott ID now, which has had its problems, and I think that there are still issues to be worked out. So, for example, the normal solution that the swap gets placed back into after being taken seems to inhibit some of the Abbott ID Now process and there’s a loss in sensitivity there with what we call VTM, or viral transport media. And so, if retailers – generally, retailers are aware of this, and so, if they are doing their due diligence and using the correct media types or just putting the swabs right into the assay.

Other assays that are being performed by some of the big box retailers, you know, if they’re not being done on an Abbott ID Now or something right there at the point of care, then they are generally being sent off to a central laboratory and being processed, usually from a PCR perspective viral detection, most of them, most of the assays, are working quite well. So I’m not as concerned about those as I am about both the use of the nasal self-test and the antibody tests that I was talking about before, have a lot more variability at the moment. But in terms of the actual manufactured test itself, separated from how it’s collected, those generally tend to be good for the viral detection via nucleic acid RNA.

MODERATOR: Next question.

Q: Hi, thanks for taking my question. So I know a lot of people have talked about this, but to switch a little bit back to the molecular tests. Last night, Mike Pence said that there are a sufficient amount of tests right now for every state that qualifies to enter phase one and begin reopening. So I’m wondering what your take is on that. Do we have enough tests for phase one and what will we need in the future? And because each place is so different, how do we even go about assessing what we need and whether we have it?

MICHAEL MINA: Yeah, we don’t have that. Testing is still limited.

We’re trying to even in Massachusetts, where we have a lot of testing per capita based on the capacity, still getting the test performed is – I don’t know what metrics he’s using to evaluate that, and it really depends on what role testing is playing in each state’s phase one opening. If you’re not including testing as a large part of phase one opening then sure, we have plenty of tests if you’re not requiring them. But I think most of us who are working on thinking about what are the safest and best ways to open up the economy again place pretty significant amounts of testing front and center to prevent outbreaks and monitor for them and and whether that’s the antibodies or virus. And we haven’t actually rolled out serious testing to the population. We have the drive thrus, which are relatively low throughput relative to, you know, companies having daily testing, for example, of their of their employees. So I would say no, I wouldn’t agree that we have enough testing.

It’s still a constant conversation, even in the hospital, about getting testing ramped up, and this is at Harvard hospitals, you know, and I think, we’re doing a good job. But when I look across the landscape of the country, you know, I recognize that, you know, at Harvard we have tremendous amount of resources and we can – even the fact that we’re even having to have the conversation about how to triage tests and where to send them to suggest to me that many parts of the country probably are not in a position that’s necessarily much better, and probably many are in a situation that’s quite a bit worse. A lot of places, for example, that are using LabCorp and Quest, if those were to be ramped up to sort of population scale testing in a manner that would allow people to get back to work safely, I think,  I don’t know that these that these laboratories would have the turnaround time necessary to prevent transmission So we have to be creative, but generally no I don’t agree with the statement from the Vice President.

Q: Alright, thank you. 

MODERATOR: Next question.

Q: Hey, thanks for holding this call. I really appreciate it. So as you might have seen, you know, Miami Dade County had preliminary results announced from their serological survey that they conducted, you know, countywide. About 1,400 people participated. The researchers involved from the University of Miami gave a 95% confidence interval that the infection rate was, you know, was 10 to 15% higher than reported cases. So about 6% of the county, you know. Obviously, there’s been a lot of questions about these kits and their biodynamics kits. You know, I’ve seen one white paper on their specificity being an issue, and I guess I’m just kind of wondering, you know, given the fact that it seems, you know, it seems that they have a pretty good method of randomizing participants. But they’re so questions about these tests. What questions should I be asking researchers to kind of figure out how reliable these data are as far as tell, you know, how much confidence can we put it in their research?

MICHAEL MINA: So, in general, when we’re looking at this. So there’s two parts. There’s one with regard to the tests and there’s one with regard to the populations being tested. So with regard to the test, if prevalence gets high, then you know  being off by 1% or 2% isn’t the end of the world, if true prevalence is much higher than that, but if prevalence is around 3% and you’re off by 2% obviously that can be, and you’re not specific by 2%, then that’s a big problem because that might triple the number of cases you think you have.

So on the one hand, I think a lot of these assays need to be vetted well, and there are a lot of institutions here and elsewhere that are doing this type of vetting and trying to determine which ones are the most accurate, the least accurate, and getting these types of specifics. Again, one of the important things to remember is that even with these specific characteristics being identified, it could be, you know, it’s unclear to me what the variability from lot a lot of these tests will be. They’re being produced in very high volume. And also what the effects of transportation and heat and things like that will end up being, as especially as it gets warmer and these things are being transported. You know, in high temperatures, you could lose sensitivity of proteins on the on the test unfold, and you know all these problems come into play. Now, obviously, the tests try to account for that and add different preservatives and things.

So I think in general, we just have to be mindful of what the test characteristics of each one are and what is the right environment to use that. In terms of – the bigger question maybe it’s really what populations are we testing. And that’s a very different question and a very difficult one to answer to make sure that you’re getting a true representative sample if you’re doing a prevalence study you want to, you want to really understand you know how, where can you extrapolate the results to. So if you’re testing in a very urban, low socioeconomic area with a lot of apartment buildings and people densely populated together, then you probably can’t extrapolate any results that you get from that out into the suburbs and vice versa.

And so we have to, you know, account for that statistically when we’re evaluating these. And that’s why we need to be really putting a lot of focus not just on these convenience samples. Having people goes to sit out in front of a grocery store is great, but it gives you one small slice of a population and you can really only extrapolate within that population and group. If you start doing that at grocery stores throughout the whole of a state, then that might actually be a way to go. Do it at grocery stores in the suburbs, in rural Massachusetts and urban Massachusetts and then you might get somewhere, but you still want to probably use databases to identify more more reasonable cohorts of people to follow. For example, how political calls are made. And there’s actually a lot of information about people from things like American census and other data sources that could be potentially used to try to enroll people into surveillance efforts.

So, I think that those two issues need to be taken very seriously. And the first question on the call today was about armchair epidemiologists and this is one of the big things that we grapple with is, you know, there’s a whole science behind constructing representative cohorts, and I think that people working on developing these kinds of surveillance efforts should probably be partnering with people with a lot of experience developing representative cohorts.

Q: And in this case, I mean, that what they’re doing is they’re using County Utility Florida Power and Light to randomly generate phone numbers and then they’re calling people, and so they acknowledged or some selection bias there and who volunteers who can drive up there trying to address for that with walk-up sites in areas where people don’t have as many cars. So it seems like they have a thoughtful approach to that. I guess my question is, you know, given some of the questions about the test, should they be doing validation on these tests?  Should they be, for instance, testing people who have positive antibody results, doing more serological testing on those people to give themselves more confidence about whether these tests are, you know, producing false positives and false negatives?

MICHAEL MINA: Sure. I think anyone who’s doing these tests – you know, no pathology lab would just pick up any test, even if it’s an FDA approved out-of-the-box kit, and start using it. You know, and that’s where I think some of the big problems here are coming in, where you have a lot of people who are doing testing now with point-of-care tests, they’re not pathologists. They’ve never thought about test characteristics in this way, in a really rigorous way, and that’s why people like like me and other laboratory medicine physicians who run pathology laboratories exist, specifically to to ensure the accuracy and and interpretation interpretability of tests.

So I do caution a little bit. I don’t think that we’re the only ones that can do this, but certainly any study that is doing this kind of work needs to do their due diligence to ensure that the tests that they’re using are accurate, and I think it’s a mistake, probably, to just assume that the whatever shows up in a box, you know, is working. So certainly all of these test characteristics need to be measured, even these point-of-care, and maybe especially these point-of-care instruments, need to be considered thoughtfully before just going and using them out in the population.

We see that with FDA approved assays, sometimes we get bad lots and that’s just a fact of how of materials and and shipping and lot-to-lot variability. So certainly I caution, anyone who’s going down this road from from, not, you know, I caution them to ensure that they’re doing their due diligence to ensure the accuracy of the test itself.

Q: Thank you so much for your answers. 

MODERATOR: Next question.

Q: Hi. Thank you for taking my question. Dr Mina, I appreciate you making yourself available. So we’re talking a lot about reopening in Florida on a local and statewide level and we’re kind of obsessed with measures, right. We want to know where we on the curve, and are we ready yet. And my question is more along – if there’s a yardstick that public officials ought to use when they’re deciding how much to lift social distancing measures? For instance, we heard a two week decline and hospitalizations and hospitalized deaths is what you want to see before you even consider reopening.

And then just a second part, which is related to this is as we ramp up PCR testing in Florida, how will we know how much is required to reopen safely? Is there a per 1,000 number, for instance, that we should aim for? And how would that number change in areas with dense populations and high case rates versus more rural areas?

MICHAEL MINA: Yes. And so it’s a great question. You know, what’s the number of tests that are needed. That’s sort of seems to be the question a lot of people are asking. And it’s very hard to know exactly what the right numbers are. It’s related to the prevalence that exists currently, it’s related to the serological prevalence of how many people have been infected, and it relates to how aggressive communities are trying to open up. And I would say that if a state or a city is just being very aggressive and opening up all their stores and restaurants and everything overnight, I would really caution to have a very robust surveillance testing system in place.

Surveillance doesn’t only have to be testing it can be clinical diagnostics and dronic surveillance. It can be app-based tools, having people’s cell phones work with DPHs, Departments of Public Health, but testing will likely be a big part of it all. And if, on the other hand, other places are opening up more slowly and gradually, then we need to sort of have testing in place for those places. So, in some states if they’re starting to incorporate the most essential employees back into the workforce, then maybe you just need to really be sure you’re testing that group of people, and if they start to transmit amongst each other or to other people in their families or others, then you  learn something, even with a pretty minimal amount of testing just among that group. You start to understand what is going to happen as you open up. So it really depends on a lot of factors and I don’t think there’s going to be a single number that is able to answer that question. And what was the first question you asked?

Q: The first part of the question was, whether, you know, two weeks of reductions in hospitalizations and hospitalized deaths is what you want to see before you consider phase one of a reopening, like was there a yardstick of some sort? Similar to the testing, like, you know, I asked per 100,000 but this would be in your indicators, if there’s something as a public official you ought to be looking for, you know, like I said a two week decline in hospitalizations are hospitalized death rates?

MICHAEL MINA: Yeah, absolutely. I think that having – two weeks is not sufficient, I think. I think probably at least four weeks because the generation time of this virus is about, you know, it’s anywhere from three to five days or so, and maybe a little bit more and that means a doubling time. So I think that having a two week window and saying, ‘okay, we’ve seen decreases for two weeks’ is probably insufficient to make too much of a case for. If we see a plateau for a few weeks and then a decrease for a few weeks, I personally wouldn’t feel comfortable until not just seeing a decrease, but until actually seeing a stoppage of cases.

And, you know, if we’re up at 10,000 new cases a day in some state for example, and you see two weeks of decline, but you still have 3,000 cases a day, that’s not exactly where you want to start opening up your economy. So you have to – I’m actually thinking more along the lines of, you want multiple weeks of very, very few cases at all. And then I would start to feel more comfortable saying, ‘Okay, let’s think about how to safely and slowly open backup’.

Q: Thank you. 

MODERATOR: Great. Next question.

Q: Hey, so listen you made some great calls on this early on. We talked back in late January, and you talked about how the 2.4 mortality rate was all wrong because the infection rate was 10x higher. That’s turned out to be true. You very early on, talked about overwhelming the hospitals and social distancing. So those are all great calls. So I’d like to ask you to make another one right now.

There’s a doctor up at St. Barnabas in the Bronx, in a low-income community, and he’s the chair of the E.R. Department, and he posits that the lockdown had nothing to do with the slow down. It was just that the virus spread, and he his point is that he’s got patients who didn’t have the opportunity to social distance because either they’re homeless or they work in essential jobs. So I’d like to hear your thoughts on that, like, did the lockdown really matter? And second, you can see all the moving parts in terms of what we’re doing, and the extent of testing. What’s your forecast? Does the disease come roaring back and create problems or does the plan that we have in place workout?

MICHAEL MINA: So I think the question ‘did the lockdown work?’ is really good one. I do think it did in most places. I think there are places where probably we might have actually reached herd immunity in certain communities. We see very high rates in some nursing homes, for example. You know, if you see a viral positivity in 30 or 40% of people or more at a given time, it’s very likely that maybe you’ve had actually 80% of people infected and, you know, just in that slice of time you’re getting 30 or 40%. So it is possible that in some very dense populations that herd immunity may have been achieved in very localized areas.

But in general, I think what we see from social distancing and what we see from the epidemic curves is, I think, that much of what the turnover of the epidemic curve was due to, was the social distancing that took place. It was very predictable, this whole – far as I’m concerned most of this epidemic has been very predictable, and including once we started social distancing that there was a delay in the – you know, there was a a couple more generations of transmission within the household for example, and just as things were slowly starting to shut down. And then a couple of weeks after that, we really started to see things plateau in a lot of the regions in this country and other places in the world that also are social distancing and then start to drop. And all of this happened in many places, for example, where serological prevelance is still lower than, quite a bit lower, than herd immunity threshold and and what we would expect needs to happen to really slow down the virus. So, I think social distancing was key.

I do think still that you know the numbers of people that are infected relative to the numbers of people that have been confirmed to be infected are wildly distinct and probably the number of people is, you know, many millions in the US versus the actual numbers we have. But even with that, you know, 4 or 5, 6, 10% even, you know, those are getting to very large numbers in terms of millions of Americans that have been infected, but, you know, sadly, it’s just still not enough for herd community to have been the reason for these curves to have turned over so quickly or relatively quickly for an epidemic like this.

What do I think is going to happen in the future? I do think that we’re going to see cases come back for the same reason. I think we will find that there are some places that have been hit very hard and we might actually see that they have formed some level of protective shield from herd immunity to a certain extent. Of course, there’s been a lot of talk about not knowing exactly whether this virus causes lifelong immunity. We do think, though, that this virus will cause some immunity and some level of protection in most people. The kinetics of the virus act and behave similarly with our immune system as we see with a lot of acute viruses. So, we don’t expect this will be an extraordinary virus in that way.  But even with some immunity being developed, we don’t know if it will totally block transmission And if you have 5 or 8 or 10% of the population infected, you know 10% is still 30 million people, which is a lot of people.

If we were up there, and I don’t think we are as a country, but if we were, that still isn’t enough to stop this virus from growing back and we can look at 1918 as a warning sign, where the second wave was really massive and killed a lot of people. And that could happen here because we know that even once we start opening things back up, we’re going to have a lot of individual cases that are still brewing under the surface. And so rather than having a few importations or even, you know, tens of importations or hundreds, when, at the beginning of this epidemic, which spurred the whole the whole epidemic across the country, we are now potentially going to drive cases all the way down. And if we’re not careful when we open things back up again, we have potentially thousands and thousands of little cases starting new transmission chains across the country that could explode if we’re not careful.

And so we have to be very very cautious when we start opening up and, you know, one of the other things that was also predictable, I think, and I remember saying it – one of the messaging points that really should have been made to people early on in this epidemic, or early on into social distancing, was that, you know, we’re entering into a period of social distancing. You will start to see things become calm, you’ll start to see hospitals not be as overwhelmed. And that shouldn’t be perceived as a reason to go outside and say that your part is done. That should actually be a reason to say your part is just, is being successful. And I wish that that messaging was really in loud and clear with policymakers pressing social distancing efforts.

Q: Okay, and so, will we need to shut down the economy again? Like will it be that kind of a full stop again, like a repeat of that do you think? Not just in the summer, but, you know, November, December October.

MICHAEL MINA: It could. I mean, we don’t know for sure. Depends on how people open back up and depending on how the weather works with this virus. Let’s say, for example, that the weather cooperates and it actually does help to inhibit spread and things go completely back to normal, but we still have brewing cases that happened throughout the whole summer. Then, the moment the weather turns back to a favorable climate for viral spread, we might be really taken off guard. And you might start to have rapid transmission again and a very quick explosion compared to the long ramp up we have leading into this current phase, so I worry that that could very well happen.

Q: Great, thank you. 

MODERATOR: – Next question.

Q: Hi, there. So it’s another forward looking question again regarding antibody testing, regarding serology testing. The big question for many is whether having antibodies confers immunity. And I’m wondering if you could talk a little bit about efforts to answer that question, whether it’s tracking people who have been infected over time or looking for biomarkers in their blood or any other means, you know of where we can resolve the question of whether showing antibodies, whether it’s a certain level or a certain type of antibody in your system, suggest that you in fact are immune?

MICHAEL MINA: So, we’re doing a lot of studies now – and we being the larger community of scientists – to answer these questions. And so some of the ways we’re answering them are to use healthcare workers, for example, who might be at higher risk for infection. And find people who are serologically positive, meaning they have antibodies, follow them over time for a number of months, and see at the end of a period of months, if we’re swapping them for example for virus or we’re serially collecting antibodies and looking for new spikes, increases, can we detect whether people are getting reinfected and whether they are – if they do get reinfected, whether they actually have viral titers that are appreciable? So, we are working to understand that. So virus titers are one piece, and then we correlate those with the antibody levels, for example.

But there are a lot of other efforts. I know at MIT, there’s some folks who are not quite sure what biomarkers to look at yet, but they’re looking for essentially proteomic and metabalomic data to try to correlate with protection and with response to the virus. Certainly viral viral titers are one thing that traditionally tend to track decently with protection and so we are trying. That’s one of the reasons I feel so strongly about the use of quantitative antibody testing, especially at this early phase of the epidemic, or the early phase of testing, is to be able to have the opportunity to find thresholds of protection. What is the value of the antibody that suggests that somebody is protected?

There’s also going to be T cell assays that will come around to look at maybe people who don’t have antibodies, which happens, 5% or more people without antibodies are actually maybe protected from other viruses. So we see it with measles, for example. People can get measles vaccines and form no antibodies, but still will show protection from a T cell response perspective. And so there’s a lot of new technologies that have come out in the last few years last five years, and a burgeoning field of technologies to look at as a result of COVID to understand new correlates of protection that go and extend beyond just antibody titers.

Another one is neutralizing antibody titers which is a different type of antibody measurement. It’s a measurement, not for the antibodies themselves, but for what dilution of the serum has to be made before the blood no longer neutralizes virus in a Petri dish. And you can extrapolate from that and say, “Okay, I was able to dilute this person’s blood 1,000 times before they could no longer neutralize virus in a Petri dish.” And that’s just for example, they would have had very high antibody titers, so there’s a there’s a lot of different approaches to get at this.

Q: Do you have any sense of the timeframe when some of these results could be coming in in a meaningful way and we could determine, in fact, who might have an antibody test that would give them that result that they in fact have immunity?

MICHAEL MINA: In terms of a time frame, I think we’re definitely looking at months, not weeks, because we have to wait to see if people will – well it will be weeks before we really understand people’s antibody responses. And those are coming in now. But we don’t know how to interpret those and whether they are protective requires just waiting and seeing if people get reinfected. So, to accrue those types of results takes months.

So I think we shouldn’t really anticipate anything before July that’s actually proving to be really useful. Maybe there will be some studies that start to hint – well, I’m sure there will be some studies that start to hint at it before that. But before we see a really robust reliable data, I think it will be a few months and, you know, if, if at all. For example, we know that even though we have levels of antibodies that we refer to as protective thresholds for measles for example, we also know that those levels don’t always truly correlate well with protection, but they give us some guidance and some guidelines.

So, we’re actually not sure if we’ll find a number for any of this, that really confers protection and that’s largely because everyone’s antibody and immunological response is, is very, very diverse and unique. And people – a value of 100 antibodies for one person might mean something different than a value of 100 antibodies for another person. There’s a lot of dynamics and biology that goes into those numbers and I won’t go into detail, but it’s not quite a straight linear line.

Q: Thank you. 

MODERATOR: Next question.

Q: This is a follow up on a couple things you just spoke about. You talked about the possibility of COVID-19 exploding this fall, you know, faster than the initial ramp up. Could you talk a little about what would happen in your mind, if the flu and COVID-19 both lined up together this fall, the challenges that that might present and, along with that, if we learned anything about that, since they seem to have coexisted for a while?

MICHAEL MINA: It’s hard to say how they will react and respond together. So there is some idea that maybe one viral infection at a given time could actually serve to protect against another viral infection happening concurrently. And that might be because if one virus alerts elicits an innate immune response, that might help shield cells against viral entry. So there’s some biological reason why there could be protection.

There’s also some biological reason why there could actually be secondary infections that can be much worse. And some of that – usually we think of that in the context of viral and bacterial infections, but they could interact biologically too ultimately. If one virus distracts your immune system and then the other virus comes in, it could cause some issues. There’s also, from an epidemiological perspective, one of the things if coronavirus starts to reemerge in the fall and people take it very seriously and shut things down, it could lead to a pretty remarkable year in terms of a reduction in influenza transmission.We don’t normally social distance as a result of seasonal flu, for example, but we clearly are social distancing as a result of this novel virus. And so, if we take aggressive measures because we start to see outbreaks emerging in the fall, for example, we could inadvertently prevent wide-scale transmission of influenza this coming fall.

And so there is the potential even that benefit that we would have both viruses be reduced And there’s all these other health things that are now coming out like we as a result of social distancing we actually see, on the one hand, we’ve seen increases in mortality in people because of COVID. But we’re also seeing decreases in in a lot of other infections and car accidents and all these other- homicides- public health metrics. So I think there’s multiple levels to that question. And we’re not sure exactly. It will really depend on what ends up happening and how we respond.

Q: A quick follow up, would you then advocate – I don’t know if that’s the right word – for some level of social distancing awareness, moving forward, for example, during flu season? I’ll leave it there.

MICHAEL MINA: Yeah, I think in the absent, you know, it’s like a public health person’s dream would be to have people do a little bit of social distancing during flu season or during other periods of infection, you know, largely the population has in generally taken it seriously, but social distancing could mean anything from, you know, I mean, it’s behavioral things like coughing into your elbow and, during flu season not touching not shaking hands. I think that those are types of social distancing that that can happen. That should probably happen any anytime. Do we really need to shake hands during flu season? I’m not sure.

And then certainly, I mean, taking seriously the idea that if you don’t feel well, don’t go to work. People generally don’t take those ideas too seriously. They still go to work and largely that’s because infectious diseases and public health have been largely neglected. For years, it hasn’t been on the forefront of anyone’s mind until now. And so, hopefully this will have some impact where people – people are now finally understanding the importance of public health and the importance that everything you do as an individual can actually add up to have pretty large population level effects, which is a very difficult concept to understand if you don’t live in this epidemiology world, but now everyone’s living it in real life. And so I think that there could be some benefits moving forward. People might take public health messaging a little bit more seriously.

Q: Really quick follow. I’m thinking, you might have this term in your in your head. Social distancing has been kind of an unfortunate label. Many of you have said, you know, physical distancing might have been a better term. You talked about behavioral preventive measures and public health in general, is there a term that you wish was being used to describe all these things from physical distancing to coughing in your elbow to not shaking hands?

MICHAEL MINA: Infection Control? I don’t actually have a – I think they’re all a little bit nuanced and I should think of one though. You know, I do prefer physical distancing and, you know, contact distancing maybe. Social distancing comes with all these. It comes with a lot of baggage, a little bit of emotional baggage and, you know, people think social distancing has to me and you know distancing from socialization and doesn’t necessarily need to. And so contact distancing or something like that.

Q: Thank you. 

MODERATOR: Looks like we have another question.

Q: Hey, Michael, just real quick. You mentioned that it might, during the flu season next year, it could come roaring back, it could spread faster, even faster than the first time around. Why would that be?

MICHAEL MINA: So the first time around, it was based on sort of initial transmission chains getting started some a limited number of imports from  abroad. For example, one import leads to other cases leads to other cases and they generally started in the cities, for example, and have kind of worked their way into the center of the country.

But now we have wide-scale transmission across the whole of the country. And we might push this epidemic down so that there is sort of smoldering flames, if you will, just below an R0 of one, or maybe even just flattening the curve so that it’s just at an R0 of one or above. And if we have that, then we might have wide-scale, low-grade transmission happening. And if all of a sudden, everyone sort of opens up their industries again, then very rapidly, you won’t have this long lead up where one person becomes two becomes four becomes eight, you know, or in this maybe it’s one person becomes four, becomes 16, you know, and things like that. In that case, you might start with 1,000 people instead of a few. And those 1,0000 people might be scattered throughout the whole of the United States and if we don’t ever sort of put out all the flames ,then we will have this smoldering number of people that you know who will all be able to ignite outbreaks all at once if everything kind of comes roaring back in August or something. So that’s  where that idea comes from.

Q: Okay. Great. Thank you so much.

This concludes the April 28 press conference.

Michael Mina, assistant professor of epidemiology (April 27, 2020)

Michael Mina, assistant professor of epidemiology (April 24, 2020)

Bill Hanage, associate professor of epidemiology (April 22, 2020)