Coronavirus (COVID-19): Press Conference with Marc Lipsitch, 03/20/20


You’re listening to a press conference from the Harvard T.H. Chan School of Public Health featuring Marc Lipsitch, professor of epidemiology. This call was recorded at 11:30 am Eastern Time on Friday, March 20.

MARC LIPSITCH: Good morning, everyone. Thank you for joining. I’m going to speak only very briefly and then mostly take questions. There’s lots of things to highlight potentially. But to me, it seems that the perhaps most pressing issue as we’ve all been hearing about is that even at the very outset of this epidemic personal protective equipment is running low in hospitals around the country. Health care workers are having to reuse it or not even have it. And calls are going out to seamstresses around the country to make the cloth masks.

This is not an adequate national response. And of the many things that need to happen, I think perhaps the most pressing in the coming days is to massively increase our production of personal protective equipment. I’ll point you to an op-ed that I wrote with the former Secretary of the Navy Richard Danzig that appeared in Bloomberg opinion this morning, highlighting five areas for action over the next 12 to 18 months. But starting immediately for the federal government to take the first one is to upscale the production of serologic testing, which would allow us to identify people who are at low risk of reinfection or should be protected from infection, and allow them to get back to work. Along with also allowing us to track the progression of herd immunity and other benefits.

The second is scaling up onto a really wartime footing the production of PPE, ventilators, and other items. There’s been some talk from Washington about doing that, but it doesn’t actually seem to be happening on a real-time basis. The third is to protect critical infrastructure and make sure that we have backups for the crucial items of food production, power production, water, transportation, and cyber infrastructure.

The fourth is to begin now to work on election planning for November, so that we can have a democratic, fair, and open election, regardless of whether contagion is still present, as it very likely will be. And the last is to make a national effort on compensating for the educational losses that are beginning to happen already, as well as the nutritional and other secondary effects of closing schools. This is the sort of thing that really does not need to happen in every state separately, in every locality separately, with papered together solutions. This is really a national emergency on all of these fronts.

So those are some of our recommendations. Those are our recommendations. I’m happy to take questions about those, but also more broadly about what’s going on with COVID. I’ll start questions now.

OPERATOR: First participant, your line is live.

Q: I had a question relating to– I know you’ve also been working on modeling this. Do you have any updates on the modeling for where things look like they’re going? And then also since I know you began doing some modeling, I believe, in mid-January, when did you first have any communications with any of your federal counterparts regarding the possibility of this pandemic really being a pandemic? Could you speak to that? Thanks.

MARC LIPSITCH: Sure. I’m not exactly sure when my first communication with federal counterparts was. I’d have to go back through my emails and figure that out. And I don’t believe it was very early– well, I should look, because I’m not sure. Everything’s moved so fast that the sense of time is one of the first things that goes haywire. But I will try to get back to you on that.

In terms of the modeling, there was a presentation by a number of groups to high ranking federal officials yesterday morning. Was it yesterday morning? No, Wednesday morning. And there was a, I would say, fairly unanimous view that for countries with an intensive care infrastructure to protect, there are two not very good options. One is to continue social distancing, to put it in place where it hasn’t happened is as intense a way as possible to try to reduce the surge on intensive care, and then perhaps to let off at some point if the epidemic comes more under control, but that that let up would be brief, because the epidemic would come back. And we would have to restart those measures relatively soon.

Different groups have different estimates of how soon. But the problem is that without significant herd immunity in the population, resurgences are relatively quick. And the alternative is not to have this kind of social distancing in place or to have it inadequately, and to overwhelm the intensive care units. Those are both really quite unacceptable solutions, but are the two scenarios that people believe are possible in the existing landscape.

The advent of a very effective antiviral drug or other form of treatment would change that potentially, although it would have to be very good. The ability to really get cases down to minuscule levels, as has perhaps happened in China, would mean that we might be able to re-implement the sort of case-based containment measures that have been more successful in some smaller countries, though the fear of imported cases would make that very, very challenging. Increasing intensive care capacity would help not only to cushion the blow and to give us a little more breathing room, but would also allow us to tolerate more cases before the intensive care units filled up in this sort of on and off scenario, and would thereby accelerate the acquisition of herd immunity.

All of these are really bad options. But making them less bad by some of those types of interventions would help. In order to do even this sort of on and off scenario well, we need massively more testing than we have in the United States. That is still inadequate. And it’s not only the test reagents, but also the RNA extraction kits and the swabs that are in shortage in many places, as well as the personal protective equipment for the testers.

So, this is a real failure of organization by our federal government, in my opinion, and really needs to be addressed as soon as possible. Because even the bad solutions will be impossible without. And we will just end up not being able to control spread, is the concern. And I think that was the general consensus of the various modeling groups from around the world that presented.

OPERATOR: Our next question.

Q: Hi. Thank you so much. There have been some early numbers on the numbers of cases in the United States, I guess, with the severe symptoms by age group. And does any of that mean anything at this point in terms of what age groups are getting it? Or is it just too premature to make any real– like you see some of these headlines saying, oh, more young people are getting it than in China. But it seems like we have no idea what the denominator is yet, right?

MARC LIPSITCH: Yes, I was, I must admit, surprised to see that report come out of the Centers for Disease Control. It’s a fairly standard principle in epidemiology, not just infectious disease epidemiology, that if you have different data sets, different groups of people who have been ascertained in different ways– for example, cruise ship passengers on one hand, and people who have been ascertained because their symptoms were severe on another hand, and people who are contacts on another hand, you don’t combine those into a single analysis. Because there’s no interpretation of the proportions of people in different groups that is straightforward with that.

So, I was surprised to see what seemed to be an analysis that put together the different US cases and made conclusions about the proportions of severe cases by age group or the proportion of different age groups among those severe cases when one of the important criteria for getting into the study itself was to have severe symptoms due to our very limited testing. So, I would say it is premature to conclude much from those findings.

Q: Thank you.

OPERATOR: Our next question.

Q: Hi, thank you very much for taking my question, for doing the call today. You mentioned that, obviously, a key to the public health response here is developing an effective antiviral. You don’t sound optimistic that that’s something that’s in the offing. I realize it’s not necessarily your area of expertise, but can you give us a kind of a state of the current research landscape at this point, given that some of the results we’ve seen so far have been mixed?

MARC LIPSITCH: Yeah. It is not something I’ve followed closely just due to a lack of time. But I understand that there was a report of an unsuccessful trial of some HIV drugs yesterday in the New England Journal of Medicine. I know that there was a report in The Guardian that a Chinese group had found success with favipiravir, a Japanese drug that was designed for flu. But I haven’t been able to find the data or the scientific source of that report yet. Maybe some others have.

I know that there are a number of clinical trials going on. I’m not knowledgeable enough to have a meaningful level of optimism or pessimism, except to say that even very good antiviral drugs are often not effective in cases that have been infected for a long period of time or have minimal effectiveness. So, tests of antiviral drugs in severe cases may fail to find a benefit, even if there would be a benefit at earlier stages.

So, I think there are a lot of challenges of getting these drugs evaluated. I’m sure that many excellent efforts are being made. And it’s not my area. So, I think I’ll stop my comments there.

Q: Thank you.

OPERATOR: Our next question.

Q: Thank you for taking my question. So, in Florida we’ve seen a high number of our cases with no known links to travel. The governor for a while resisted describing that as community spread and has since said there is evidence of community spread in Broward. But we’re seeing this phenomenon in Miami as well. What would your recommendations be for local health officials here as to how to best deal with the spread of this virus if we don’t have enough testing, really, to know where it is or where it’s coming from?

MARC LIPSITCH: I would also add that, at least based on what I saw on television last night, the networked thermometer company whose name is escaping me right now also shows that Florida is a big hotspot of atypically high fever activity at the moment. And maybe you could chase that down, because I don’t remember even the name of the company. And that was widespread in Florida based on the map that was shown on Rachel Maddow last night.

And it’s not unprecedented, I should say, for the scientists to be getting their information from news programs in epidemics. That was one of the features of SARS and, to some extent, also in 2009. So, the news travels both ways.

But to answer your question about countermeasures, I would say first that high levels of syndromic disease at this point of fever and cough, which in most states are monitored by emergency departments and other systems should be indicative that something is very strange. And the fever data from the company add to that concept.

It’s not cold season. It’s really very, very late for flu season. And so, it’s a fairly safe assumption, although not certain, that a significant proportion of that disease is COVID-19. So, I think treating it as if that was the case would be very wise. And certainly, that’s what other jurisdictions around the country have done. And that means implementing social distancing measures in a fairly intensive way, in an intensive way, and preparing to build up capacity and building up capacity for a health care surge.

Florida has, obviously, a number of elderly people who are going to be at high risk for severe outcomes. And one of the very clear lessons from Italy, and from China, and from other places that have been hit hard is that there’s a long delay of three or four weeks between implementing control measures and seeing a downturn in demand on intensive care. Because infections now lead to intensive care needs two to three or four weeks from now. So, it is really bad public health practice to say, we’ll wait and see if we have a stress on our system, and then we’ll act. Because then you have three more weeks of exponentially growing demand. And that is not good.

So, I would say, without knowing the details of the data in Florida, that there are a number of signs indicating that more stringent measures are needed. And also, that if the testing remains limited, there are ways to do surveillance with tens or hundreds of tests per day among mild cases of respiratory infection in order to ascertain the trajectory of the epidemic. And those sorts of testing combined with syndromic data should be very helpful in verifying whether there is indeed an unusual level of this syndrome of fever and cough, and what proportion of it has to do with COVID-19.

OPERATOR: Our next question.

Q: Hi, Marc. Thank you for taking my question. As you know, many people continue to dismiss the threat of COVID-19 and resist social distancing in part because they continue to compare it to the flu. I wonder if you can address the comparison between the two to overtly explain why COVID-19 threatens to overwhelm health care systems here and abroad while flu doesn’t.

Using Italy as an example, in 2016-2017 in Italy, there were 5 and 1/2 million flu illnesses, 24,000 deaths, compared to 41,000 cases of COVID-19 and 3,400 COVID-19 deaths. I understand that COVID 19 is at least as infectious as flu and potentially 10 times deadlier, but at this stage, why are the smaller numbers of COVID-19 overwhelming Italian hospitals, which are a harbinger for the rest of us? Is it the speed with which the cases are blowing up? Is it because progression of flu is slower? Is this an example of how flattening the curve makes a difference? Can you just kind of talk about that, please?

MARC LIPSITCH: Yeah, I think it’s a combination of things. I know you emailed this question, and I’ve been thinking about it. So, the total number of deaths for flu is stretched out over a whole flu season, which is typically a couple of months. The deaths from COVID probably are not even the full number that will die from the number of cases we have now. Because it takes a long time, several weeks at least, typically, between becoming infected and dying.

So, it is indeed the regional concentration. So that certain health care systems are being more overwhelmed than others. It’s a temporal concentration of having them all build up in a very short time period. And it is, indeed, the fact that, per case, the risk of needing intensive care and the risk of dying are considerably higher for COVID than for flu.

It’s also the case– I don’t know how those flu deaths are counted in Italy. There’s a whole industry and business of trying to figure out how many people die from flu. Because most of them are never diagnosed as flu. We think many of the respiratory and circulatory deaths in the winter are due to flu. I’m not sure how the counting was done in Italy.

But that’s another piece of it, is that the COVID cases are in intensive care with respiratory distress and related symptoms for a long period of time, whereas at least a number of the flu deaths are probably acute myocardial infarctions, heart attacks, strokes, and other things that sometimes don’t keep you in intensive care for as long. So, it’s not just the number of people, but it’s the duration for each person.

And based on the limited data that I’ve been able to look at from China, it looks as though people stay in intensive care really for quite a long time, typically, with this disease. That’s part of the answer. I’m not sure it’s the whole answer. But those are some of the factors.

Q: Thank you, Marc.

OPERATOR: Thank you. Our next question.

Q: Hi, Marc. Thanks for talking with us this late. My question is this. Over the last two weeks we’ve seen the maximum recommended size of gatherings go down from 1,000, to 250, to 50, and finally this week to 10. And I want to ask is, is there any objective evidence-based way that we can arrive at these recommendations? And can mathematical models predict the results of these interventions?

MARC LIPSITCH: That’s a great question. I think we can do some sort of semi-quantitative predictions where the key idea– and this has been emphasized by some of my colleagues– I can’t actually remember which ones– on a recent exchange. It might have been Alex Vespignani who made this point. The number of transmission opportunities in a gathering is proportional to the square of the number of people. Because if I can transmit to you, if I can transmit to each of the other nine people in my gathering and each of them can, then there are sort of 10 times 9 opportunities for transmission.

And so, the emphasis on making it smaller is really partly motivated by that, partly motivated by the notion that, if super spreading is important, then I think it still remains to be seen how important it is for this virus. A big event is an opportunity, if someone is very infectious, to infect quite a large number of people. But I think that, essentially, in my view at least, the declining number of recommended people is a way of signaling being more and more serious about the need to socially distance. And I’m not sure that there’s a particular number that is magical.

Q: Thanks.

OPERATOR: Our next question.

Q: Hi, Marc. How are you? Have hospitals and medical practices been given any clear, practical guidelines on how to define non-essential surgery? And for patients who are choosing to move ahead with, say, a cancer surgery or reconstructive surgery, could having surgery at this time potentially raise one’s risk of contracting the virus, either through environmental contact, or through surgical incisions, or even during the recovery period from surgery when immune function is maybe less than optimal?

MARC LIPSITCH: I don’t think I’m qualified to answer that except in very general terms. That hospitals are going to be places where there are people setting virus more and more, both because they’re gatherings of large numbers of people. And some people who are asymptomatic will be there. And also because, obviously, sick people come to hospitals. So, I think the general notion of trying to avoid hospitals is even better advice than avoiding shopping malls or bars.

I don’t know enough to give you a good answer about the particular risks of different kinds of surgery. So, I will defer that question.

Q: OK, thank you.

OPERATOR: Our next question.

Q: Hi, Marc. Thanks for taking my question. I have a question about the Science paper that appeared Monday about the undocumented infections that drove 86% of the cases in Wuhan, China. I wondered if you’ve done any thinking about that and what it implies for the general characteristics of this infection and how it’s being given epidemiologically?

MARC LIPSITCH: Yeah, I think it’s clear that undocumented infections– which is a funny word. Maybe I would just say non-ascertained infections. It almost sounds like we’re talking about a different issue when you say undocumented. So, I would say infections that have not been ascertained are clearly playing a role in different settings. The size of that role is quite dependent on the setting. And there was a particular model structure that was able to estimate that 86%.

But I think increasing amounts of evidence, most of it still indirect, from analysis of who transmits to whom among places where contact tracing is still working, or was still working for a period, show that a substantial portion, almost up to half, of transmission seems to happen in the pre or right around symptom onset period. And that is based on a preprint from the group in the Netherlands led by Jacco Wallinga at the RIVM, which is their National Institute of Health– one of the best modeling groups in the world. And their estimate is– I think it’s around 40-plus– between 40% and 50% of transmission is pre or right around symptom onset.

So, it’s clear that that is a problem. It’s also clear that in Wuhan there were cases who just didn’t get tested because they were so overwhelmed as a system, which might be pushing it up to even a higher level. So, I think the practical implication is we need much, much, much greater testing capacity, so that we can ascertain cases.

And sort of counter to that is the example of the town of Vo, Italy, where they had such massive testing capacity relative to their small population that they were able to stop transmission simply by testing even healthy-appearing people. So, it just highlights that, without testing capacity, we have at least one hand tied behind our back in the control efforts.

Q: Thank you.

OPERATOR: Next question.

Q: Hi, Marc. I want to backtrack to something you said earlier about serologic testing and getting people back to work. I wonder if you can talk a little bit more about that. This would be a whole different– if I understand it, a whole different range of testing to find people who have already been infected and recovered. And do you have any idea at this point in the epidemic here what that population may be in the United States?

MARC LIPSITCH: Yeah, that’s a great question. I do not know the answer to that. I will just elaborate a little bit more on what the concept is and what the challenges of it are.

The concept is that if someone has previously been infected, they will have an antibody response. Potentially even if they were never symptomatic at all, they will, nonetheless, have a protective immune response, which can be detected by measuring antibodies in the blood or sometimes through other body fluids, so it can be done at home, according to some procedures that have been proposed. The real question, of course, is whether those antibodies are– whether the presence of antibodies really does signal that you are immune to becoming infected and becoming infectious to others. So before doing that on a large scale, it would be necessary to check that.

But the value would be potentially tremendous if we had a workforce that was growing as the epidemic progressed that could safely go back into various jobs, including health care workers. But I think, at the moment, we don’t really have an estimate of that number. Short of having widespread serologic testing, one thing that I think in some places is being done is to keep records of who has tested positive at what time, on what day. And so, as we begin to understand the natural history of this infection better, it may be possible to just use that as a proxy for having been infected. I mean, it is evidence of having been infected. As a proxy for being immune.

But at the moment, this is a somewhat speculative strategy. But if we’re going to keep our economy working at all, I think it will be very helpful to have some people– and if we’re going to keep health care working at all, it’s going to be helpful to have some people that we know are immune, if such people exist.

Q: And so, would that necessarily have to come– since we’re so far behind on testing for active cases, this is after that?

MARC LIPSITCH: Well, you don’t need one to do the other. I mean, you need it to validate that the people who show antibodies on your serologic test, in fact, did– at least some of them did shed virus before, so that you know that it’s detecting something real. But there’s no reason why we have to wait for one to do the other.

And in fact, there have been reports from China of home-based sero diagnostic kits that you can– I think it must be with saliva. I need to read them more carefully. But there are innovations happening in other countries. And I know there are some efforts underway is here as well. But on that, as with the virus testing, we’re behind. Both at once.

OPERATOR: Our next question.

Q: Thank you. I’d like to ask even a bit more about the serologic testing. So what level of effort would that require just in terms of the science of it? Do we basically already have the ability obviously, and we just need to ramp up? And if so, what kind of an investment are you imagining?

And then I just have to really ask the really obvious question, which is, you’re coughing. Are you OK? It sounds like a dry cough.

MARC LIPSITCH: Thank you, yes. I have had this dry cough for about a month. And I think it’s from fatigue. I have checked my temperature repeatedly. Thank you for your concern.

Q: Good.

MARC LIPSITCH: And it doesn’t seem to be getting worse. So, I think it’s something else. I think it’s lack of sleep. In terms of the serology, I am struggling to keep up with the information. But Florian Krammer at Mount Sinai in New York has recently deposited a preprint that shows a serologic test that was developed in his laboratory for this, which is, I think, not very high throughput, but is effective. And he’s been very active on Twitter explaining what the strengths and limitations of that are, and the uses of that. So, I would recommend looking at that.

As I say, my understanding is, from news reports, that both China and Singapore have developed technologies, including more high throughput and sort of user friendly technologies. There was an article in Science magazine in the news section at the end of February about that. And I’m struggling now to keep up with the developments in China. But my understanding is that there are some quite usable approaches that just would have to be manufactured on a large scale.

Q: Thank you.

OPERATOR: Our next question.

Q: Marc, thank you for your time here. And I appreciate your article in The Atlantic, which was so helpful. We represent hundreds of large health systems, and nursing homes, and senior living companies. So, my angle is wanting to understand what this means for the health care system.

I guess– so many questions. But the one key one is, this coming surge, do you have any sort of sense of what we’re talking about in terms of how much capacity– how big a bottleneck will we see, say, in ICU beds, do you think? And do you have a sense– will it play out that in certain geographies like Manhattan it’ll be quite constrained, but maybe in some other geographies maybe not quite as constrained? And when do you think the sort of peak will be hitting acute care?

MARC LIPSITCH: Many good questions. Very hard to answer for a number of reasons. So, let me tell you what I think can be said. Comparisons between places are very hard to do. But we have a preprint on the Harvard Dash server, if you type in It’s the first preprint that comes up, because turns out more people are interested in this topic right now than other areas of scholarship, which I wish– I can’t wait to become obscure again. But it’s the first one that’s listed on there. Or you can search for it– that compares what happened in Wuhan to the US intensive care bed capacity.

So, Wuhan shut the city down on January 23 when they had about 500 or so confirmed cases and 23 confirmed deaths, or 20-odd confirmed deaths. They had their peak demand for intensive care four weeks later, right around the same day in February, around February 23 or so. That was on a per capita basis equal to the number of ICU beds per capita in the United States, empty and full. Which is about, if I remember correctly, about 2.8 per 10,000 adults.

So, if you took that literally and said that a city that implemented very intense social distancing measures when there were 500 confirmed cases in the city would have that kind of spike, you might be in the right range. Differences include, of course, that they were probably testing and confirming more cases than we. So probably, in fact, the relevant number for the US would be smaller than that in terms of what number of confirmed cases would correspond to the same point in the epidemic. Yet another reason why our testing inadequacy is going to cost lives.

Other aspects that might not be comparable include that Wuhan probably was able to do better social distancing or more effective social distancing than a typical American city, due to the governmental system and also the degree of– well, yeah, the, essentially, consequences of the kind of government there is there. And there are also, obviously, many other differences.

My colleague, Caroline Buckee, who will be on one of these calls in the next few days is working on a county level surge model to try to look at this in more detail based on local age structure and local ICU bed capacity. And I mean, to add on yet more levels of concern, I think this is what we might be looking forward to in the next few weeks. But if we take social distancing measures away at some point, we may experience repeated ones, repeated surges like this.

So, the last thing I’ll say is that, if you look around the country, there are clearly areas with a lot of known cases. And then there are areas like Florida with a lot of fever that’s unexplained and not a lot of confirmed cases because of limited testing. And then there are areas that seem to be less affected. And I think that is what you expect as an infectious disease epidemiologist, that communicable diseases take off with some randomness. So, some places take off early and some places take off later.

And because exponential growth is like compound interest and time is what matters, you could very easily have two places that did everything the same, but just one got lucky and the epidemic didn’t take off. There was no introduction that led to more cases for a month or so after someplace else that was unlucky and had it start early. And the place that was unlucky would have the same problem a month earlier.

So, I think as we see things getting worse in Seattle and New York and some other places, it would not be the right conclusion for other parts of the country to say, oh, we missed it. It’s not going to happen to us. It’s completely expected that it might happen, just at a later date.

Q: That’s great. Thanks so much.

OPERATOR: Next question.

Q: Hi, I’m was wondering, you were talking about people who might be particularly infectious. And I was wondering if– we’ve seen news reports of a case in South Korea. A woman who allegedly was– I think they refer to her as patient 39. The stuff that was reported about was pretty anecdotal. Have you seen anything more academic about that case or other cases that really kind of suggest how significant the potential is?

MARC LIPSITCH: No, I haven’t. And again, everybody is struggling to keep up. That is not something that I have been able to keep up with. So, there may be stuff out there, but I haven’t seen it.

Q: OK, thanks.

OPERATOR: Our next question.

Q: Hello, I would like to know if we already know if kids play an important role in the chain of transmission.

MARC LIPSITCH: If kids play an important role?

Q: Yes.

MARC LIPSITCH: I believe that is still uncertain. There is better and better evidence that children do, in fact, get infected, and that they shed virus, at least enough so that viral tests turn positive. And the best understanding as I believe at this time is that we failed to see that early on because children were milder, and therefore, didn’t get tested. But analyses from Shenzhen in southern China in particular have suggested that, that children can get infected and shed.

What role they play in transmission is hard to say. And you could imagine it being greater than flu. So, for flu, one reason to know the answer to that question is, of course, whether closing schools is important. If this were flu, we would say yes.

If we think about flu as an anchor and try to compare it, you could make arguments in either direction. You could say, well, if children are more mildly ill, then they are not coughing and sneezing. And therefore, they are less likely to transmit the infection, even if they have virus in their nose and throat. On the other hand, they are also less likely to take precautions. And people around them are less likely to take precautions around them if they’re mildly or asymptomatic.

So, I don’t even believe that we know which way that particular aspect would play in terms of infectiousness. So, I think we are still at an early stage. And perhaps the way we’ll find out is when we compare places that do and don’t have school closures among their suite of interventions. Although those places are likely to differ in many other ways.

A better way, actually, to find that out would be in targeted epidemiologic studies of households where children become infected and adults are monitored. But you have to ascertain those children somehow. So, it’s challenging. It, again, relies on very widespread testing.

Q: OK. Thank you.

OPERATOR: Next question.

Q: Hi, thanks so much. I’m just wondering, so much of this is relying on testing, widespread coronavirus testing. And I realize there’s a number of hiccups that caused a delay in getting our testing rolled out sooner. But we’ve been hearing sort of for the last couple weeks from the administration in different places that testing is going to ramp up. And yet, we still haven’t seen. Is there an understanding that you can share of why we’re still not testing as widely as we really ought to at this point?

MARC LIPSITCH: It’s a great question. I think because the resources to make testing ramp up have not been put in place, is one problem. And so, I think the regulatory blockages have been, as I understand it, largely solved, although not completely. But that part is less the blockage. And now it’s simply laboratory capacity and throughput, and the availability of all the pieces that you need to do testing.

So, you need a swab. Those are in short supply, an actual shortage in some places. In fact, you need two swabs, at least. You need the RNA extraction kits. Those are also in shortage right now. You need health care workers with adequate personal protective equipment to take the test, to take the swabs. Those are in short supply. I mean, the PPE is in short supply.

So, I mean, the administration has made repeated assurances that we’re scaling up, and then failed on multiple levels, at least three or four different levels to make that happen, make the conditions possible for that to happen. So, I think it’s an example of politicians promising something that they are not delivering, which happens in other spheres of life as well, I’m told.

Q: Wait, and so if it’s a failure there, I mean, what could or should they be doing to help curb these shortages, for example?

MARC LIPSITCH: Well, as we recommend in our piece this morning in Bloomberg, treat this as a true national emergency and activate– those companies that can produce PPE should be producing nothing but PPE, or nearly nothing but PPE. Those companies that can produce the RNA extraction kits need to be organized in a wartime kind of footing to manufacture those. And similarly, for the swabs and the other pieces.

And then getting the capacity within labs is a harder challenge, but there are– I mean, we are the world leader in biotechnology. This should not be an insurmountable problem, especially when you see a small town in Italy that can do it that has no industry locally. And I mean, that small town has no particular biotech industry. But other countries are organizing themselves. And it’s just an extraordinary failure by our country not to do so.

Q: Thank you.

OPERATOR: Next question.

Q: Marc, thanks. A separate question here. So, I’m curious of your sort of take on the– your recommendation, I suppose, on to what degree we will see sheltering in place or something approaching complete lockdown. I noticed Neil Ferguson in Britain kind of did a quick change from sort of, let’s sort of let this play out in the population and acquire some herd immunity, to, let’s go into as close to complete lockdown as we can very quickly.

Do you think that’s like– what do you recommend? I guess I’m curious, what do you think is likely to happen here as far as lockdown and– yeah, that’s it.

MARC LIPSITCH: Yeah. I think from a disease control perspective that’s absolutely right. And I’m not sure if I’d share that description of Neil Ferguson’s evolution of position. But I’ll leave that aside.

But I think from a disease control perspective a near complete lockdown is appropriate at least until we have time to assess the situation and build up our capacity to test and to take care of patients with PPE and that sort of thing. I have refrained from making a blanket recommendation, because I know what I’m talking about in the epidemiology realm. I’m not an expert in the secondary and tertiary effects of locking cities down and locking places down. And I think it’s a decision that should take into account those secondary and tertiary effects on mental health, on well-being, and on economic activity.

I am sort of struck– so I’m going to punt that one for the moment and say, I think there’s a very compelling case from a disease control perspective. And those places that do choose a lockdown or a near-lockdown need to really be careful about trying to maintain the health and safety of their population on other fronts at the same time.

For example, I’ve been told that in Paris you can’t go out for a walk unless you have a dog that you’re walking. That may be anecdotal or apocryphal. But that is not a sensible disease control approach. Being outside is actually probably better than being inside. And until there’s evidence that people are mobbing the parks and having large gatherings in the parks, those kinds of overreactions I think will cause a lot of secondary problems.

So, I think there should be a real effort to share ideas about how you do a lockdown if you’re going to do it, how you keep the grocery stores and pharmacies running, how you keep plumbers, and electricians, and other people servicing what needs to be done, keep the infrastructure going. We are not used to this. And I don’t feel it’s my position at this stage to say, we need to do something that could be so destructive if done badly. I will just say that, from a disease control perspective, I think we are really heading for big problems in terms of health care overload. And from that perspective, it’s a good idea.

Q: Yeah. I guess the follow up to that, Marc, is, do you still feel your overall prediction about a pretty good swath of America getting infected by this over the next 12 to 18 months– you said 40% to 70%. Do you still sense it’ll be in line with that?

MARC LIPSITCH: Yeah, I have reduced that from 40% to 70% down to 20% to 60% about two weeks ago. But doesn’t seem to have stuck as much as the first. And that’s obviously a very wide range.

I’ve gotten a lot of questions about that. My current thinking is as follows. Clearly, China has shown that you can, in the short term, even in a place as critical as Wuhan, bring cases down to nearly zero, detected cases to zero, with several weeks– about two months, I guess, from January 23 to the present– almost two months of lockdowns. If somehow that could be maintained, then over a long period, like over 12 to 18 months, then those numbers would not come to pass. Because the disease would stay under control, very likely.

The problem is that I don’t understand how anybody, but especially a democratic society, is going to maintain that level of control for that period of time. So, the alternative is to have cycles of cracking down and letting up. And that would slowly accumulate. Cases would accumulate immunity in the population and would approach those kinds of levels.

If there’s some other solution, like if a vaccine magically appeared, which I don’t think is going to happen in the next year, that would change the game. If a treatment magically appeared– or not magically appeared, but if a very effective treatment appeared that would allow people to get infected and get immune, but not have severe complications, then we would probably decide to let the infection spread more and just treat those who get infected.

So, the premise of that idea is that long term control of the epidemic either requires vaccination of about half the population or more to acquire immunity, or some other way of acquiring immunity. And the only way we have now is natural infection. So, it’s not about sort of the short term. It’s about how you stop an epidemic from spreading in a more permanent way. And that’s where those numbers come from. It’s based on models of how infectious diseases spread under mild control, or no control, or intermittent control.

Q: Thank you, Marc. One quick follow up, if you don’t mind. I don’t know if you saw the Tomas Pueyo analysis. But do you agree that if our health care system is overwhelmed then the mortality rate of perhaps 1% would be higher than that?

MARC LIPSITCH: I haven’t seen that analysis. But I think that’s true. I mean, the health care system is helping. And treatment, obviously, does help. It is also true that among those in intensive care, a significant fraction do go on to die. So, it’s not perfect, even when it’s functioning at a good level.

But nonetheless, clearly, treatment is beneficial.

Q: Great, thanks so much.

OPERATOR: And at this time, we have no further questions in queue.


This concludes the Friday, March 20 press conference.