The following is a rough transcript which has not been revised by The Jim Rutt Show or Samuel Scarpino. Please check with us before using any quotations from this transcript. Thank you.
Jim: Today’s guest is Sam Scarpino. Sam is a director of AI and Life Sciences at the Institute for Experiential AI at Northeastern University. He’s a professor there of the practice in Health and Computer Sciences. In 2020, Scarpino co-founded a data sciences initiative called Global Health, which was backed by Google and Rockefeller, and he just told me it’s now backed in addition by the Welcome Trust, which is great. He’s also an external professor at the Santa Fe Institute and an external faculty member at the Vermont Complex Systems Institute, and he also holds an appointment at The Roof Institute. Welcome Sam.
Sam: Thanks for having me back, Jim.
Jim: Sam is a returning guest. He’s my go-to guy on epidemiology, epidemics, pandemics, and all that stuff. For those who want to hear more about Sam, back on Currents 099, we had an episode titled, Sam Scarpino on Preparing for the Next Pandemic and Currents O47 way back in the COVID hotspot on the epidemiology of COVID-19. Both extremely interesting and to the point episodes. And as always, of course, links to those episodes and everything else we talk about within reason will be on our episode page at jimruttshow.com. Let’s get into it.
Today we’re going to talk what motivated me to reach out to Sam again, and we’re going to talk about what people call the bird flu, which if you’re following the news, there’s quite a little bit of flutter all around about it these days, but we’re not going to call it the bird flu. We’re going to refer to it by the scientific name for the current variant, which is H5N1, so that our episode can have greater value over time because I do see people come back and listen to these things even five years later sometime. And by then we may have H9N6 and it might come from parrots. Who knows? So we’ll go with the scientific name. So tell us a little bit about H5N1. What do we know about it so far?
Sam: Well, we know quite a lot compared to many other viruses. So the global health community has been preparing for influenza pandemics for decades. We’ve seen a number of them going all the way back to 1918. H5N1 has been on the radar since the late ’90s when it spilled over from birds into humans and caused a number of fatalities. One of the big concerns, and I’m sure we’ll come back to this, is that historically the fatality rate has been somewhere in the 30 to 50% range, which is typical of avian influenza that spills over into humans. One of the things that is not entirely unprecedented but somewhat new about what we’re dealing with today is the role of dairy cows, which have historically not played a major role in the transmission of really any kind of Influenza A, not just H5N1, but any of the human bird or swine flus have not really involved in ungulate or cow intermediate host.
Jim: Interesting. Yeah, the fatality rate’s a key one. In our last episode we talked about the fact that COVID-19, as disruptive as it was, was not the big one. While it ended up being quite contagious, I think the last number I found was at an R0 of maybe 3.5, which is pretty contagious. It had a case fatality rate of what? Less than 1%, something like that.
Sam: It was about 10 times worse than seasonal influenza, but still under 1%, so not anywhere in the range of what people would’ve considered the big one. And that’s actually one of the things that I think has been underappreciated about the current H5N1 situation given the number of cases we’ve seen in the US in the ’60s effectively. We know that the case fatality rate isn’t 30 to 50% because we would’ve seen a number of fatalities already, but if we put our Bayesian reasoning hats on, we could still easily have case fatality rates in the five to 20% range, which would be a hundred times or more worse than COVID-19, which itself is already 10 times worse than influenza. So it’s not like we’re out of the woods just because the case fatality rate is not in the 30 to 50% range.
Jim: Of course, we don’t know either. It’s going to mutate before it develops, the ability to jump to humans, if it does, and whether it intensifies or as more often of course becomes less virulent as things mutate, thinking about it, second law of thermodynamics and all that, we don’t really know. But the fact that I think it’s just a reminder about how as disruptive as COVID was, because its case fatality rate was quite low. Its actual jabbing right at the guts of our civilization was not what some of these future pandemics could be, and this is a quite reasonable candidate. Before we then jump in a little further, why don’t we talk very briefly about the math of epidemiology. I know the SEER model is one such model I’ve seen. Maybe talk to us a little bit about how we should think about comparing one pathogen to another and what the attributes are that make them more or less dangerous.
Sam: So when we’re thinking about a pathogen, there are a few features that we want to pay close attention to. One of the key features is how readily it transmits from person to person. The second feature is how many people are susceptible in the population and how many of those susceptible people are likely to be in close contact with individuals who are infectious. One of the things that is happening right now and has been the case with previous H5N1 outbreaks is that the virus is both not readily able to transmit from person to person, and when it spills over into humans working on a farm, those individuals tend to not be particularly well-connected to big cities, large urban areas. And so there’s a dual benefit, one on the biological side and one on the sociological side. A common way that we try to mathematize this is constructing a model that allows us to calculate terms like the basic reproductive number, which we’ve heard a lot about for COVID over the past few years.
This number is effectively a theoretical construct. It doesn’t really have meaning in the real world, but helps us make comparisons between different kinds of pathogens. So this number is the expected number of secondary infections from a single infected individual in an entirely susceptible population. And if we’re looking at this on the math side of things, we normally assume that population has an infinite number of people in it, which is one of the reasons I say it’s not particularly helpful in the real world, but it does help us gauge something of the risk level. And this are not encapsulates both features of the population of hosts. So as I mentioned, the number of contacts that people have, the susceptible population size as well as features of the pathogen, how infectious it is, and the infectiousness of course is itself a construct involving both how likely it is to actually infect someone and how infectious that person becomes before they either clear the infection or succumb and die, which is what we would think of as virulence.
So if a pathogen is really virulent like Ebola, you may not have a lot of transmission because the individual is incapacitated and often ends up dying before they have much of an opportunity to transmit. So the other reason that the are not as a little challenging to use is that it rolls up a whole bunch of different socio-biological features of a pathogen into a single number. For influenza, one of the other concerns is that it doesn’t, at least for seasonal flu, there’s not a huge effect of super spreading. And this may sound like a good thing, but it’s actually not. So for COVID, had an R-naught of probably somewhere in the three to five range, but a lot of that was propped up by these super spreading events where someone would show up at a religious service or a restaurant and infect 10 or 15 or 20 other people.
And you imagine that if you’re going to have an average number of secondary infections in the three to five range with some individuals transmitting to 10, 20, 30, that means that most individuals don’t actually transmit. And so you imagine you have all these zeros being strung together, and once we move out of the infinite population regime, it’s actually not that unlikely that it would just go extinct on its own, any one individual transmission chain. Whereas for influenza, for reasons that we don’t understand very well, it tends to march along pretty deterministically, two to three people getting infected for every single infectious person.
Jim: Well, thanks for bringing that up because I was going to point that out then Our first conversation, you talked about the curious network attributes of COVID that lots of people were zeros and lots of people were 50s or 20s, which is that difference between the basic Poisson process, simple propagation versus the super-spreader model, are those differences common? Which is the more common across the kinds of pathogens that we worry about?
Sam: There hasn’t been a robust study across pathogens really quantifying how Poisson the process of transmission is versus having a heavier tail. So we think of the Poisson distribution because it allows us to model count data, so that distribution means that we can have numbers of individuals, not proportions of individuals. And the special thing about the Poisson distribution, I know this, I’m just reminding everybody who’s listening, is that its first and second moment, its mean and variance are equal to each other. So there’s really no concept of super spreading in the Poisson distribution because the mean really accurately characterizes the entire distribution.
If we think about the normal distribution, we might not want to use that because it’s continuous instead of count data, but we need both a mean and a variance to tell us something about that distribution or something like an exponential distribution where the higher moments can be quite diverged from the first moment of that distribution. So we tend to think of most pathogens that are spilling over into humans as being dominated by super spreading. Ebola, SARS-1, certainly SARS-2 earlier and many other pathogens that are much more established like seasonal influenza as being less dominated by super spreading. But again, to my knowledge, there hasn’t been a really rigorous comparative study to back that up. That’s just anecdotal.
Jim: But we can assume with some reasonable level of confidence that H5N1 is likely to be a more Poisson or easily calculatable person-to-person similar contagion rates.
Sam: I think that’s the assumption. Now, that certainly isn’t going to hold at the very earliest stages. So as I mentioned, the virus is not spilling over into typical parts of our social network to the extent that we have anything typical of any of our social networks, but it’s spilling over into agricultural populations. One of the concerns is that we would have a super spreading event that would cause it to jump into a more urban setting where you might have much more regular transmission following something that looks like this Poisson process. So one of the concerns, and one of the reasons why so many of us have been advocating, for example, for vaccinating cows against H5N1, is that I basically think about it as a pandemic roulette wheel that every time there’s a spillover, we’re spinning that wheel and it’s really just a waiting game until it ends up in the wrong individual who ends up in the wrong place at the wrong time. You have an early super spreading event that then kicks off a much more regular, harder to control the Poisson process of transmission.
Jim: Vermont dairy farmer goes to his sister’s wedding in Brooklyn. Oops.
Sam: Well, I mean probably in Vermont, even the sister’s wedding might not be too large, although a lot of people travel to Vermont. We actually saw this during the pandemic where in Maine there were a number of really significant super spreading events because there would be a wedding that brought people in from all over the US to a small town. Small towns are also not typical in terms of their social network because people tend to be much more tight-knit. There’s only one grocery store, maybe only a couple of churches, one school. And so once something gets moving in a place like that, it’s going to spread very, very rapidly and it’s really conducive for these super spreading events. So on one hand, maybe Vermont’s not so risky because it’s rural, but on the other hand it actually might be much riskier because people are going to travel in from all over the US for a wedding and you have a much tighter knit social network, which means that you’re at higher risk of a big outbreak because you just have so many more connections between individuals.
Jim: Well, let’s go back to where we branched off, which was you said that the Poisson or even distribution of contagion is actually bad news relative to the super-spreader model. Why is that?
Sam: Well, the reason is… Well, two reasons. One is, as I mentioned earlier, if you have super spreading events, that means that you balance them out with lots of zeros in the transmission chain. And so any one individual chain is more likely than not to go extinct on its own. And we actually saw this during the early days of SARS-CoV-2, there were lots of introductions in the US that we can see in the Pathogenomic data that went nowhere. And so you get a bit of a benefit from the perspective of just the stochastic process, the chain dying out. So that’s one. The second is if there are features that we can identify that cause the super spreading events, we may be able to intervene on those features with pretty minimal disruption to day-to-day life and control the outbreak. So a really good example of that is Ebola. We know that if you have safe and respectful burials, most of the transmission happens at funerals because the infectiousness of the individual grows exponentially and continues actually to grow after they pass away because the virus is still replicating.
You can control Ebola pretty effectively with minimal disruption to society by ensuring that you have safe handling, respectful handling of bodies after an individual passes away. And again, we saw a little bit of this during the pandemic where if you say reduce the number of individuals that go to a house of worship, you can control the pandemic without any lockdowns or anything like that just by bringing down the total number of individuals that are there at any one given time. This was actually something that worked quite well in Vermont. Most of the people go about their normal day-to-day life. Maybe you have to shift when you go to church by an hour to the next service, but you can prevent those kinds of super spreading events from happening. So you get two benefits from super spreading. One is just because of the stochastic process, it’s more likely than not to go extinct. And the second is if we can identify what’s causing it and sometimes we can’t, but if we can identify what’s causing it, then we may be able to control the transmission with pretty minimal intervention.
Jim: You talked about transmission. How is H5N1 transmitted?
Sam: So this is a pretty interesting question again, in part because we’ve been so concerned about influenza pandemics, we know quite a bit about how flu transmission works. Birds actually harbor most of the genotypic and phenotypic variability for influenza. So we hear about H1N1, H3N2 infecting humans. We hear about H5N1, H7N9 coming from birds, but there are dozens and dozens of combinations of H’s and N’s that regularly transmit through bird populations. But in birds, it’s primarily a gastrointestinal disease, so it infects their gastrointestinal system. That’s part of the reason why it transmits well in enclosed bird populations. But for humans, we know there are some gastrointestinal pathogens like norovirus that can spread pretty rapidly, but for the most part, if it’s going to cause a human pandemic, it needs to be in our upper respiratory tract. And what happens is if a bird flu like H5N1 finds itself inside of the human lung, it actually can’t bind very well to our cell receptors in the upper respiratory tract.
It can bind pretty well in the lower respiratory tract. There’s a difference between the cell receptors in the upper and lower respiratory tract. That’s why it often becomes very severe in humans, but doesn’t transmit readily. One of the things that researchers have been following, and it’s been in the news lately because of some mutations that have shown up, is we actually know quite a bit about what has to change in that influenza. It primarily has to do with the H, or the Hemagglutinin that has to change such that it goes from a bird-like flu, a gastrointestinal infection, a lower respiratory tract infection to a human-like flu infecting our upper respiratory tract.
Jim: Now, interesting question, going back to something you said earlier. Birds, we’re related to birds. We go back to somewhere in the pre-lizard world, but we’re pretty damn closely related to cows. So if there is a significant outbreak in large mammals like cows, does that increase the probability of the kind of mutation that could lead to a more human easily spread version of H5N1?
Sam: So the really interesting answer is that we don’t know. For pigs, we do know that it increases the risk and the reason why is that pigs have chimeric cells that have both the bird type and the human upper respiratory tract type cell receptor on the same cell, which is why they call pigs an evolutionary laboratory for influenza host switching. It’s been very well studied. Cows, there are some published studies showing that cows can get infected by H5N1 in particular, some of these go back 20 years or so, also showing that you get reduction in milk volume, which is what actually keyed in the veterinarians almost a year ago now that there was a problem in north Texas because the milk volumes were dropping. But there have not been very many studies around influenza transmission in cows, and I’m speculating here because there have not been very many studies.
But what is known is that cows have a number of innate immunological responses to Influenza A, which H5N1 is a part of that suggest a evolutionary history with the virus that may be somewhat unique. And what I’m thinking about here, and again, this is just pure speculation, is if you’ve ever gone on safari, you see the water buffalo with the bird sitting on its back, that ungulates and birds have lived together over millennia passing pathogens back and forth to each other. And so it may not be a surprise that cows in particular and ungulates more generally have some innate immune responses to influenza that make them much less likely to get infected and transmitted.
And so the short answer is we don’t really know what to expect if it transmits from cow to cow and then into human. As you pointed out, it’s probably not good just because cows are so much more closely related to humans than any bird species is going to be. And it’s again, a big part of the reason why I’m not really in the F around and find out camp. I’m in the let’s do something about it and not have to be faced with another pandemic camp, especially if we can vaccinate the cows. It’s perfectly safe. It doesn’t involve any disruption to the human populations, and in fact, it would probably be a huge economic benefit to the dairy farmers who I’m sure are suffering right now with the reduced milk production volume and the other measures that they’ve had to put in place.
Jim: How widespread is H5N1 in dairy in the United States now and cows, let’s say not just dairy, but in all cows? Do we have any data on that?
Sam: We have some. So the USDA is looking pretty actively in the beef cattle population and hasn’t found any evidence that it’s in beef cows. So we think that it’s not particularly common outside of dairy cows. We know that California is really active in trying to find infected herds, and I think they’re at like 700 now that they’ve identified. So it’s a huge number of herds that have been infected since we detected it in California, which hasn’t been that long ago. In many other states, there’s variable degrees of active surveillance for infected herds, but when people have looked, they found it.
We also see a lot of evidence for H5N1 in the wastewater in human populations, which probably is coming from milk byproducts, from infected farms, and in many states where they have H5N1 positive wastewater samples, there’s no reported herds that are infected. So my assumption, which I think is shared by many of my colleagues is that despite the huge number of herds we’re seeing, it’s just the tip of the iceberg. That being said, I’ve not seen any estimates or any real credible attempts at modeling how many individual cows or number of herds have been infected. That would be a very useful question to answer, may in principle be answerable, at least in the ballpark, but I haven’t seen any real serious attempts at it.
Jim: If avian flu is a gastrointestinal thing in birds, is it possible that milk could be a vector to gastrointestinal version of H5N1 in humans?
Sam: Well, we know that raw milk has led to infections in cats, and cats are very susceptible, and it’s also incredibly virulent, cats. So you’ve probably seen in the news both domestic cats and more recently a large number of exotic big cats at a refuge in Washington died. Some of that is because of we think poultry that wasn’t cooked properly that was fed to the cats, but we think a lot of it is raw milk. This has led to the USDA significantly ramping up the testing of raw milk, and so I think the hypothesis, we don’t really know how it’s getting from cow to cow or farm to farm. That’s another big open question. The hypothesis maybe is that it’s shared milking equipment and we know that the milk has detectable H5N1 in it. We know that the raw milk has culturable virus, meaning that it’s actually infectious.
We know pasteurized milk is safe. It’s been very well tested quite rigorously by the FDA and USDA on that side of things. The same with eggs. But I think you’re right that the prevailing hypothesis is that the milk is playing a big part of the vectoring process. That may be why it’s not as serious in humans. A lot of the cases that you’ve heard about the pink eye or the conjunctivitis, we think it’s farm workers getting splashed with milk in the face and leading to that kind of infection, which is not likely to transmit from person to person and also not likely to find its way into the respiratory tract and become severe.
Jim: I’ll just use this as yet another opportunity to pound the table against raw milk. I know some of my listeners are part of the raw milk movement, but let me tell you, my wife who’s a biologist has looked into this and I’ve looked into it at least at the second order, and dude, whatever the benefits are, they ain’t worth the risk.
Sam: No, that’s definitely true. I mean, I think unfortunately because there were lockdowns and school closures that didn’t need to happen during the pandemic, public health has gotten a bit of a bad rap, and I’m not here to defend some of the lockdowns and school closures. I advocated for some of them at certain times when I thought they were justified. But public health, I think you could make an argument, has been one of the biggest economic drivers in the history of human civilization. And here I’m talking about sanitation, pasteurization, vaccination, things that have really allowed us to grow and thrive and reap the nonlinear benefits of spending time together with each other inside of a society.
We know that raw milk carries a whole range of different pathogens that can infect humans from bovine tuberculosis to strep and other kinds of things in addition to bird flu. So while there may be some health benefits associated with drinking raw milk, again, I think the jury is still really out on that one as well. We know for a fact that the risks far, far outweigh the benefits, especially for individuals who are maybe immunocompromised, who are pregnant or trying to get pregnant, kids. We know that cats are dying because of it, so the costs just pile up.
Jim: Interesting comment about public health. I make this comment all the time that something like 80% of the improvement in life expectancy since 1890 has been public health, not the high-tech medical miracles. They help at the edge and certainly particularly antibiotics have helped at the edge, but a whole lot of it’s learning not to shift upstream from your well and the amazing amount of investment that went into metropolitan water and sewer systems between 1875 and 1914 thereabouts. This is really important. That also, I think I’m going to pull forward a talking point that I had that sounds like because you raised it, let’s talk about it. What did we learn and mislearn from the COVID-19 experience that may be relevant to this possible H5N1 and other pandemic? It is interesting that we have socially learned some not so good things.
Sam: Well, and actually just very quickly since you mentioned antibiotics, one of the things that people ask a lot is what keeps me up at night, and one of the things that keeps me up at night is antibiotic resistance. And that’s actually quite relevant in the context of influenza because 1918, we think most people actually died of opportunistic bacterial infections. That was in the pre-antibiotic era. They got flu, they had suppressed immune systems. As a result, they ended up with a bacterial infection, strep pneumo, and they died from that, not from influenza. And so one of my big concerns is that we’re actually faced with a 2009 H1N1 that’s in and of itself, not particularly severe, but we’re in a post antibiotic era where there’s lots of resistant bacteria floating around, and you have a lot of deaths that occur because of that.
Jim: We got to throw out a ruck proposals I’ve been trying to sell to people for years, which is I believe that every antibiotic ought to come with a tax calculated to generate enough revenue to do the research, to find more antibiotics because the use of antibiotics in cattle feed and some kids got a runny nose or something when it’s actually a viral thing, but they got a doctor and [inaudible 00:27:18], there really ought to be a significant tax on antibiotics such enough to fund the research because it’s not a super profitable part of pharmacology. So the pharma companies don’t spend as much as they could or should on research on antibiotics. So anyway, back to Sam. Sorry.
Sam: No, I don’t want to go down the rabbit hole here, but that’s actually a really interesting idea because there are so much antibiotic use that the tax wouldn’t need to be very large in order to generate enough revenue to make it attractive. Actually, one of the things that some of my colleagues and I have been discussing is on the artificial intelligence drug discovery side, there’s a lot of interest in trying to repurpose drugs. Say a candidate went through phase one trials, it’s safe in humans, but it didn’t have the effectiveness against whatever they were trying to use it for. It’s sitting on the shelf at a pharma company. Some of those probably have potent antimicrobial properties that we might be able to identify with AI.
And if say the pharma companies were allowed or willing to provide a very limited authorization around antimicrobial use for that compound, basically releasing part of their patent protection, we might be able to get the big foundations to pay for the phase two trials and the governments to pay for the phase three trials, and we could open up a whole new pipeline of innovation into the antimicrobial side. So it’d be fun to talk more about that. But going back to your question, actually, there are some interesting things that we learned. I was going to mention this earlier, but one of the problems we had early on with COVID-19, part of the reason why some of the fatality rates were so high early on is that we were treating it like we would treat influenza. And COVID, I’m not a medical doctor, but from my understanding, COVID is a very different kind of respiratory infection than influenza hits.
And if you were following the case in Canada where the person had very severe flu and recovered, they were treated like we would treat an influenza patient with certain kinds of supplemental oxygen, with the antiviral drugs that we have and that person recovered. So one potential benefit compared to COVID is that we actually know a lot more about how to treat influenza now. We know a lot more now about how to treat COVID. Another thing that came out of the pandemic are these mRNA vaccines. There’s already mRNA vaccines that are being developed for H5N1, there’s no guarantee that they’re going to work, but they’re already in development. A couple of other things on the positive side that came out of it, a lot of people mask these days. I know that it’s still very controversial. It’s by no means universal, but pre-pandemic, I don’t know that I would ever see anyone in most cities, most airplanes in the US wearing a mask.
And now, every day, and it’s not just because I live in Boston, but all over the US when I travel, every day I will see people wearing masks. And so we did learn something about masking. We also learned something about wastewater surveillance, which has provided us a whole new surveillance modality. Now, on the flip side, we also know that we had lockdowns, school closures, mandates around vaccines and masks that have generated a lot of pushback in part because while we know they saved a lot of lives, many of them were left in place for far too long, and we incurred too much economic and educational costs on the back end. So I think from my perspective on COVID-19 and a flu pandemic, there are plenty of things in the plus column that we actually learned and I think will deploy with high effectiveness if we’re faced with N5N1 or other avian type of influenza pandemic, but plenty of other things on the negative side.
Jim: You mentioned it as unfortunate skepticism about public health full stop.
Sam: No, I agree. I mean, the most consistent predictor, we actually published some papers on this. A number of other folks have published them as well. The sort of most consistent predictor of whether someone is going to take a public health action, vaccinate, wear a mask, really anything is whether they indicate that they trust the government or the agency that’s providing the recommendations. It’s not about political affiliation, it’s not about socioeconomic status. It’s none of those things. It’s whether they trust the institution that’s giving them the recommendation or not. And we know that not just public health, but trust in faith in government and science has fallen off a cliff over the course of the last few years coming out of the pandemic.
Jim: And frankly since the ’60s it’s been on a decline. We look at the general social science survey data. It’s amazing. Levels of trust have fallen, but the public health was probably the last to fall and took a precipitous dive. Could we have done anything smarter in the public health arena from COVID-19? From your experience, you’ve looked at this data and you’ve been thinking about this hard for a long time.
Sam: I mean, in hindsight, there are plenty of things that we could have done differently. I think one of the things that people don’t remember well enough, and I think our brains have evolved to bury things like this pretty quickly. In Boston and New York City, when we closed things down, had school closures, et cetera. We were converting convention centers into hospitals. We had refrigerator trucks because the morgues were full. Things basically, not quite five years ago now, were incredibly dire in a number of cities in the east coast of the US and perhaps Seattle and a few other places on the west coast. We were only left with bad options at that point. I think in retrospect, we did not need to have those rolled out throughout most of the rest of the US the way that they were. In retrospect, we should have had better plans in place for how to get things reopened faster, especially schools.
But again, with where we were in terms of the infection risk, the fatalities, we only had bad options. Now, I think there are some things on the positive side coming out of the pandemic. So with wastewater surveillance, we in principle should be less surprised. If we experience another kind of SARS-CoV-2 like event SARS-CoV-3, within a matter of days after identifying the genome, we could have tests being validated in the lab and within a matter of weeks or less, we could have those rolled out to the hundreds and hundreds of wastewater surveillance sites all around the US. So we would not have been surprised the way that we were with SARS-CoV-2, and we would’ve had a much more precision ability to intervene. We also learned things like the importance of ventilation. So part of the reason why some schools were slow to reopen is that they were installing new ventilation equipment in the schools so that the kids could be back safely without masks on.
We have a lot of that now. We really understand more about ventilation. There’s been big federal investment. We also learned a lot of lessons about school reopening. We’d never really closed schools like that before. We’ve never really closed businesses like that before. So my hope is that we can be much more precision going into it in the future. But certainly in hindsight, if I could go back and change one thing, it would’ve been pushing a lot harder to have reopening plans that got people back to normal much faster, especially in periods like in the summer of 2020 when the number of cases was pretty low, all things considered.
Jim: The other thing, when I look back at it I go, wonder why it took them so long to bring these factors into policymaking relatively quickly. We knew that the fatality rate was strongly age graded and risk factor graded. People say about vaccines, I go, “Hell. I’m old and I got some risk factors. Of course, I’m going to do the vaccine.” Whether you’re a 22-year-old male in good shape, the math might be different. And it did not seem that the relatively rapid learning on age and risk factor grading was applied to policy very quickly compared to, say, Sweden, who decided that the theory is to isolate the higher risk and let everybody else go about their business more or less. The US and most of the rest of the world did not come to that conclusion. As I recall, I haven’t looked at the most recent data, but somewhere along the line, it looked like the overall fatality rate in Sweden was comparable to other countries of its same socioeconomic level.
The other one, which I think was just, again, everybody was scrambling. We didn’t know, but this idea that the initial shutdown would stop the pandemic. Looking back at it, what they should have said was, to your point, hospitals are overwhelmed, morgues in the freezer trucks. This may be hard to explain to a math illiterate country, but that what we really need to do and about, all we can do is flatten the curve unless we’re going to go to an extremely stringent lockdown. And if they had communicated it that way, people would not, I think have taken away, oh, those don’t know what they’re doing. Why are they doing this? That didn’t work. What the fuck.
Sam: I think on the Sweden piece, we know that there are other differences. Average household size, et cetera, makes some of their measures a little bit easier than it would be in the US. One other thing that we have now are rapid tests. And so one reason why I might get vaccinated is because I’m worried about long COVID. We didn’t know about long COVID back in 2020, but that might be a reason why a younger person would get vaccinated. Another reason I might get vaccinated is I don’t want to pass it to parents or individuals that are high risk, but I do have rapid tests now. I woke up with a bit of congestion yesterday, and I was able to get a rapid test that tested for COVID, flu A and flu B on the same little test strip and see that all of those are negative.
And that means for me, maybe I will mask if I am going to be around high risk people or something, or isolate for a day, but I know that I’m probably not going to infect anybody. We have these tools now, which is quite powerful. To your point, one of the things that I think we don’t give enough credit to is the public. So I would be on a lot of shows, sports radio and stuff in Boston, and one of the questions that I got asked around Thanksgiving, I think it was Thanksgiving in 2020, is we’d had this recommendation about not gathering in large groups to celebrate, but many restaurants were open. And the question I got was explain to me how that’s compatible. And I said, “Well, they’re not compatible. Either they’re both safe or they’re both not safe.”
And I think to your point, the public doesn’t know enough to make policy level decisions around school closure. That’s why they elect people that have that knowledge, have that experience. They recruit experts, they have all these panels, et cetera, but the public is sure, smart enough to understand whether these policies make any sense or not, or whether they’re having the impact that they were supposed to have, and also whether the policymakers and the scientists were being transparent as possible about the assumptions going into this, what we actually thought was likely to happen, et cetera. And so I think more open and regular dialogue with the public increases trust and allows for the kinds of honest good in intentioned mistakes that are going to happen as we learn about something that’s brand new.
A really interesting example of this, Dr. Nirav Shah, who’s now at the US CDC, was in charge of the response in the state of Maine, and Dr. Shah was incredibly popular and trusted in the state, in part because he basically went on TV every afternoon and told everybody, here’s everything we know. Here’s everything we don’t know. Here’s what we’re doing. Here’s why we think this is the right thing to do. Here’s what we’re going to do differently if we start to see these different signals and was just out there being open and honest with people day after day. And again, the public doesn’t need to have a PhD in EPI or a master’s in public health or any of that stuff to understand whether someone is being open and honest with them and introspective around these complicated policy decisions.
Jim: And good faith is huge. I mean, again, I don’t know the ins and outs of it, but at least the stories that flitter around the internet, we know what they’re worth, that maybe a lot of the masking was in bad faith. They knew it wasn’t going to work, but they said do it anyway just so people would be doing something. I don’t know. Do you have a view on that?
Sam: From every politician that I interacted with, and I spent a lot of time mostly with local mayors around Massachusetts, folks in Maine, Indiana, New Hampshire, just providing pro bono advice. I think they were all just trying to figure this out. They didn’t want kids dying. They didn’t want people in their jurisdictions dying. They wanted to get things open as safely as possible, but there was just not clear guidance on what to do. The governor of Massachusetts basically abdicated in many ways, responsibility and left it up to the cities, which in some ways sounds like a good idea. Let the cities do what they think is most important. But if Boston decides to do one thing and Cambridge decides to do a different thing, it doesn’t matter, it’s going to whatever the worst thing is is what’s going to propagate across both of those cities. People are going to move around.
So the reason we have governors and the reason we have federal government is that you often need to have coordination across cities, across states for a response like this. And that’s why it’s important that we have these kind of coordination. I think it would make us more effective, especially at pivoting off of strategies that aren’t good. I’m sure there were bad actors, there’s always bad actors, but the folks that I worked with, and there were lots of people across all different aisles of politics, they wanted to do the right thing, and they were acknowledging that none of us really knew what we were doing here and were pivoting, trying to follow the science, trying to follow the economics and trying to do the best they could.
Jim: Well, that’s good to hear. That’s good to hear. So I’m going to actually pull another question up from late in the podcast since we talked about, we’ve been talking about the lockdown and other things. What have we learned about how we might respond if it is the big one, let’s say it’s this R-naught of 3.5, Poisson distributed approximately with a case fatality rate of 10%, which is in the middle of year five to 20 and is probably at the higher range of the Spanish flu, maybe something like that. As you’ve point out, we can monitor a lot more now. We can just have a hell of a lot more transparency, what’s going on. Let’s say we’re at a relatively early stage equivalent of say, February 2020, and we’ve concluded that this is the big one. What could we do?
Sam: Well, one of the things that we saw time and again during the pandemic when we looked at the data on how people were responding, if people are afraid and they have the privilege to do it, meaning they can stay home from work, they can pull their kids out of school, they do it, and that transcends politics, that transcends, doesn’t necessarily transcend socioeconomics because there’s some difference in terms of how people can actually stay home and still keep their jobs. And many of the essential workers were individuals that were doing, picking up garbage, working in custodial services, things that are really essential services, they don’t have the ability to stay home. So I think what would happen if we had this 10% fatality is that most people would stay home even regardless of what the government said to do, especially until we knew more about what was going on.
One of the things that we would see right away, and this is now top of mind for a lot of people, hospitals run at 90, 95% capacity on terms of their beds on any given day. And that’s not because of a pandemic, that’s because of the need for surgeries. We call some of these things not really elective, but what we mean is that these surgeries, the person is not going to die today if they don’t have the surgery. It doesn’t mean they don’t have cancer and they might not die in 10 days if they don’t have it, or two weeks or a month, is that if we see that kind of fatality rate, that means almost certainly that the hospitals will be overwhelmed because those individuals that are passing away, at least initially, are going to be doing it at the hospitals.
And so we would see widespread disruption in terms of our ability to provide care outside of the big one inside of the hospitals. And I think that that would lead the government to have to make really hard choices. And again, this is part of the reason why so many of us are advocating for more precision public health in terms of leveraging wastewater surveillance, pre-staging vaccines, ensuring that we have plans in place for how we’re going to do the right kinds of targeted intervention to bring things under control without the whole scale shutdowns or people shutting everything down themselves because they get scared.
If we don’t have those, we’re again going to have only the bad options like we were left with in 2020, about five years ago. And that’s partly what we’re worried about with the H5N1, is that this thing has been kicking around now for about a year. It’s getting worse and not better. We need to have coordinated action plans in place for what we do if this thing starts spreading effectively person to person so that we can stop it and prevent it from becoming the kinds of things where again, it will basically be taken out of our hands what the decisions are going to be. Because if there are too many bodies for the morgues, if the beds are full, people are dying at home. This is the kind of situation we’re talking about with a 10% fatality rate, if we start to see it really take off from an epidemic perspective.
Jim: I’ve had the good fortune or bad fortune to be on an informal panel advising one of the three letter agencies on ideas on how to respond to a weaponized smallpox attack on the United States. This was back in the ’90s, and this group of supposedly smart folks, none of us were experts, but we were all thought to be smart people. It’s why we were brought together. We came to the conclusion there was only one possibility, which was that every household in America needed to have a month, a 28-day supply of food, and that there ought to be somebody whose job it was to analyze the data, and if the conclusion was that we were in a runaway weaponized smallpox attack, which by the way would kill 50% of the population probably, if not dealt with, would pull the whistle and basically say, everybody must stay home for 28 days.
Some of you will die because you’re not going to get your medications, but there are no essential services. Everybody must stay home. A rudimentary amount of the National Guard will be patrolling the streets. Anybody outside will be shot. Sorry. That sounds kind of crazy, but we couldn’t come up with any other thing that was likely to actually work.
Sam: I mean, I don’t know what else we would be able to do aside from paying people to stay home.
Jim: Or if they don’t have food, pay them doesn’t do any good. That was our quick analysis was the need for essential services undermines the ability to actually stop cold, super dangerous pandemic. And so you had to have the ability for everybody essentially to stay home, except people working at the water works, on the electric grid, and that’s about it. Even, sorry, no ambulances. Some of you’re just going to die, but it’s better that than 50% of the population dying.
Sam: I mean, again, you imagine though, just back of the envelope, if we had to pay 1% of the people to stay home, let’s say that costs $25,000 for a month per person. That sounds like a lot, I guess.
Jim: Probably more like $4,000 a person would be more accurate.
Sam: That might be two to 3% of the 6 trillion federal budget to do that. But again, if you could stop the big one by paying two to 3% of the federal budget, we would benefit from that on the back end for sure. But I mean, one of the things that you mentioned, the problem here is you’d really have to eliminate it, and then you’d have to have a system in place to keep it out. Because what will almost certainly happen with an H5N1 is if it starts spreading, it’s going to rapidly become an international problem because of how well-connected the US is, and basically, we won’t be able to eliminate it. There won’t be any way that we can have the kind of border screening in place that we would need to prevent it from coming back into the US. We saw what had to happen internationally, which is two-week hotel quarantines for everyone coming back into Australia and New Zealand, hard border closures.
We couldn’t do that in the US. It wouldn’t work. It would be too expensive, and I don’t think there would be a political appetite for it. And so even if we spent the three to 5% of our, and I guess close it to $7 trillion budget to eliminate it from the US, even if that worked, which it probably wouldn’t, we wouldn’t be able to keep it from coming back in again. And so I think either we need to stop it now or we’re going to have to have plans in place for rapid vaccine rollout and precision intervention neighborhood by neighborhood to prevent hospitals from getting overwhelmed while keeping things open.
Jim: I do wonder why can’t we close the United States down? We are a pretty autarkic country compared to almost all others. If we said, sorry, nobody, zero, do the New Zealand thing for six months or a year, nothing bad would happen to the United States. We’d have less crappy Chinese toys, probably. That’s about it.
Sam: Well, certainly from an import perspective, it would be a lot more than that. I mean, oil and gas would be huge.
Jim: Well, we’re self-sufficient in oil and gas now.
Sam: We could be self-sufficient with enough lead time for a period of time, but certainly in terms of the cost at the pump, because that’s not really about supply and demand, that’s about people betting on the futures markets. I mean, that would be a big mess. For me, I think the actual real question, and I’m not a constitutional lawyer. I mean, I know that the federal government has pretty broad authority over the border. I don’t know what would happen if they closed the border down from the perspective of US citizens that have the right to come back into the country. They may be able to do that. I can’t imagine that that would be popular by any stretch. And so I don’t see how it happens politically.
Jim: And not fast enough. My take, that’s what on this red team thing is you’re not going to be able to make these decisions on the fly. You have to have a file that says, when this occurs, you do this. Because we have studied this extensively, and this is the only known answer, even if it is catastrophically expensive, and that might include closing the border down and bad shit happens. You mentioned 3% of the budget. World War II. The US at its peak was at about 30 to 40% of its GDP being spent on war making. Totally unprecedented. In Germany, it was 70%. In Russia, it was 80%. So in extremis, countries go all in, and even Russia only lost about 10% of its population in World War II, which was the highest of any country. So a loss on the scale of a Spanish flu or a 10% kill rate by that math would be spending five years worth of GDP or a few years worth of GDP. $20 trillion would be justified. But again, politically-
Sam: But then there’s also the question of whether anyone would give us the loan for the money.
Jim: Well, of course. Fortunately, we just print it for a while.
Sam: That would work for a period of time. Again, I think-
Jim: Well, that’s another side conversation, but I can tell you there’s a trick to get out of it on the other side. But if it really was, again, it’s epistemic problem, how do you know that it’s the big one? But once you do, it seems to me you should be willing to spend any price to avoid it. And because it’s so hard and so controversial, unfortunately, particularly with the political backwash, which has come back from the COVID, good, bad, and the ugly, it seem to me that if you could convene a group of the smartest people to work for a year on, what would we actually do if someone made the call that this was the big one, and what can we do now to prepare for that? Let’s say we can allocate just 1% of the federal budget, which would be about 6 billion a year to preliminary.
It’ll be more than that. It would be 60 billion a year. Fuck, that’s real money. Hell, that keep me in beer and hot dogs for a while. Say 60 billion a year to prepare for calling the play that it’s the big one to pre-stage, build large surpluses of the most basic food, making the water systems as no hands needed as possible, et cetera. And then every year you’d update the red file on where we’re at, what our affordances now, and then when the big one is called, you pull the ripcord, pull out the red file, and you execute.
Sam: Well, I actually think from my perspective, one of the big challenges would be, and you pointed this out, how you would know it’s the big one? And whether you could make that decision in time to still use any of those options in your playbook. I’ve actually been working with a nonprofit that came out of the pandemic called CAPTRS, where they’re building professional war games to try and challenge decision makers around how they actually, what kinds of data sets they look at their decision-making process, not around chain of command in terms of the response, but how you actually get to that decision that this is the big one, or it isn’t the big one, and we need to make a different call. Because I mean, my real concern is that by the time we have enough information, and we have decided it’s the big one, that even these $60 billion a year options are not on the table anymore.
It’s something that’s much larger. Of course, in the case of smallpox, my assumption would be that the vaccine would work. We would just need enough time to ramp up production and get it out to people. As we saw with COVID, there’s no guarantee that even if we have a vaccine that works well enough to protect people from severe disease, that it will stop the transmission the way we have for a measles or for a smallpox. So again, even if we had the big one for smallpox, there would likely be an end game with a tight timeline based on production such that you could model out the economic costs. But for an H5N1, there’s no guarantee that we don’t end up in an evolutionary arms race with this thing, and it’s years and years and years before we’re into something that’s more manageable.
Jim: Which is more or less like COVID, where yes, we get ahead a little bit and then it catches up. What is the current state of play of vaccine prep? I mean, we obviously get our flu shots every September, so we know how to make flu vaccines. We also know that they work plus or minus. What is the current state of play on vaccine preparation for H5N1?
Sam: So the federal government has a strategic stockpile. They have to be what’s called fill finished, meaning actually taken out of the stockpile and prepped to be in people’s arms. My understanding is that that has already happened. For a portion of the stockpile, it’s probably only enough to really provide protection for a key set of first responders, which would include medical professionals and others. But that has been activated. The results now suggest that it’s a good match for the circulating H5N1. Part of what happens with seasonal flu, there’s two components associated with the vaccine not working as well as we’d like. One is just trying to use the crystal ball to figure out what exactly the flu strain is going to be each year. Sometimes we get that right. Sometimes we get it wrong.
But the other piece is that as we grow these up inside of chicken eggs, which is the most common way of manufacturing the virus to get the antigens to build the vaccine, it picks up mutations and it shifts off of that human-like strain to pick up some sort of bird-like mutations. And the resulting vaccine is less effective in humans as a result of the manufacturing process. We have some new manufacturing techniques, one that involves tissue culture, which is probably what most people are getting now in terms of the flu vaccine. The second, as I mentioned earlier, is that the mRNA manufacturers are now working actively to try to develop mRNA vaccines against the H5N1. So my expectation is that the vaccines will be a good match.
The challenge is going to be that there won’t be enough of them for months and months and months and months, maybe a year after this thing kicks off because we’ll be prioritizing first responders as we should. The second is that over that period of time, while we’re ramping up the manufacturing of the vaccine, the virus is going to evolve. It’s going to evolve resistance to natural immunity. It’s going to evolve against the vaccines that we’ve deployed. And it may be the time, by the time we get it rolled out that the vaccine’s not as good of a match anymore. Now probably will reduce severe disease almost certainly, even if it doesn’t protect against infection. We might find ourselves in a situation that looks a lot more like COVID in terms of the red queen where you’re running as fast as you can just to stay in the same place.
Jim: And it is a good point just to remind people it’s still worth getting if you’re at a risk, still worth getting the COVID vaccine. I do, and all the new ones come out and not so much it will keep you from getting it, but the chances of getting a debilitating hospitalization grade case is still reduced by about a factor of at least 15. So it’s worth it if you’re at risk.
Sam: I was thinking whether or not I was going to get the mRNA booster again, just because I have side effects. It puts me on a commission for about a day, and I was weighing, I’m young, I’m healthy. I’ve had a bunch of doses of these. I had COVID for the first time over the summer, maybe I don’t need to do it. And then I actually had the Novavax and I had zero side effects from that. It’s basically just like the flu shot. It’s a bit of a sore arm. And so one of the other things we have going for us now on the vaccine side is that we have a vaccine that’s as effective as the mRNA vaccines, but you really lose that side effect that I think we’re putting a lot of people off around getting the boosters.
Jim: That’s interesting. Novavax. I don’t know. I think I got the Pfizer last time. I had the usual… For me, every time it’s the same. It is basically the next day I get pretty tired by about 2:00 in the afternoon, just have to go to bed. So it’s like… But that’s it. After the day after that, I’m fine. So it’s well worth the cost for me. Everyone was different.
Sam: Better than a ride on a ventilator or worse.
Jim: Exactly. Exactly. And then the other one, I now have three friends who have long COVID or one of them is just maybe finally kicking it, but two of them them had it for years. And I mean, that’s a non-trivial fucked up thing. And then one or 2% of people that get COVID end up with long COVID. So that’s another good reason, even if you’re young and healthy.
Sam: I think the jury’s still a bit out on this, but it’s part of the reason why when I had COVID back over the summer, I took Paxlovid because that is very effective at bringing the viral load down real quick. And part of what they think may be going into the long COVID is that peak viral load. Again, the science is still out on that, but it wasn’t quite like an antibiotic in terms of feeling amazing within 24 hours. But I had pretty serious symptoms that were basically gone within 48 hours of starting those meds.
Jim: My doc’s instant Paxlovid dude. So the instant you get a diagnosis, we have a local pharmacy, they’ll just dish it right out. So I need a prescription of course. But yeah, very important. Now, we’ve talked about a bunch of dire bad scenarios, things that could work, what happens if it is the big one? We’re not sure how we can stop it at all, actually, but fortunately at the moment, the best thinking seems to be that the 2025 risk is relatively low. I went on medic calculus this morning, is that they pronounce it medic calculus. It’s basically an online prediction market. And the prediction for a 2025 outbreak in the US of avian flu of any sort that’s highly virulent is 3%. Is that right, do you think?
Sam: I don’t know that that gives me a lot of comfort though. I mean, if you told me there was a 3% chance that one to 5% of the US population could die in a year.
Jim: Yep. Oh, I know. This is this thing that our brains, humans were not evolved for the low probability, but continual rolls of the dice. And some people think there’s a one or 2% chance of a nuclear war every year, which is probably correct. A scary ass thing. And we just keep rolling them sevens, but one of these days it’s going to come up snake eyes.
Sam: Well, one of my PC advisors, they probably stole this from somebody else, but they used to say it’s one of the reasons why probability was the last branch of mathematics discovered, because our brains don’t work that way and it’s hard for us.
Jim: I love that. I love that. That’s great.
Sam: A couple of thoughts around that. So the first is I paid almost $12 for a dozen eggs over the holidays. They’re already pretty significant economic, and those prices will come down because people don’t buy as many eggs after the holidays, but there’s pretty significant economic consequences. We know what’s happening on the dairy side, significant reductions in milk volume costs that are being incurred by the federal government and the states also the farms, I’m sure. So there’s already pretty significant economic consequences, and there’s not really any evidence that those are going to get better. They might just get worse as we move into seasons where cows are moved around for feed down into Texas. I mean, we’re only about a month out from when the outbreak started down there. So we might just see this thing rolling around the country and be dealing with that on the economic side of things.3% is, it’s not 30%, but it’s also not 0.3%. And as you said, if we’re thinking about nuclear war, we do a lot to try to prevent that from happening. And that’s maybe now half the risk of a flu pandemic.
So we are seeing some coming from the federal government, we heard that the Biden administration’s going to put $300 million in towards preparedness and mitigation. It’s the kind of thing that you were talking about is making sure hospitals have plans in place. They’re ready to go if this kind of thing happens. We have tools now from wastewater surveillance. We’ve got, we have vaccines. The USDA is starting to run trials around vaccinating cows. I think we should be offering the vet, not mandating, but I think we should be offering the vaccines to farm workers if they want them. If it were me, I think we’ve got more than enough evidence to be taking some pretty decisive action to get the economic costs on the agricultural side back under control and bring this 3% down to 0.3% or something without incurring, we’re not talking about billions of dollars in costs here. We’re talking about maybe a billion dollars over the next year that would be spent. And I think we’d get all that back, at least on the agricultural side from an economic perspective.
Jim: And as you mentioned earlier, it sounds like your recommendation of the strongest short-term move would be to vaccinate the cattle?
Sam: We need to make sure that the vaccines are safe, those trials are ongoing. I fully suspect that they will, it’s not unprecedented at all to vaccinate cows against diseases like this. We don’t get it out of the dairy herds, then it’s going to be a persistent threat, and it’s just a matter of time until there’s a human outbreak, it’s just going to keep spilling over and spilling over and spilling over. One of the things that this is not a benefit because it’s bad for the birds and it’s bad on the agricultural side, but it’s incredibly virulent in poultry. This is why we call flocks. If H5N1 or other avian flu is discovered, and so it’s not like it’s a persistent threat because we eliminate it by culling, and you don’t have the situation where it’s just the roulette wheel spinning, the roulette wheel spinning, the roulette wheel spinning, and that’s what’s happening with these cows just being persistently infected.
And it’s also probably spilling back over into wild birds, just how it’s moving around the country. It’s spilling over into poultry populations, incurring huge costs and exposing more farm workers. So I don’t see how we get out of this without getting it out of the dairy populations. And we’ve tried with the movement restrictions, that seems to have helped in some places with the testing requirements, but I think it’s becoming more and more clear that this doesn’t work without vaccination.
Jim: Any final thoughts?
Sam: I think just the final thought is that there’s still a lot we don’t know, and part of what the federal government is investing in is helping us better understand how this thing is moving around, how it’s evolving, how well the vaccines are going to work. 3% is not an insignificant risk. It’s not a panic in the streets risk, but it’s exactly the moment where we can take measured action. Not talking about lockdowns, we’re not talking about mandates, but we’re talking about targeted precision action, leveraging things like wastewater surveillance to bring this under control before it becomes a 10 or 15% problem, or before it kicks off.
Jim: All righty. Well, this has been wonderful. Just what I was hoping we’d get is a great tour of the horizon of what the factors are, what the knobs are, what can be done, what we still don’t know, which is always a lot. So thank you, Sam Scarpino.
Sam: Thanks so much for having me, Jim.
Jim: It’s been great. Thanks.