Transcript
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 Friday, April 3.
Previous press conferences are linked at the bottom of this transcript.
MICHAEL MINA: Just for anyone I haven’t spoken to before, I’m an assistant professor of epidemiology and also of immunology and infectious diseases at the School of Public Health. I’m also one of the medical directors at Brigham and Women’s Hospital where I oversee parts of the viral diagnostics lab, in particular, the molecular viral diagnostics lab for clinical care. And so, that sort of puts me square in the middle of diagnosing coronavirus in patients in terms of the laboratory side of things.
And I’m also – separate from Brigham and Women’s Hospital – I’ve also been working hard at the Broad Institute to work with their team there to get high throughput testing available to Massachusetts and that’s now running.
So, I think I can speak to both sides, both- the testing has sort of been more where my where my mind has been in the last few weeks. But also, I can certainly speak towards the epidemiology and have a slightly different bend in my thinking because of my clinical hat than some of my peers. So, I’m happy to just take questions.
MODERATOR: Alright, our first question.
Q: Hi. Thanks for doing this. I just wanted to ask on the testing – I guess if we’re talking about reopening, starting to reopen some things, easing up on social distancing you know, what level of – how much more do we need to increase our testing to be able to get to that point?
It seems like, unless I’m wrong, the latest numbers I saw were like 100,000 tests a day right now, which seems pretty good, but are there underlying issues still, of like a backlog or supplies that are sort of making those numbers misleading? I guess just where do we need to get our testing to to be able to start reopening things?
MICHAEL MINA: Yeah, so it’s a great question. I think that the answer to that question is more related to sort of the unknowns right now about what this virus is doing, in my opinion, and we really don’t know if the virus has infected less than 1% of the population at this point or 15% of the population.
And of course, if it’s infected 15% of the population then, despite having huge amounts of hospital admissions and death, we are also dealing with a virus that’s less and less virulent and pathogenic than we are currently thinking. And if it’s still less than 1% of the population, then we’re a long way off from gaining herd immunity. And it is a severe infection and we have to do pretty intense surveillance to be able to open anything up.
So, I think what we actually need to do first, before we can kind of even figure out how much testing is really needed, is we need to know what order of magnitude we’re in, in terms of the spread of this virus. And that’s where I think that this next phase, which I’ve been seeing more and more than media and it’s certainly lining up with what we’re doing and in the actual clinical sphere, is moving to serology and getting exposure histories of people to know just how many people in a population have actually been infected.
Versus just, you know, we can continue testing and really drive up testing to get virus. And I think that that might be needed if we’re on the track where this has not infected a huge proportion of the population and it’s really quite a virulent virus. We’re going to have to, in the future, be doing intense surveillance for the virus and people’s nasal secretions and that might be a place we’re at. And if that is where we’re going, than 100,000 tests today isn’t close to what we need.
100,000 tests a day would essentially take six months to get 5% of the population tested and what we have to bear in mind is that many people are going to be requiring multiple tests. The vast majority of people that get a test are going to be negatives. But without having the serology, we won’t know anything about whether they have already been positive only missed the window or they’re going to be positive tomorrow.
So, I think it’s very important to keep that in mind that when we’re making these calculations, we have to fold into it the numbers of repeat tests which is going to be most people, ideally, if we were using this for an intense surveillance program.
I don’t know how likely it is. I think what we’re going to see is a big change in technology over the next couple of months where instead of doing everything by PCR, we’re going to have all of these antigen rapid diagnostic tests. They’re kind of the pregnancy model of test.
And those, I think, are what will be able to be deployed much more widely, potentially to the tune of somebody being able to order a pack of 10 and use one every week. And that is maybe the model that we would have to go with if we try to actually control this virus through testing for virus itself.
I think there’s other – we’re obviously seeing right now that there’s other ways to control the virus without testing. But it’s highly detrimental to the economy and to our lifestyles and pulls up the social fabric of society, frankly, so we don’t – we obviously can’t just keep social distancing.
But this is at least what we can do at the moment, given the lack of testing. Social distancing is essentially what we would be doing if we knew what the test results were for each person- we would be having them individually social distance, which would be better, but at the moment we’re kind of stuck in this rut that we’re in.
Q: Yeah, that makes sense. If I can just ask one more. Do you think we’re on – is there something that the Trump administration needs to be doing more to get us on that path towards serology testing or towards more testing? Do you feel like we’re on that path or do they need to be doing something different?
MICHAEL MINA: I think at this point, everyone knows that serology as needed. So, if we go back a couple of months, when I was talking with some of the big manufacturers of the tests that are for viruses, some of these larger manufacturers and the tests we’ve heard about in the news, there was still a lot of question about whether or not viral tests were going to be useful for this virus, whether they’re going to be needed in the United States.
And that’s where I think that the message from the White House and for the national government could have been much better and more accurate early on to tell companies, yes these are going to be needed. At this point, I think that all companies who have the capacity to build these serological tests are doing so as fast as they can.
Whether Congress can appropriate funds to sort of push them into the private sector to really be able to ramp up serological tests even more and get the R&D developed further, I don’t know at what capacity these companies are currently working on. But I do know that most of the big ones that do produce serological tests are working hard. Whether or not our president, you know, steps in, I think that that we’re going to see them developing it.
What I’m worried about though is that we’re going to run into the exact same problems that we’re running into with the viral testing and the moment DiaSorin or Abbott comes out with serological assays, the reagent supply is going to be limited and that will potentially become a huge bottleneck to testing. So, we have other we have smaller organizations and academic labs all trying to develop immunoglobulin assays as well, or antibody tests as well, and I think it’s hard to know exactly what role the White House should be having because everyone at this point is knowledge or sufficiently knowledgeable to know that this is a crucial need.
Q: Okay, thanks.
MODERATOR: The next question
Q: Yeah. Hi, thanks for doing this. I was wondering if you could just talk a little bit about kind of the range of kinds of serological test we’re going to be seen. And if there are going to be any differences in what they can tell us about how strong different people’s immunity is. And then if the different kinds of tests will be more or less useful across different segments of the population. So, either people who’ve had symptoms or people who, you know, have had no contact with other exposed people. Thanks.
MICHAEL MINA: So, I think that the types of serological tests that are coming out, it’s going to be kind of analogous to what we’ve been seeing happen with the viral test where we’re going to have these point of care tests, and we’re going to have laboratory tests, and it’s essentially going to reflect what we already see at the population level, I mean, in the status quo right now for serological tests that are used clinically.
So, we will see first and foremost, will see these ELISA-based tests and these are sort of tests that are done in the in the laboratory and these are going to start, probably with population screening is my guess and we really need to know population levels of immunity at the moment to understand the trajectory we’re on as a as a globe.
And to date, there’s really not a good serological study that’s been done. But, these are going to be these tests that are done in a laboratory sort of been on plates and the samples have to get shipped to them, but pretty soon after that we’ll start seeing those same kinds of tests be introduced as a true clinical diagnostic – not a diagnostic – I want to make that clear that these are not diagnostic tools – but as a clinical decision-making tool. It will be a data point for physicians and others to be able to use, both to take people off protective measures in the hospital and to let people know that their serostatus can indicate anyway, whether they’ve been infected.
It does not necessarily indicate if they are immune and can’t get the infection again. That’s going to take more rigorous evaluation in controlled studies, not necessarily controlled studies, but in terms of infections, but really monitoring people in a careful way using good epidemiologic tools.
And then we’ll also see these sort of rapid tests come about, and those might become the most useful for distributing. I wouldn’t be surprised if we start seeing them sold at CVS and things like that. And then they should be the most helpful to help just the everyday individual out of the healthcare system know their status and be able to have some idea of where they’ve been in terms of their infections, if they still see it.
So there’s two different antibodies that we’d also look for. IGM is something that comes up very quickly after infection and IGG is something that comes up a little bit later, usually a couple weeks later, and sticks around potentially for life.
So, it’s good – if you can get a test that accurately measures IGG specific to this virus and not to other viruses, then it can work as sort of a record of your immune system going back potentially years. And it works. It allows us to quickly be able to get cumulative case counts of what has happened in the past.
Coupling that to IGM gives us some idea of whether or not we’re still in the midst of an epidemic. If you get 3% of the people that you’ve been testing, or 4% of people that you’re testing, for example, with IGM then you know that you’re still in the midst of an outbreak or an epidemic. But if everyone just has IGG, then you can be pretty sure that the outbreak has passed.
Q: Yeah, thank you. That’s helpful. Thank you.
MODERATOR: The next question.
Q: The disease detectives, the people who track down contact – how important are they in tracking and slowing down the progress of COVID-19? Secondly, do they have a greater chance of success in rural states, such as Vermont or Maine, than they would in Massachusetts?
MICHAEL MINA: Got it. So, the people – yeah, so the question is, do they have a better chance of success of being able to contact trace and follow the path of an epidemic in a rural versus of metropolitan area?
Q: Yes, and also, how important are they, generally, in terms of trying to slow down the progress of COVID-19?
MICHAEL MINA: Yeah, so, I mean, the disease detectives – I guess I would consider myself a disease detective in many ways, but then there’s the epidemic intelligence service, which comes from the CDC and that’s a terrific cohort of people who work on contact tracing or get deployed quickly during new outbreaks.
I would say that they become increasingly less – individual disease detectives doing tracing become increasingly less important as the incidence or prevalence of a disease ramps up. And that essentially gets to the whole question of, you know, what’s the importance of contact tracing every person? At the beginning of an outbreak, it’s absolutely essential. Maybe you can actually prevent that outbreak from spreading at all, if you can quarantine and capture everyone who’s had it.
So that’s when disease detectives who are doing the contact tracing are just immensely important and that is one way to control the infection, but by the time you get an infection transmitting like it is in the United States and elsewhere across the globe, at this point, each individual person and their contacts becomes – you really have diminishing returns and the amount of effort that would go into tracking any one individual and their contacts at this point would be immense relative to the gains, except in places that are extremely high risk for mortality or severe disease, places like nursing homes and hospitals, where we really do want to ensure that we’re tracking every single person that we can who might have an infection or be exposed and get them to stay away from vulnerable patients or keep infected patients away from others.
So, I think that, in general, the window of time when disease detectives and contact tracing was most impactful has now passed. But what we’re going to see is, once we are kind of in the process of trying to smash this epidemic down, not just like flatten the curve, but really try to suppress it almost entirely with the idea that we can reset and start over. And at that point, I think we’ll see disease detectives and contact tracing become increasingly important again.
And it will require a huge amount of resources and additional staff to really do that. As we start to sort of reopen the economy and allow people to start moving back into their jobs, we are going to have to be doing pretty intense monitoring of populations. And that’s where I think we’ll start to see the crucial need for each of those contacts to be traced again in the future.
Q: My other point – do they have a greater chance of success in rural states, such as Vermont, for example, where they’ve got a team of about 60?
MICHAEL MINA: Yeah, I think, generally, if you can put more – if the ratio of people who are available to work on these questions is – if the ratio of those to the populace is higher, than I think that that’s generally good.
And in a rural area, we would anticipate that the virus will spread more slowly and be potentially in more discreet pockets. So, you might have a small town or you know, a family in rural Vermont or rural Ohio or wherever it might be.
At that point, I think that it’s a little bit easier to attempt to trace the family units or the social units. Versus in a place like New York City, where one infected individual can go in and infect large number of people, if they get on the subway, for example. And you’d never be able to trace those individuals.
So, I do think the successes would be potentially greater in a rural area. There are other logistical challenges of being in a rural area but slightly separate.
Q: Great, thank you very much.
MODERATOR: Hi. The next question. What is the ideal rate for public health officials to know or feel reasonably certain about the number of people affected in a community, and to develop effective interventions? People have brought up South Korea, which has tested one out of every 200 people. Is that the ideal rate? If not, what is the ideal rate? How many people should we be testing per capita?
MICHAEL MINA: Yeah, it’s a great question. So, in South Korea, if we really look at what they actually did – so it might have been one in 200 across the country. But what we should really focus on there is, how many people during the course of the major outbreak that they had, which was really centered in Daegu just, I don’t know, I think it’s a couple hours south of Seoul.
They actually tested around 5% of that population in around 25 days, which is a huge number of people. That would be hundreds of thousands of tests per day in the US or more, maybe 500 – I have to do the math. So, they test about 5% in 25 days and that seemed to be effective, whether – I don’t know, though, that we should all – I’ve been supportive of that type of model where we test sort of a large fraction population, but you can do it in more sophisticated ways as well.
With some good epidemiological expertise and study design, you can actually figure out what’s the minimum number of people that really should be tested to know what the outbreak is doing and including trying to make sure that you have representative populations included.
And the key thing is to know what sort of test we’re talking about. If it’s serology, then you generally need fewer people to a certain extent because individuals will accumulate. If it’s a viral test, then you might need pretty large numbers to get above just a couple of people who might be positive at any given time.
And so, I would say that it’s not necessarily a proportion of the population that we’re really trying to get out as much as setting up surveillance systems appropriately to know just sort of how to study it. But I think the problem that I’m having with answering the question is, we don’t know.
We can’t do a sample size calculation at this point in time to know what the proper number is until we actually have some better idea of what the virus is doing and I can’t stress that enough. We still don’t know if this virus has infected say 300,000 Americans or 15 million Americans. And those are pretty distinct numbers and until we really understand that difference, it’s very, very difficult to know just how many people we need to be testing.
In the healthcare system, it’s a little different but there we’re really doing targeted testing for control measures and that’s not just purely surveillance. That’s really knowing who needs to be on precautions, who doesn’t, who needs to stay out of the hospital, who doesn’t. And there we’re seeing actually quite high numbers of the population that are being tested symptomatically, or who are at a very high risk, positive. And so, in the healthcare setting, it’s not it’s not so much surveillance.So it’s not the appropriate way to answer the question.
And until we get a good idea of what this virus is doing, we could throw a bunch of numbers around to say how many people really need to be tested. I can pretty confidently say that it’s quite a bit more than 100,000 a day.
But what that number looks like will depend on just how widely this virus turns out to be transmitting and we won’t necessarily know that without serology or without really well-designed studies of the general population, including many asymptomatics or just individuals without any symptoms, whether they’re infected or not.
MODERATOR: OK, the next question.
Q: Hi. On that question in terms of, you know, we don’t know how many people are infected, 1 to 15%. There’s this emerging debate between lawmakers and faith communities in terms of allowing faith groups to continue to worship amid the pandemic. And now the arrest of two pastors, one in Florida, one in Louisiana for continuing to hold large gatherings, despite regulations implying otherwise.
And then meanwhile, we have these cases like in Sacramento County, California, where a third of the coronavirus cases there are attributable to faith communities, and in particular, one church. And in North Carolina they had a similar incident, and, of course, in South Korea in the early stages of the outbreak, there were attributed to, in many ways, to this one faith community that prompted the government to test everyone in that faith community.
But one of the reasons that that’s been tricky here in the United States is that some of these faith communities are very cloistered and actually are very distrustful of government and are – you know, there’s active antagonism between authorities and health officials in trying to get these people tested. So, I’m curious, is there concern within the broader health community about not only these faith communities continuing to worship amid the pandemic but also whether or not we could get a fair assessment of how many people are infected if they’re not reporting back until they show up at the hospital.
MICHAEL MINA: Yeah well, it’s a great question, certainly, you know, whether it’s faith-based communities or otherwise, political or whoever might be a congregating group. These are these are clear issues and, you know, the hardest thing is it’s not just – the regulations aren’t just about protecting those individuals who are there, but every extra person who shows up in the hospital puts everyone else at risk.
And, in particular, for showing up at the hospital with an infectious disease puts everyone else at risk and removes needed supplies from other people who also need them for similar reasons.
And it also puts people at risk because the normal function of the hospital starts to degrade as more and more people get infected and so every person that shows up with coronavirus puts the person who was just in a car accident at greater risk of not getting support and help they need in the hospital.
So, I think that we really have to look at these sorts of decisions to limit congregations of people not in much more of a public health sense than then an individual health sense, even amongst the groups. And so, I’m in support of limiting those kinds of congregations from happening because the ramifications extend well beyond those individuals.
Getting at good sense of who’s actually infected when these things happen because they stay away from hospitals, for example, in many ways, unfortunately, that’s essentially our policy today, regardless.
When people are infected, we’re pretty much saying and hoping that most people, unless they really need to seek medical care, that they remain at home if they’re very mild and, you know, just maybe have mild symptoms and really don’t need to go to a physician. We’re essentially asking people to weather the storm and hope that they don’t progress. If they do start to progress at all, then certainly getting medical attention is important. But this is one of the major reasons why, whether it’s because people are distrustful of government and sort of hospital systems in general, or just because our hospital system is strained, we are not getting accurate assessments of the population that’s truly infected.
We’re missing huge, huge numbers across the whole populace, for various reasons. But the biggest reason is we just don’t want people coming in. We don’t have enough tests to test over and over who thinks they might have a symptom.
We also don’t have enough material to do that, like these swabs have become a big problem. And we also don’t want people to have to leave their house if they are sick. And so, we’re kind of stuck in this situation where it’s very, very difficult to get a good accurate estimate of who is infected, regardless of the reason.
And that’s sort of the policy of today. I think the policy of tomorrow is going to change, in particular as we start to see home, do it yourself, sort of, at home kits. Maybe you sample yourself and then mail it in, or you take a blood pressure from your finger and mail in a drop of blood.
I think we’ll start seeing much greater ability to survey the population once we have those in place. We might even have things like drop off boxes where you can drop your sample in a box or something in a plastic bag with a package around it or something along those lines.
So, that will change. But I think that the problems that you’re describing, are, unfortunately, consistent regardless of reason.
MODERATOR: Next we have a follow up question. If the turnaround time for someone to receive a test result is a week to 13 days, is a test still valuable is taking that long to come back? Someone can still get COVID during that time.
MICHAEL MINA: Yeah, this is a huge problem. It’s an enormous problem. It’s valuable to let somebody know that they have been positive. It doesn’t have a tremendous amount of value – if it comes back negative, it’s really not too valuable. We know that there’s a lot of people who are negative one day and then turn positive the next.
So, I would say it’s still valuable, but from a diagnostic – and the way that these tests are generally being used, which is to diagnose people and then to let them know to stay away from other people, or to let hospital staff know to gown up and PPE.
What this long delay is doing from a lot of the major send out laboratories is really it is limiting the utility of these tests to be able to be used in the most efficient way, which is to very quickly be able to say this person, you know, is not infected with coronavirus and you can reserve your personal protective equipment for those patients who are.
And it’s very much harming our ability as a country at the moment to really be able to take the care that we need and to be able to most efficiently preserve the limited resources we have at the moment, in terms of protective gear.
There’s labs that are – I think that one of the ways to get around this, and the labs are starting to do it, is to triage patients and say that only these individuals will kind of jump the line and this might be individuals for whom there’s high suspicion and there’s a high amount of PPE being wasted on them if they’re if they’re negative. And so, get the result back as quickly as possible.
Unfortunately, it’s pretty difficult to do that type of triaging well. It will cause some samples to just keep being delayed longer and longer.
And it’s a real problem that we have found ourselves in. Some laboratories are working a little bit differently. So, in Boston, we decided that we didn’t, we essentially didn’t want to go that route and have to rely on Quest, and Lab Core, and Mayo because, we presume that these big labs will likely have the kinds of delays that they do. So, we took a slightly different approach and worked with the Broad to create a whole different – I don’t want to say private – but essentially, a different high throughput testing operation on robots and everything, so that we can more efficiently get samples back to Massachusetts and to New England-based places. And we’ve been able to keep turnaround times below – certainly below 24 hours. And lots of times for some hospitals, below 8 hours.
And that’s just some efficiencies that we’ve built into it. But the whole reason behind it was to ensure that these hospitals don’t have to run into the same issues.
Unfortunately, you know, it’s this problem where the wealthy and high resourced rise up and have better advantages than others. And it’s really hard to know how to equitably distribute these types of tests. And right now, we’re just working with Boston-based hospitals, which of course, I think, as the world knows generally tend to be pretty high resourced hospitals in the first place.
It leaves small clinics and small hospitals a little bit out to dry. We do take tests from anywhere in Massachusetts, though and we work with the state laboratories to make sure that happens.
But for the most part, these big backlogs that are happening, they were both expected. I think anyone who wanted to bury their head in the sand and say that this wasn’t expected is naive because there’s just such immense demand for both the actual reagents and for just getting into the queue in these in these labs.
It would be hard for the most efficient lab in the world to be able to really process these well.
MODERATOR: The next question. Do we have a full picture of how many people have died from COVID-19? Do we know if there are people who have died from coronavirus without being tested? And if someone dies of suspected COVID-19, do they conduct autopsies to rule out that cause of death or is testing limited only to the living?
MICHAEL MINA: So, it’s variable. And it also depends in part on what the resources are. Certainly, hospitals are attempting to do everything they can to diagnose a death as COVID if there’s suspicion. If there’s not suspicion, then in general we can’t treat every death as though it’s suspicious of COVID.
But I think overall, relative to our ability to detect cases and viral acquisition numbers, which is pretty poor, I think our ability to capture a large majority of deaths that are occurring due to COVID has been better than our ability to diagnose people in the first place.
How many we’re actually missing is really hard to know. I’ve heard many stories so far of there being suspicion of a death being due to COVID but not having tests available, and autopsies being done, potentially, assuming that there was COVID and so people gowning up appropriately. But then, never having a test being performed on those on those bodies.
And so, we might be missing a large fraction, but I think that, in general, the patterns that we’re seeing and the overwhelming number of deaths that are being recorded, I think that they would probably represent the majority of deaths from COVID.
And we’re probably missing some fraction, but I don’t think it’s going to be an order of magnitude might be, who knows. I mean, it could potentially be half, I kind of doubt that. I think probably it’s more in the 10s of percents.
MODERATOR: Next question.
Q: Yeah, thank you Michael for doing this press conference. I’m pivoting a little bit to supply chain. So lately, there’s been a lot of interest in decontaminating and reusing N-95 masks. There have been various methods proposed to do this – UV light, hydrogen peroxide vapor, steam heat in an autoclave, or dry heat in an oven. What are the pluses and minuses of these methods? Is Harvard considering using any of them? And if so, which one? And finally, how much of a dent will this make in the shortage of masks?
MICHAEL MINA: Yeah, so there’s, I think there’s a lot of different options for decontaminating these masks, at least from the perspective of decontaminating for the virus. And you know that in some ways, the simplest option would be to take a mask and let it sit out, if you have a sufficient supply. And you can put them on sort of a two-week rotation or something. You know, you could just let them sit there, and we know the virus generally tends to die after 10 days, probably earlier, but just to be sure.
So, the hospitals around here are certainly considering these types of measures and we’ve been discussing different approaches. Things like gas decontamination. Even things like ozone, for example, could potentially be used, UV light, heat.
They each have their pluses and minuses. Some of the issue with heat is what it will do both to the fibers, but also to the to the rubber bands that hold these masks on your face tightly. If that starts to get dried out and stretch because of the heat, that could cause problems.
Things like solvents and alcohol and those types of decontaminants, we worried that those would degrade the integrity of the fibers pretty quickly, and so we try to stay away from that type of thing. But, they could potentially work for number of reuses.
Things like gases, I think are some of the better options because they probably could do a good job at decontaminating without causing much degradation to the fibers and the filtration.
But these are – there’s a number of groups, looking at it. Our infection control group is now just starting to study it and I think there’s a group at Stanford who has been really doing a good job of looking at it.
It gets a little bit further away from my exact involvement, but I’ve been on a lot of these discussions where infection control units are trying to figure out, you know, just what’s the best way to employ these different options.
Q: Yeah. Can you address how big of a dent will this make in the in the supply problem? I mean, supposing you could find a method that would let you reuse masks five or 10 times. Would that mean you need five or 10 times fewer masks?
MICHAEL MINA: It depends on what cycle they could be in. And if you could do that every single night, you could potentially – these masks are pretty durable, so if we could find a way to decontaminate them and get them back into use the next day, for example, but with that same person, they could really become – it could be a huge help to the overall supply chain.
I think it’s one of the crucial things. The real question is, will other companies get on board to help make the masks in a more integrator supply than is currently available before these types of good infection control and sterilization methods come on board widely. I think it’s not going to be every place that can do it.
Although, I don’t know, say something like a microwave actually turned out to work really well and didn’t degrade the mask and was able to kill off virus, you know that that’s something that would be very, very available everywhere.
I think that for anyone who’s working on these problems, the best thing to do is not to look for solutions that could only take place in very specialized hospitals, but really, the solutions that will make the greatest dent will be those that can be employed at small clinics in particular in developing countries, where a lot of them are still using very rudimentary materials. I think would be optimal and that type of thing could make the greatest dent
In a place like the United States, it could be that, you know, in the next couple of weeks, we end up getting millions and millions of extra masks, because some company starts building them in huge numbers, but I do worry that that many places across the globe won’t have the availability of these masks and this type of reuse is going to be incredibly important.
So, I would just push anyone who’s really working on it hard not to focus just on what they can do in their high-income hospital, but I think it can make a really big difference at this given the state of things where PPE has become an absolute crucial need. I think it would make a really a very large difference. Quantifying that is difficult.
MODERATOR: Okay. The next question.
Q: Hi, thanks again for doing this. It’s really helpful. I wanted to ask nursing homes. In Massachusetts, I think the Globe reported last night, that more than 80 nursing homes and assisted living facilities have cases here. If your parent were in one of them, would you would you get them out? I mean, is there a way to keep people safe? Are these places just careless, or is it impossible to keep people safe in a group setting like that?
MICHAEL MINA: The easy answer that I have to that is if my parents were in a nursing home, I would do everything I could reasonably do to get them out at the moment.
But a lot of people don’t have the luxury of being able to do that. And although of course now that everyone’s working from home, it’s maybe a little bit easier to actually practically do that. And I do think that as many people as we can get out of these homes is probably better.
I think that this is an extraordinarily transmissible virus. I think it’s more transmissible than then we recognize. And actually preventing it from spreading within nursing homes is an extraordinary feat that would require – you know, we still have to have employees going out into the world to get to and from work and keeping a virus that’s potentially very, very transmissible out of any setting like that is going to be very difficult.
And we’re seeing that all over the place. We see it – we can’t keep it out of hospitals. We can’t keep it out of nursing homes. If we were testing businesses and industry, I think we would find that we can’t keep it out of those either.
So, I would never want to say that they’re being careless. I can’t say that some of them aren’t being careless. But I think that the transmissibility of the virus just exceeds what we’re able to do. And in that sense, taking extreme caution – going into rooms, keeping each individual safe while in those settings is absolutely important.
But we did see the problems with taking this approach on the cruise ships, for example, and in many ways there’s a lot of analogies to made between cruise ships and nursing homes. We’ve seen, of course, Seattle have devastating effects in at least one of their nursing homes and others around the country. So, it’s a real concern. And I think as much as we could potentially do to get people out of them would be optimal.
It’s extraordinarily difficult. Some people don’t even have families anymore and so what to do with those individuals is not well known, but I do think that placing excess resources into nursing homes for surveillance is warranted at the moment.
And in a place like Massachusetts, I think we should probably be considering doing pretty massive surveillance and, every few days, swabs of employees and maybe pooling those swabs together to be able to sample them quickly and look for any positives and if there’s a positive, then you quickly sort of pull those pools apart and look at them to find out which person was actually positive and isolate them.
But I think that there’s sampling procedures we can do that can make this much more efficient and it potentially is warranted to really do intense, intense sampling to prevent catastrophes.
MODERATOR: Next question.
Q: Thanks. I just wanted to ask a question about – in terms of trying to understand where the next big hotspots are going to be in the US, which lead indicators are most important to look at and which modeling groups are you kind of looking to? You know, who’s got the best projections right now, in your opinion? Thanks.
MICHAEL MINA: I think as far as indicators, if we’re just using testing as an indicator, it’s very, very difficult because the distribution of tests is so haphazard at the moment across the country. And who exactly is being tested and how to interpret one place that has 20% positive versus a place that has 5% positive. I think it all just comes down to how they’re being sampled.
Certainly, the best or some of the best indicators in my mind are simply the demographics and sort of density of populations. We know that cities are hit early and hard if you don’t take pretty decisive action pretty quickly.
And then we also know what I’m particularly concerned about is the is the. We know that comorbidities really lead to excess deaths, much higher rates. And so certainly the southeast of the United States and much of the Midwest where we know that areas that have much higher problems with cardiovascular risk and diabetes, these are places. I think that have potential to be hit very hard, places like Memphis and New Orleans, Atlanta to certain extent.
So, I think that those are those are places where the co-morbidities of the populace could potentially lead to significant numbers of excess deaths over what we’ve seen elsewhere, including really in most countries. We haven’t had places that are as unhealthy as big swaths of the United States are, being hit quite so hard yet, but I think it’s all coming soon.
And I think some of the early indicators are certainly uptics. We can try to use testing. And if we had better testing, I would feel more confident. Testing as in many ways is what I focus on and how to use that epidemiologically is also what I focus on, but I have little faith in the types of tests we’re doing now and how it’s being reported.
We know that it’s not really- it’s not serving as a terrific indicator of what’s happening at the population level right now. And so, in that sense I think really monitoring hospital admissions. unfortunately, kind of misses the boat. By the time you’re seeing hospital admissions, you’re at least a couple of weeks behind.
But it’s what we can look towards for more accurate readouts of where this is hitting currently and where to really start in reinforcing control measures.
Potentially, testing will continue to become more widely available. Of course, we know that it’s primarily available in the east and west coasts and places where big academic centers tend to be and that isn’t exactly helping the huge portions of the United States to get tests to be widely available.
Over the next few weeks, I think we’ll see testing continue to increase, but again we can only scale up the actual instruments so much, so until there’s new types of tests, we’re going to have trouble doing that a great speed.
And the groups that I think are doing the best – I think that are focusing specifically on the United States. I mean, there’s not a lot of groups who are doing it. There’s a lot of models out there, but I think of the folks here in our groups at Harvard and the Center for Communicable Disease Dynamics, I think, are putting together some good models. The Imperial and Oxford groups, I think, do a good job. And Johns Hopkins. Some in Seattle.
These are sort of the big powerhouse groups of modelers and where they end up concentrating and I think we can look to the models that are being built today to understand what’s happening, but we also have to be very wary of all of them right now because all of them, in some way, are based on the data we know.
And that, from my perspective – I feel from wearing not my epidemiologist hat but wearing my laboratory and testing hat – I feel that the data we know from testing anyway is generally flawed. And to the extent that I worry that it’s largely misleading, more than anything else.
But there’s nice study designs and models that are now really leveraging and utilizing things like hospital admissions and excess deaths to understand where the real outbreaks are happening.
And those models I think will start to see an increasing number of as people really take heart the fact that the growth of the epidemic based on test numbers is a flawed metric.
MODERATOR: The next question.
Q: Hey, thank you. Thank you for doing this. Can you discuss – I mean this is spinning us for a bit into the future – but can you discuss what level and type of testing and contract tracing would need to be in place before any regions might consider, be in a position to consider relaxing at the shelter in place standards?
MICHAEL MINA: Yes. I was touching on this a little bit earlier, I think that the very first thing we have to do is just get a sort of order of magnitude understanding of how many people have actually been infected. We really don’t know if we’ve been 10 times off or 100 times off in terms of the cases. Personally, I lean more towards the 50 to 100 times off.
And we’ve actually had much wider spread of this virus than testing and the numbers are really giving us the moment. And that’s generally a feature of just extreme limited numbers of tests.
So, until we really know that, I think we truly can’t make any good decisions that are based in real data about what kind of contact tracing, what level of effort is going to be needed.
And it kind of splits. In a way, the more transmissible this virus is, if today, we actually have 10 or 15 million people in the United States infected or who have been infected in the past, that puts us on a very different path than the one that most presume we’re on at the moment.
And if that’s true, then on the one hand, it’s much more difficult to contain the virus. But on the other hand, it means that it’s really just a razor thin slice of infected who are resulting in this sort of large number of hospital admissions and problems, and this could ultimately shake out to be not as bad of a virus as we are currently thinking. But just the sheer absolute numbers of people getting infected is what’s driving the overall hospital admissions and the healthcare crisis at the moment.
And on the other hand, if it’s a smaller number of people who are infected, that means it’s more virulent, it means we come out of this with less population-level immunity, less herd immunity. And it means that we have to do, perhaps, much more intense contact tracing to ensure that individual outbreaks across the population don’t take hold. And essentially, it becomes a game of suppressing the virus as much as possible until a vaccine comes out or, really, riding the wave of just having R0 right about one and that’s the whole idea of flattening the curve, to essentially try to slowly accrue immune people who are at less risk, so younger individuals in particular.
But those are both to say that just getting an actual number of what type of, even what magnitude of response in the future we need to have, is going to fall out of the serological surveillance studies that are going be coming in the next weeks and months.
Until we get those and until we really have a good understanding of just what population proportion has been infected and then trying to plan exactly what the response in the future, come June or July or August, is going to look like for surveillance is a pretty difficult task.
We can look at other countries – and I think given the testing situation in the US, starting to really focus intensely on what other countries are experiencing can help. Iceland is right now doing a really nice job at trying to sample, I wouldn’t say a random selection of the population, but through volunteer testing, they’re sampling people. And they’re finding as many as 1% of the volunteer individuals are showing up positive for virus.
And so, overall when you read the reports, it says, you know, 0.5% to 1%, less than 1% essentially, are positive. But that has to be taken very carefully. That doesn’t mean that less than 1% of the population has gotten infection. That’s daily incidence.
So, 0.5 or 0.7% of the population has the virus in their nose at any given time, for even just a couple of weeks, that accumulates very quickly and puts us more on the track that we have millions and millions of people in the United States infected are already.
And so, I think there’s some evidence accruing that we might be on that path. But we really – but even those studies have a tremendous amount of bias and we have to do the appropriate serological studies and surveys to really understand this.
Q: Great. Thank you so much.
MODERATOR: Next question.
Q: I really appreciate your information here. We were looking specifically at Ohio and Michigan. And the states have, you know, similar populations and are neighboring. But Michigan’s numbers are about four times as many cases and more than four times as many deaths and we were wondering if you’ve looked at that at all, or if there was any information that what Ohio did to restrict movement early is really paying off now?
MICHAEL MINA: I haven’t personally looked at that. It’s possible. I also wouldn’t want to make too much of a comment on it at the moment until I do look at the data, because I don’t know if those numbers are based in deficiencies in testing or differences and testing and the way that the testing is being performed to look at numbers of cases and deaths and positives.
I think it would take a little bit more evaluation tool to do a comparative analysis between the two before I would want to make any assumption that it had to do with the policies that were put in place earlier in one place over the other.
They can be – the differences in the magnitude of testing has led to pretty extraordinary differences in what ends up being reported and what’s been known. And there’s incredible heterogeneity across the states at the moment that needs – I think, from an understanding of the virus perspective, it can actually really be utilized to our benefit.
The variability actually can be beneficial to understand what’s what works, what doesn’t. What testing approaches seem to give more or less accurate numbers. But I think before I can really comment on that, I’d have to see the data.
Q: Okay. Thank you.
This concludes the April 3 press conference.
Joseph Allen, assistant professor of exposure assessment and director of the Healthy Buildings Program (April 2, 2020)
Caroline Buckee, associate professor of epidemiology and associate director of the Center for Communicable Disease Dynamics (April 1, 2020)
Bill Hanage, associate professor of epidemiology (March 31, 2020)