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 12 p.m. Eastern Time on Friday, October 16th.
MODERATOR: Dr. Mina, do you have any opening remarks?
MICHAEL MINA: No.
MODERATOR: All right. Let’s get to those questions. First question.
Q: Hey, Michael, how are you? Look, here’s a comment from the strategist at Credit Suisse. While concerns about a widespread second lockdown have increased, improvements in treatments have led to a large decoupling between deaths and infections compared with the initial stages. We think any second lockdown is likely to be short, confined to specific localities or high-risk groups. Agree, disagree and show your work.
MICHAEL MINA: Yeah, I feel like I would say that there’s a couple of things, certainly mortality has reduced, but we shouldn’t mistake it for being greatly reduced within the individual age strata. We have gotten a little bit better at ensuring that we’re not allowing the virus to enter into nursing homes, which so far have accounted for almost 40 percent or bit more of all of the deaths. And so if we can really keep the virus out of there, it does. We do have a declining mortality rate. But if we were to allow this virus to spread much more widely, I think we would really have a massive amount of difficulty to prevent spread into these very vulnerable locations. So I do agree that probably we will see more targeted lockdowns instead of just massive lockdowns. And that’s maybe for no other reason, except that the public might not be willing to undergo larger lockdowns, in particular, if we don’t have pretty immense economics sort of bolstering during that period of time through congressional appropriations or something like that. But we don’t really know yet. I would say there’s a very good chance that in the next couple of months we will just see cases continue to climb and climb and maybe it will happen sufficiently gradually that people will be uninterested in taking it seriously or something like that. And I think that the problem with COVID fatigue is real and we’re just going to become complacent with hundreds of thousands of additional people dying. And that’s more or less what we’re seeing happening. You know, maybe the change in leadership will affect that significantly. But I do think that we will see both of these things happen. We’ll see more targeted lockdowns, maybe less lockdowns overall. But in association with that, we’ll just see a greater willingness to let this virus continue to spread and ultimately could end up killing more people that have already died. So it’s a complex issue that really does take into account enormous complexities around the economy, people’s well-being outside of COVID, and then, of course, the damage due to COVID itself.
Q: Great. Thank you.
MODERATOR: OK, great. Next question.
Q: Hello, professor, so I have two questions. The first is, because recently many countries around the world have seen a large rebound of the epidemic. So does it mean that the second wave is coming? In the next stage, what new approach or strategy should America or the world adopt?
MICHAEL MINA: Yes, the short answer is yes to the first question. I think what has happened is, the population at large has mistaken continued transmission for there not being a second wave. And just this idea that we’re just on the first wave, which is sort of true. But there’s a real danger in that thinking, and that danger would be that it’s not going to get worse, that we’re already at a peak of a wave and that wave won’t grow. But I think firmly the other direction that we will see continued increases in cases that could, especially in the US, where we’re actually almost already back at the peak that we saw nationally in the summer, in the early summer, when cases started to really spread through the South and the Midwest and California, we’re almost back at that same same level. And I don’t believe that we’re anywhere close to that rise abating. And so call it what we will, whether it’s a second wave, a third wave, an increase of the first wave. There’s no good definition for what defines a wave. But certainly, I think what we can say is that cases will go up. We’re seeing them go up and in Europe and in many of those places, cases truly did get down to fairly low numbers first, and now they’re skyrocketing again in many places. We’re seeing it in the U.S. So we’re seeing it in other places as well. And so we’re pretty much there. And I think we have a long a long winter ahead.
And the second question, what could we put in place, I’ve talked a lot about these in the past, I think number one is we have to stop being complacent as we wait for a vaccine. Our policymakers have put so much energy into talking about a vaccine and pinning our whole response essentially on the vaccine. It’s discussed daily. It’s been discussed daily for months. And it will continue being discussed daily. And it’s important, but we are not in anywhere close to a position to pin our mitigation strategies on a vaccine that doesn’t yet exist. So that’s where masks and distancing, and then other normal pieces of hygiene are the routine things we could be doing every day. And I’ve discussed a lot in the past that there is a strict strategy that is as ambitious, if you will, as a vaccine. But unlike a vaccine, we could be building it now. We could be deploying it now. And this is very frequent testing to the masses, flood communities that are being hard hit with this virus, flood them with frequent testing so that as many people as possible at any given time know their status and can isolate and quarantine their neighbors without finding out that they were positive two weeks ago or without going their whole course of infection without knowing that they were positive at all. So we have a tool that at least could be used in concert with social distancing and with masks, and that’s very frequent rapid tests that can be built into many, many millions per day and distribute it widely. They would be flexible in a way that PCR is not. PCR, you’d have to have a lot of logistics surrounding any programs that are focused on PCR testing in a laboratory. These would be tests that could be used at home, at school, at work before you get on a plane, whatever it might be. Any opportunity we can take to identify positive people and ask them to isolate is a transmission chain severed and potentially tens, hundreds or thousands of cases that do not end up occurring down that transmission chain as a result of finding that one individual one at a time. And so I think we can bank on herd effects, which are very similar to vaccine herd immunity that could be derived from frequent testing as a true robust program. That is ambitious for sure. But it’s no more ambitious than the speed that we’ve put our focus on vaccines, but there’s not a time limit. With vaccines, we have to wait until the studies and safety studies get done. There are no safety signals here. We just have to figure out how to how to build them, deploy them and message them appropriately.
Q: Thank you. Actually, I have one more question about the Great Barrington Declaration. I know they’re writing many discussions around this declaration. And my key question is that the declaration stated that the risk raised by COVID-19 is almost a thousand times higher for the elderly than the young. So for children, the risk is even lower than in fact, flu. But there are also increasing evidence of young people suffering longer term health consequences, consequences from moderate or even mild COVID-19 infections. So I want to know that because I already read many articles and I want to know that if there is a consensus among scientists about the different risk among different age groups, you’re already making consensus on this issue. And we should find age-based chatter around this topic.
MICHAEL MINA: Yeah. I think there’s a lot in that question. Certainly, I’ve been saying since February that we should be taking advantage of what we know about the age associated risks of this virus. There is some likelihood that younger people will have long term consequences, but they’re probably at the end of the day, going to be rare. But that shouldn’t be misconstrued for this idea of just going for herd immunity. I think it’s a terrible idea what was described in the Great Barrington Declaration. But I also think that completely discounting it is a terrible idea as well. I think we have to be really serious about understanding what the economic consequences and the social consequences to society of this virus and our reaction to the virus are. And how do we weigh them appropriately against our individual strategies to contain the virus? It’s not easy and there’s no clear answer of what’s right and what’s wrong in this case. And in some ways, it might differ by country and might different by social structure and by how individual countries want to respond and what their risk tolerance is, what their age stratifications are in their communities. So there’s no single answer. I would say there is consensus that this virus is more lethal to the elderly than young. And in young it is less lethal than flu. But I’m really concerned that if we did just go to herd immunity pathway, we really couldn’t necessarily keep the elderly safe. If this virus was everywhere, we just wouldn’t be able to keep it out of the nursing home. So, sure, it might pull off the Band-Aid quickly. But at great devastation to populations at risk. And I don’t think many populations are willing to take that on as a consequence of controlling the virus or not controlling the virus itself. So it’s a difficult question. I don’t think I think that the media has kind of pinned it as two sides and scientists have painted as two sides in many ways. And, you know, I think that it needs to be dealt with in a more nuanced way. And I’ve tried to do that myself, working with economists like Jim Stock at Harvard. We’ve published a couple of papers on this particular issue of saying that there has to be a middle ground. We have to consider the economic devastation. At the same time as we consider the containment of the virus. And how do we do that in a comprehensive way? The only way that I know to do that is to start bringing the epidemiologists together with the economists, the social behavioral scientists, and have us all work together. And largely, we haven’t seen that happen, which has been a real disappointment. There are some others, but in general, the discussions have remained fairly disparate.
Q: Thank you.
MODERATOR: Next question.
Q: Hi, Dr. Mina. Thanks again for doing this. We’ve heard a lot of news about these massive quantities of the Binax Now that are going out to Texas, Arizona and the schools across the country. But in your opinion, do we have enough to make a difference with cases starting to rise nationally? Are you optimistic that we have enough to make a difference?
MICHAEL MINA: No, I don’t think we have nearly enough. One hundred and fifty million sounds like a lot, but there’s a few parts to this. It’s not a lot. First, it sounds like a lot to the average person on the street who says one hundred and fifty million is a big number. But they’re going to be built at 30 million a month so far. So that’s enough for about one person to have one test per year. It’s just not a lot in this country. If we were a small country of a few thousand people, sure would be a lot. But we’re not. We’re 330 million. And so one hundred and fifty million is what they purchased. Those hundred and fifty million aren’t going to be deployed tomorrow. And if they were deployed tomorrow, it would use up all of them so they couldn’t be deployed next week. So what we really need is a very consistent supply of 10 or 20 or 30 million a day, and we don’t have that. And so, you know, when the vice president got up and announced in the Rose Garden a number of weeks ago for a second time, it was already announced in August. So I don’t know why they had a whole second announcement except to distract from taxes and such. But when they got up and announced the use of them, they said things like, all teachers will be safe to go to school. We’ll be able to really ensure the safety of teachers in K-12 schools and things like that. And that’s just total fabrication. It’s stretching the truth beyond where it really should be stretched. And it’s just hyperbole. They’re saying words that don’t really match facts and, in this case,, to actually get enough tests to keep K-12 schools safe and to keep teachers safe. If testing is one of the major strategies, it would have to be frequent. You can’t just test somebody once and say that they’re negative, not with a virus like this, it doesn’t work that way. And so in general, I say that it’s a good start.
But I’m extraordinarily frustrated that the federal government is just continuing this this this position where their only strategy is to purchase what becomes available through industry and corporations in the private market. This is effectively a war against a virus that’s damaging our society in a way that very few things in the past century have done. And it’s that economic hardship that’s closing in businesses. It’s deaths. It’s everything. Why are we not treating this as we would a Third World War? You know, this should be treated with that same sense of urgency. And part of that would be just building the damn tests ourselves. We know how to build these things. The same way that we build missiles and bombs. We should just be taking the onus on us as a country, as a federal government, and not just paying a company and saying whatever you can make us, great, that’s enough. We should be taking the initiative to build them and ensuring that everyone can go to school safely and then we can actually control this virus. And we just haven’t really seen any movement in that direction.
MODERATOR: Do you have a follow up?
Q: Yeah, I do. Just quickly, were you surprised that what happened at the White House after the Rose Garden event? I mean, we’ve been following your work for a while and you’ve been saying testing is not a replacement for masks and other infection control measures. But were you surprised that it stretched so far?
MICHAEL MINA: No, not really. This is how super spreading events work. You get one person who’s really aerosolizing virus, and this is the natural consequence. What I do think is really important to recognize is that the frequent testing was used as one of the only barriers of this virus up until the Rose Garden event. And the White House was actually able to stave off outbreaks like this earlier on. And I think that a lot of people misconstrued the Rose Garden event as evidence for frequent testing not being useful. But frequent testing as a strategy to combat a virus is all about reducing odds of transmission. It’s not about a bullet proof way to prevent all transmission. And that’s why it has to be couched in all of the other measures we should be taking like masks and social distancing. But it can be along with those. It can provide sort of the one of the strongest barriers, not completely impenetrable, obviously, as we saw, but a very strong barrier to reduce the risk of these events happening. And that’s really what we saw from March through October. But I look at the White House as just rolling a roulette wheel every day and eventually it was going to land on the wrong the wrong box. And that’s exactly what happened. But they could have really improved their odds if they were also wearing masks. You know, maybe that same person, if they had been wearing masks, we wouldn’t have seen a super spreading event like this happen despite that one person squeaking through and not getting caught by the test for whatever reason. We don’t even know enough information about whether the person was even tested, you know, was the super spreader the president? We don’t know. And we don’t know if that president or that person was being tested beforehand. So we can’t really say too much about that. But the short answer to that question is, no, this isn’t surprising at all. If you continue to have community gatherings like that, you refuse to wear masks, you can cut your risk through frequent testing, but you cannot prevent these types of things in absolute.
MODERATOR: Are you all set?
Q: Yes. Thanks.
MODERATOR: Next question.
Q: Hi. This week, the CDC issued some, I think they call them, interim guidelines for testing in schools. And I wonder if you had a chance to look at those and what you thought of them. I was having a hard time figuring out what the takeaway was. I mean, they did mention the possibility of frequent testing, but a lot of the guidance was sort of couched in terms, if you could do this, you might do this. So I’m curious what your take was.
MICHAEL MINA: To be honest, I haven’t seen them yet, which to me suggests that they probably weren’t particularly clear or useful. Otherwise they probably would have landed on my desk at this point. If somebody doesn’t mind sending them to me, I’d be happy to comment on them via email for later. But my guess, knowing the types of recommendations that have come out thus far is there’s not a strong stance because there’s no real stance that the CDC can take at this moment, because they can’t provide or guarantee testing with any sort of regularity to any school across the country or really even create programs that include it, because different schools will have access to different amounts of testing and will have different resources. And so far, federal government has not done much to improve that fact to try to help schools gain equitable access to testing in a way that they can control. And it’s one of the things that I find so appealing about if the federal government would take it upon themselves to build up these rapid frequent tests, which, of course, new evidence continues to come out. There’s been a lot of good data now showing just how powerful and effective and sensitive and even specific that these new rapid tests can be. I think one of the real appeals there is that if the government did build them in sufficient numbers for K through 12 schools, for example, then they could actually come out of policy that really serves as more than just some handwaving to help schools understand what to do. But at this point, they don’t have control over any testing infrastructure. They don’t really have control over any programs that they could suggest because they just can’t guarantee that any one school would actually have access to what it is that they are suggesting doing. And this is a failure of our country at the moment. It’s not necessarily a failure of the CDC. They don’t have a lot of options here, I would say. And so I think they’re in a very tough place to create any useful recommendations because they don’t have the tools that they can guarantee to anyone. If they were to make recommendations that are helpful.
MODERATOR: Do you have a follow up?
Q: That was helpful. Broadly, some people seem to be ready to draw conclusions from the first month or six weeks of school in some places, that opening schools turns out to be safe because, you know, we haven’t read about too many wild outbreaks in schools. Are you drawing any conclusions at this point? And if so, what are you basing that on? And if not, what are you waiting for?
MICHAEL MINA: No, I haven’t drawn any conclusions about this particular issue, I do think that schools are areas that can transmit widely. We’ve seen evidence of that, for example, the summer camp in Georgia showed very well that you put the wrong person at the wrong time into the wrong place. And you can end up with serious super spreading events that then can transmit to the wider community. My hope is that schools can figure out how to remain safe. But I do think that as we start to see community spread, increase in cities and in communities, that just the chances of new cases coming into schools and leading to transmission will inevitably increase. And we’ll probably see them. Will it be so frequent and devastating that schools have to shut down? I hope not. But I do think that all evidence suggests at this point that children can transmit and can acquire the virus they can get. They can have viral loads that rival any adult. And in the distributions are quite similar from what we’ve seen in terms of the violence. And so I don’t see any reason why schools couldn’t be areas of transmission. I do think we should be trying to do everything we can to stop cases from entering into them. But I wouldn’t want to say that. I don’t think I’ll end up saying at any point that schools are immune to transmission within their walls.
MODERATOR: Are you all set?
Q: I’m all set. Thank you.
MODERATOR: Next question.
Q: Hi, thanks for taking my question. On August 29th, comments of yours were published in The New York Times regarding cycle thresholds for PCR tests for COVID-19. The gist here is that a CT should be published alongside positive results to give a better idea of viral load. Locally, these comments have been taken out of the original context by conspiracy minded people proclaiming that all PCR tests are invalid because they detect dead virus by running these sorts of cycles. I don’t think this is the intent of your original comments. Could you please clarify your comments in the Times?
MICHAEL MINA: Sure. I would say that article led to massive confusion. At the same time, though, it did get the message out and it has really changed the conversation at the federal level and among the Department of Health and really globally about whether or not we should be considering cycle thresholds. And so I would say while there was a lot of confusion and conspiracies developed from that article and from the research that we’ve published that led to that article, in large part, it has been effective. So I’ll say that it’s been effective in making people think, including the FDA and CDC, about cycle thresholds and their importance. But the conspiracy theories have largely said things like PCR, because it’s picking up dead virus, it’s wrong. That’s not true. And the they’re not false positives that are getting picked up. And I think a lot of people, because the headlines that picked up that story said 90 percent might have even said it in the story, but I think it was a little more context, that 90 percent of the PCR results from surveillance are false positives. That’s not true. They’re true positives. They’re really picking up the viral RNA. What it means is that we just need to be doing more frequent testing if we want a chance at catching those same people early. The important thing to recognize is that those people who are getting a test, who are coming up positive, but at a very low RNA load or very high cycle threshold, they likely had very high viral loads earlier. They likely were at a transmissible state, maybe a week or two or three earlier. So the problem isn’t so much that the test is incorrect, it’s that we’re not interpreting it correctly, if we go and assume that every positive means that somebody is currently at the beginning of their infection and transmissible. And cycle thresholds can really help us to better interpret the value. This is back in March.
We published published a paper in March or April, that said, you know, to appropriately use the SARS-CoV-2 test, use the cycle threshold to guide you effectively. That’s the message. And we can use it both in clinical space and in public health to know, does somebody need a full 10 days of isolation? Do all of their contacts from yesterday need to be quarantined. And in large part, no, but we have strategies and we know that there is very low viral load. And we see that three days in a row. And that means that we can use that information to say, hey, maybe this person doesn’t need to be isolated. They have no known exposure, but they are positive. But with such a low amount, that means that they’re probably over their infection weeks ago. We can make those assessments from the cycle threshold. But this should not be misconstrued as saying that the PCR tests are wrong. I would say that we’re just not using them for all of the power that they have. We’re throwing away one of the only pieces of data that we really have. For a lot of people who get sound on surveillance, who don’t have a known exposure, who don’t have symptoms, we have very little to go on besides a plus sign. And the cycle threshold is what we have to go on there to better interpret their plus. And so I think that article was taken out of context or was misconstrued. It shouldn’t be that we should not be going up to a cycle threshold of 40. That’s a wrong evaluation. We should always be using the test to its maximum to detect as many people as we can. But then once we know that somebody is positive, we should then take the next step and interpret it appropriately and say, you know, is there more that we can learn from this? The easiest example is not at the low ends of viral loads, but if somebody is found to have a very high viral load, those would be people that if I was a contact tracing unit or a public health department, I’d want to take that person who’s detected today with a very high viral load and put them immediately to the top of my list, because they’re the best bang for the buck, if you will. The most efficient use of resources to stop new transmission is if you know that somebody is currently highly infectious, potentially highly infectious. And you know that, then you have a very limited number of people that you need to necessarily contact trace to ensure that they weren’t infected. So we’re just not interpreting the CT values at the moment. And that’s really what that was about.
Q: I do have a follow up. In a clinical or diagnostic setting, there can be considerable pre analytical variance in the viral load of any given sample due do things like sample error or sample storage conditions or the phase or severity of infection that patient might present with. How do variables like this influence the discussion of the use of CT value as part of the testing regime?
MICHAEL MINA: Yeah. So it has to be taken in context. You know, I’ve been astounded at the number of people who bring up this point and see this as a deal breaker. Nobody ever said that anything in medicine is perfect. If we could completely rely on every data point that we got in isolation, then we wouldn’t need doctors. Google would do everything. We would have no physicians anymore. So these data just need to be assessed in the context of each individual, the same way that we normally, we get a positive person. I like to give the example of love listening to a heartbeat. Doctors all listen to heartbeats. The sensitivity and the variance that’s associated with listening to somebodies’ heart is vast. You know, maybe you catch somebody and maybe you don’t, with a murmur or something. But it doesn’t mean we shouldn’t be incorporating that information into our clinical context. The same thing goes here. Like I said, the easiest piece, if for no other reason, we should be identifying those with very high viral loads because a bad swab and a poorly performed test isn’t going to make the viral load erroneously high. Only a high viral load can make the test come up with a really high viral load. Unless the sample makes up for something. That’s at least one direction that we can go, and we can say, look, that’s an important piece of information. And then on the other sides, we could say if you have a very low viral load, then what’s our next step? Just like physicians normally do, we get this result and what do we do next? Do we believe it outright? Probably not. It’s not a good idea to just completely take these values at face value because there is a lot of pre analytical variations in the swab, it’s in transport, all these things. But, for example, we can improve that. Our understanding, if we have a positive today with a low viral load and we test them tomorrow again and they have a very high viral load, then you know that you struck gold and you found somebody very early in their course of infection and you have removed them from infecting others through isolation. But if they’re very low, low viral load, and then it happens again tomorrow. And if you want to be extremely cautious, do it again a third day. And if they remain at a thirty-eight, thirty-eight, thirty-eight, then that is extraordinarily strong evidence that this person is no longer infectious and most likely, in just pure mathematics, bear this out, the most likely scenario for that individual is that they are over the course of infection. So I just think if we continue to try to make everything black and white and say this can’t work because of this. But you know that the natural human predilection is to poke holes in ideas, and that’s needed, but in the midst of a pandemic, I think we should be trying to think more constructively. And so this is a good example of where we have to say, OK, there is variance. How do we overcome that? How do we still use this data to our advantage to know how best to contain this virus? And I think that we have a lot of options.
Q: I do take this point. But the thing that I guess is the natural follow up to that is if you’re a contact tracer, how much information are you actually getting about each kind of PCR kit for COVID, when you’re receiving like a positive test to investigate? I think that if you’re going to use a CT value, you also have to include a fair degree of technical information and training for people, probably in order to make that value meaningful, because PCR kits aren’t necessarily cross comparable. They don’t always have the same CT threshold.
MICHAEL MINA: So this is completely normal. I think people have lost their minds here. This is why every single lab in the country makes their own distributions and makes their own normal values. We know the metrics. The idea that an instrument isn’t comparable across the instruments is totally normal. We deal with that every single day in laboratory medicine. Even just basic chemistry, cholesterol, you know, anything along those lines, we have normal values that are specific to labs and specific to instruments. This is how FDA approval’s work. All of this is totally normal for people who are used to laboratory medicine. So I’ve been really surprised that this has continued to be an issue. We know how to deal with this. We publish normal values. If there’s any given moment in time, there’s only so many labs that are necessarily sending their results to to a Department of Health for contact traces, for example. This can be part of the same way that a physician, anytime we get any result in a patient’s chart, even if it’s just something like a cholesterol, it comes with the normal value. Does this fit on this machine, in this lab? Is this an outlier? And we could do very similar things like that with these values. But the important part, too, is I think the differential between labs has really been totally overblown, because what I’m talking about in terms of how these values will be used, is on the numerous orders of magnitude separating a low RNA load from a high. And so, I think that that’s really important to take into consideration between instruments or labs, maybe you have a difference of four or five cycle thresholds. So that’s a 10 or 20- or 50-fold difference. But that’s all within the same region of a viral loads in terms of what actions you would take next. The differences that I’m talking about that would really cause contact tracers to go in two different directions, are getting a viral load of 10 billion versus getting a viral load of 10. And these quantitative PCR instruments are not going to be that far different that we can’t distinguish those two. And I think that that’s really very, very important to understand how many orders of magnitude difference we’re talking about.
Q: Right. Thank you very much. The background for these questions, I used to test primers. So, you know, I’m trying to get this out of you. Thank you for your time.
MODERATOR: Next question.
Q: Yes. Thanks. Two questions about using the percent of positive virus tests as a marker for infection incidence or even prevalence. You know, people have been using the percentage of positive tests as an indicator. And I’m wondering how widely is that done? And is that a good or a bad thing?
MICHAEL MINA: Well, we don’t have many metrics. I personally look at that and think very little of those values. I mean, I’m astounded that policies are made surrounding the fraction positive. The fraction positive is wholly dependent on what you’re sampling procedures are. You know, sometimes we see twenty percent positive and sometimes we see point two percent positive. If you’re testing only symptomatic people, then maybe five percent positive is low. If you’re testing random people who walk by you on the street, then point five percent is very high. And so I think it’s just so incredibly hard to know how to use these data. And we’ve been working on a different area. We just put out a preprint yesterday or three days ago, maybe, that shows a different strategy for how to understand epidemic trajectory and whether cases are increasing or decreasing. That is not based on total case counts and it’s not based on a fraction positive at all. It’s actually based on the CT values and it’s based on understanding how we can look at the distribution of cycle thresholds across positive samples on any given day. And it can actually give us a growth rate of how quickly the virus is going up or going down in the population. And so it’s another area where we can really use cycle thresholds to help inform on these issues because of the issues that you bring up, which are now that fraction positive just isn’t, you know, it’s such a furphy term in this way because we just don’t know how to interpret it and any public health level for the most part.
Q: Have you seen any problems arise because of overuse of that statistic?
MICHAEL MINA: I don’t think I can pin any of specific problems on it. I do think, for example, in New York City they’ve said that they’re going to put things on hold if the positive rate goes above three percent, for example. But I don’t know if that’s actually useful. I don’t know how they’re necessarily finding who they’re testing when they say the three percent of the community is it truly random people, is it not? Because in my opinion, three percent positive in the community is extremely high. If it’s truly random, we would be on herd immunity in 20 days or something. And so obviously, they’re not talking about random sampling. So I think that I haven’t seen any specific issues where I’d point to and say, look what happened, because they’re using this value. But I wouldn’t say that people are placing a lot of emphasis on when it’s really a value that’s extraordinarily difficult to make any sense of.
Q: Thanks very much.
MODERATOR: Next question.
Q: Hello. Sorry to bring you back to vaccines, but that’s where I’m going. So today the news broke that Pfizer’s CEO says assuming positive data, the company will apply for an EUA in November. One researcher I spoke to already says it’s really unlikely that Pfizer will be able to reach that goal and they’re just being optimistic. What’s your take?
MICHAEL MINA: Yeah, I agree with that. Theoretically, could they? Yes. And first I wanted to say don’t apologize for talking about vaccines. My normal life is vaccines and my clinical life is testing, my normal research life is vaccines. So I do appreciate talking about vaccines. I think that it’s theoretically possible, but I completely agree that, you know, the CEOs and the industry companies, if you follow kind of how they normally talk and this is diagnostic companies, vaccine companies, they always shoot for the stars in terms of timelines, especially because it gets people jobs. It gets people energized. It gets investors energized. Normally there’s not so much critique about it because it’s not normally a global importance in the way that these vaccines are. But this is really common language and common sort of ways to talk about it like we see all the time. For example, with the diagnostic companies, who I work with a lot in the clinical realm, they’ll often be a whole year off in terms of when they put a time line up and they say, oh, we’re going to have this new test, it’s going to be FDA approved in six months, it’s usually not for another 18 months, for example. And so I think this is how our market works. Maybe Michael Brush has something more to say on it. But I think that this is really common. And could it happen? Sure. Is it likely to happen? Probably not. And so they’re usually careful to use their words carefully, which is we might be able to apply for EUA by August or, you know, our goal is to but they’re not painting themselves into a corner with it.
Q: OK. Thank you. And one follow up I have, in his role as the designated senior representative, Dr. Fauci is hands on in overseeing the companies that are participating in Operation Warp Speed, and thus he has a role in deciding whether the vaccines are ready to seek approval from the FDA. But when it comes to Pfizer, since they didn’t apply for government funding, it doesn’t apply to them. So can you talk to me or expand upon that? So how does that change their approval process? And who’s overseeing them?
MICHAEL MINA: Well, they’re overseeing themselves. They have safety monitoring boards. And this is normally how vaccines get approved. Usually it’s on the companies and more or less, the company is alone to do the testing, do the due diligence, ensure that they have what are considered to be external advisers and external safety monitoring boards that are doing their due diligence to ensure safety and efficacy. People are evaluating efficacy, appropriately on safety and capability. So that’s I would say, that it doesn’t change a lot in the context here. You know, if you’re part of project warp speed and you’ve received funding, there might be some internal mechanisms that will help once things get through to the FDA that would have been considered sort of an ability to kind of get to the front of the line if you will, for evaluation. Even though there are a lot of vaccines being produced at the moment, there’s only going to be a handful at most that go up for EUA authorization at any given time. I think that any of the first vaccines to go up for authorization are probably going to get immediately put to the EUA line, if you will. So I don’t think in effect, it’s going to have much impact on the timeline. And Pfizer has got enough money that they didn’t really need taxpayer dollars at this point.
MODERATOR: Next question.
Q: I can assure you there’s never any hype in the stock market. Another data question. So we’re all talking about the death rate now as if it’s a great thing. But I wonder if you can tell us, is that improving because we’re really getting better at treatment and at protecting people? Or is it more just a measurement issue in the sense that so right now, in the seven-day average, about 60,000 cases produce a thousand deaths. And in March, it was 30 K to basically 2.25 K. So is that more testing or are we really getting better at treatment and protecting people? What’s your take on that?
MICHAEL MINA: It’s a little bit of both. We have gotten better at treatment. There’s no doubt about that. Even really simple things like how to position patients in the ICU when they’re on the bed, have really made an impact. With each of those, none have been the silver bullets. But each of them is improving and they collectively are improving our ability to know how to take care of the patients and keep them alive. But I would say that’s not at all the only thing that’s lowering the perceived death rate. And certainly, there are biases, or I should say we’re becoming less and less biased, if you will, in terms of how we’re calculating precent deaths of cases. And that’s because we have improved to a certain extent, access to testing. More people are getting tested who are not necessarily symptomatic. So all these things are increasing our ability to determine these rates. I think one of the best metrics that we can use to understand this question is among hospitalized patients. Are we getting better at keeping people alive? And that’s where I think if we start to separate out these two items, or at least control for them in an epidemiological mathematical models, we can start to infer better what proportion of variance or what proportion of the overall reduction in mortality is truly a reflection of improved outcomes in the hospitalized patients. And what fraction is due to sampling biases maybe becoming a little bit more neutralized.
And then the other, which can be accounted for and can be considered a sample bias, is that we are seeing different demographics getting infected now, particularly as younger people have taken a more prominent role in the numbers of infections that are happening, or at least the numbers of infections we know that are happening because we’re testing younger and younger people. Then that also is causing the mortality to shift a bit. And I think in general, what we’re seeing is we’ve all focused so heavily on one number, infection fatality rate of one percent, point five percent. You know, initially it was three percent. And we could have looked way back then and recognized, hey, it’s not an average. That’s probably not right. We’re not capturing nearly all the cases that are actually occurring when we use serology. That gives us a much better window into how many people have actually been infected versus died. But I think that we should never, especially with a virus like this, we should never be just talking about a single mortality rate as though an average makes any sense. You know, I’ve been surprised to see how many biostatisticians and epidemiologists and physicians have been willing to talk about a single mortality rate, when it’s sort of like the first thing you learn in bio stats 101 or any statistics courses if you have a highly skewed distribution. And the mean doesn’t necessarily reflect the overall distribution. And you have to take that into consideration. And we’ve largely kind of lost that. I think in general; it’s got a little bit better. We hear more and more people talking about low versus high. But we still often hear one number for the infection fatality rate, and that’s probably not smart.
Q: So what is the death per hospitalization rate? I like death per capita. I think that’s a good way to adjust it. And also, in terms of the skew in the distribution, what were you referring to just there? Like, what does that mean for hospitalization and then the skew that you have to account for? Sorry.
MICHAEL MINA: Yes. I actually don’t know the death per hospitalization at this moment. But in terms of the skew that that I’m referring to, is the skew in who is actually dying, how many people, if we’re really trying to understand a balanced infection fatality rate, for example, we would to apply the appropriate weights. Then we see this sort of exponential increase in deaths across ages. So it’s a very, very skewed distribution in terms of what ages are dying. And that age distribution does not reflect the age distribution of the country of people. Yes, exactly. And so we should be taking a more discerning approach to not just be calculating cases known and dividing deaths overall. We should really be doing either stratifying the death rates or always talking about it in an age stratified manner. We’re really trying to understand, though, the true population distribution of infections and appropriately waiting deaths based on that. And we generally haven’t done that.
Q: Yeah, that’s a really interesting point. Thank you very much.
MODERATOR: Okay, Dr. Mina. I think we’ve hit almost one o’clock. Any final words you like to share with us before we go?
MICHAEL MINA: Well, I guess for those who are still on. I’ll just put a plug in for that paper that I mentioned that Nicole put the pre-print under. I do think that it’s still a pre-print and we’re going to come out with a published version in the near future, so this doesn’t have all the analysis. But I actually think that this should, if thinking about the CT values or new ways to understand the trajectory of an epidemic, this is really a completely novel new method that has never really been done before. But I think it has potential to really change the way we’ve generally looked at three metrics throughout this whole pandemic case counts on a daily basis and frequent positives among those being tested. This just gives a whole new resolution to understand for public health people and for the population at large to just be able to monitor outbreaks and epidemics. And I expect that we should see public health departments around the world probably starting to use these methods like this that we come up with. And so I would encourage people to follow this idea and see if it starts to be used more frequently.
This concludes the October 16th press conference.