Jay Bhattacharya: The Case Against Lockdowns | Lex Fridman Podcast #254

The following is a conversation with Jay Bhattakaria, professor of Medicine, Health Policy and Economics at Stanford University. Please allow me to say a few words about lockdowns and the blinding destructive effects of arrogance on leadership, especially in the space of policy and politics. Jay Barakaria is the co author of the now famous Great Barrington Declaration, a one page document that in October 2020 made a case against the the effectiveness of lockdowns. Most of this podcast conversation is about the ideas related to this document and so let me say a few things here about what troubles me. Those who advocate for lockdowns as a policy often ignore the quiet suffering of millions that it results in, which includes economic pain, loss of jobs that give meaning and pride in the face of uncertainty, the increase in suicide and suicidal ideation, and in general the fear and anger that arises from the powerlessness forced onto the populace by the self proclaimed elites and experts. Many folks whose job is unaffected by the lockdowns talk down to the masses about which path forward is right and which is wrong. What troubles me most is this very lack of empathy among the policymakers for the common man and in general for people unlike themselves. The landscape of suffering is vast and must be fully considered. In calculating the response to the pandemic with humility and with rigorous open minded scientific debate, Jay and I talk about the email from Francis Collins to Anthony Fauci that called Jay and his two co authors fringe epidemiologists and also called for a devastating published takedown of their ideas. These words from Francis broke my heart. I understand them, I can even steel man them, but nevertheless, on balance they show to me a failure of leadership. Leadership in a pandemic is hard, which is why great leaders are remembered by history. They are rare, they stand out and they give me hope. Also, this whole mess inspires me on my small individual level to do the right thing in the face of conformity despite the long odds. I talked to Francis Collins. I talked to Albert Burla, Pfizer's CEO. I also talked and will continue to talk with people like Jay and other dissenting voices that challenge the mainstream narratives.

And those in the seats of power.

I hope to highlight both the strengths and weaknesses in their ideas with respect and empathy, but also with guts and skill. The skill part I hope to improve on over time, and I do believe that conversation and an open mind is the way out of this. And finally, as I've said in the past, I value love and integrity far, far above money, fame and power. Those latter three are all ephemeral. They slip through the fingers of anyone who tries to hold on and leave behind an empty shell of a human being. I prefer to die a man who lived by principles that nobody could shake and a man who added a bit of love to the world. This is the Lex Friedman podcast. To support it, please check out our sponsors in the description. And now here's my conversation with Jay Battacaria.

To our best understanding today, how deadly is Covid? Do we have a good measure for this very question?

So the best evidence for Covid, the deadliness of COVID comes from a whole series of seroprevalence studies. Seroprevalence studies are these studies of antibody prevalence in the population at large. I was part of the very first set of seroprevalence studies. One in Santa Clara county, one in LA county, and one with Major League Baseball around the U.S. if I may.

Just pause you for a second. If people don't know what serology is and seroprevalence, it does sound like you say zero prevalence. It's not. It's sero. And serology is antibodies. So it's a survey that counts the number of antibodies specific to Covid.

Yes.

People that have antibodies specific to Covid, which perhaps shows an indication that they likely have had Covid. And therefore this is a way to study how many people in the population have been exposed to or have had Covid.

Exactly. Yeah, exactly. So the idea is that we don't know exactly the number of people with COVID just by counting the people that present themselves with symptoms of COVID Covid has, it turns out, a very wide range of symptoms possible ranging from no symptoms at all to this deadly viral pneumonia that's killed so many people. And the problem is, if you just count the number of cases, the people who have very few symptoms often don't show up for testing. They're outside of the ken of public health. And so it's really hard to know the answer to your question without understanding how many people are infected, because you can probably tell the number of deaths. That's even though there's some controversy over that. But that. So the numerator is possible, but the denominator is much harder.

How much controversy is there about the death? We're going to go on million tangents. Is that okay? We're going to. I have a million questions. So 1. I love data so much, but I like almost tuned out paying attention to Covid data because I feel like I'm walking on shaky ground. I don't know who to trust. Maybe you can comment on different sources of data, different kinds of data. The death one, that seems like a really important one. Can we trust the reported deaths associated with COVID or is it just a giant, messy thing that mixed up? And then there's this kind of stories about hospitals being incentivized to report a death as COVID death.

So there's some truth in some of that. Let me just. Let me just talk about the incentives. So in the United States, we passed this CARES act that was aimed at making sure hospital assistants didn't go bankrupt. In the early days of the pandemic, the couple things they did. One was they provided incentives to treat COVID patients, tens of thousands of dollars extra per COVID patient. And the other thing they did is they gave a 20% bump to Medicare payments for elderly patients who treated with COVID The idea is that there's more expensive to treat them at, I guess, the early days. So that did provide an incentive to sort of have a lot of COVID patients in the hospital, because your financial success of the hospital, or at least not lack of financial ruin, depended on having many COVID patients. The other thing on the death certificates is that reporting of deaths is a separate issue. I don't know that there's a financial incentive there, but there is this sort of, like, complicated, you know, when you fill out a death certificate for a patient with a lot of conditions, like, let's say a patient has diabetes, a patient that, while that diabetes could lead to heart failure, failure, you have a heart attack, heart failure, your lungs fill up, then you get Covid, and you die. So what do you write on the death certificate? Was it Covid that killed you? Was it the lungs filling up? Was it the heart failure? Was it the diabetes? It's really difficult to disentangle. And I think a lot of times what's happened is people have erred on the side of signing. It's Covid. Now, what's the evidence of this? There's been a couple of audits of death certificates in places like Santa Clara county, where I live in Alameda County, California, where they carefully went through the death certificate, said, okay, is this reasonable to say this was actually Covid, or it was Covid incidental. And they found that about 25%, 20, 25% of the deaths were more likely incidental than directly due to Covid. I personally don't get too excited about this. I mean, it's a philosophical question, right? Like, ultimately, what kills you? Which is an odd thing to say if you're not in medicine, but really it's almost always multifactorial. It's not always just the bus hits you. The bus hits you, you get a brain bleed. Was the brain bleed that killed you, would it have burst anyway? I mean, the bus hit you, killed you. Right.

The way you die is a philosophical question, but it's also a sociological and psychological question, because it seems like every single person who's passed away over the past couple years, kind of the first question that comes to mind, was it Covid, just because you're trying to be political, but just in your mind?

No, I think there's a psychological reason for this. Right. So we spent the better part of at least a half century in the United States not worried too much about infectious diseases. The notion was we'd essentially conquered them. It was something that happens in faraway places to other people. And that's true for much of the developed world. Life expectancies were going up for decades and decades. And for the first time in living memory, we have a disease that can kill us. I mean, I think we're effectively evolved to fear that, like the panic centers of our brain, the lizard part of our brain takes over, and our central focus has been avoiding this one risk. And so it's not surprising that people, when they're filling out death certificates or thinking about what led to the death, this most salient thing that's in the front of everyone's brain, would jump to the top.

And we can't ignore this very deep psychological thing when we consider what people say on the Internet, what people say to each other, what people write in scientific papers, what everything. It feels like when Covid has been brought onto this world, everything changed in the way people feel about each other, Just the way they communicate with each other. I think the level of emotion involved, I think in many people, it brought out the worst in them for sometimes short periods of time, and sometimes it was all therapeutic, like you were waiting to get out, like the darkest parts of you. Just to say, if you're angry at something in this world, I'm going to say it now. And I think that's probably talking to some deep, primal thing, that fear we.

Have for formalities of all different kinds.

And then when that fear is aroused and all the deepest emotions, it's like a Freudian psychotherapy session.

But across the world, it's something that psychologists are going to have a field day with for generation trying to understand. I think what you say is right. But piled on top of that is also this sort of. This impetus to empathy, the empathized compassion toward others, essentially militarized. Right. So I'm protecting you by. By some actions. And those actions, if I. If I don't do them, if you don't do them, well, that must. Must mean you don't. You. You hate me. It's created this, like, social tension that I've never seen before. And we. And we. And we start. Have started. We started. We. We looked at each other as if we were just simply sources of germs rather than people to get to know, people to enjoy, people to get, you know, to learn from it. Colored basically almost every human interaction for every human on the planet.

Yeah. The basic common humanity. It's like you can wear a mask, you can stand far away, but the love you have for each other when you look into each other's eyes, that was dissipating and by region, too. I've experienced, having traveled quite a bit throughout this time. It was really sad. Even people that are really close together, just the way they stood, the way they looked at each other, and it made me feel for a moment that the fabric that connects all of us is more fragile than I thought.

I mean, if you walk down the street or if you did this during COVID I'm sure you had this experience where you walk down the street if you're not wearing a mask, or even if you are, people will jump off the sidewalk as you walk past them, as if you're poison. Even though the data are that Covid spreads indifferently outdoors, if at all, really outdoors. But it's not simply a biological or infectious disease phenomenon. Epidemiological. It is a change in the way humans treated each other. I hope temporary.

I do want to say on the flip side of that. So I was mostly in Boston, Massachusetts, when the pandemic broke out. I think that's where I was.

Yeah.

And then I got came here to Austin, Texas, to visit my now good friend Joe Rogan. And he was the first person without pause. This wasn't a political statement. This was anything. Just walk toward me, give me a big hug and say, it's great to see you. And I can't tell you how great it felt because I, in that moment, realized the absence of that connection back in Boston over just a couple of. Over a couple of months. And we'll talk about it more. But it's tragic to think about that distancing, that dissolution of common humanity at scale, what kind of impact it has on society, just across the board, political division, and just in the quiet of Your own mind, in the privacy of your own home, the depression, the sadness, the loneliness that leads to suicide. And forget suicide, just low key suffering.

Yeah, no, I think that's the suffering, that isolation. We're not meant to live alone. We're not meant to live apart from one another. That's of course the ideology of lockdown is to make people live apart, alone, isolated, so that we don't spread diseases to each other. Right. But we're not actually designed as a species to live that way. And what you're describing, I think if everyone's honest with themselves, have felt, especially in places where lockdowns have been sort of very militantly enforced, has felt deep into their core.

Well, if I could just return to the question of deaths.

You said that the data isn't perfect.

Because we need these kind of seroprevalence surveys to understand how many cases there were to determine the rate of deaths. And we need to have a strong.

Footing in the number of deaths.

But if we assume that the death, the number of deaths is approximately correct, like what's your sense? What kind of statements can we say about the deadliness of COVID across different demographics? Maybe not in a political way or in the current way, but when history looks back at this moment of time, 50 years from now, 100 years from now, the way we look at the pandemic 100 years ago, what will they say about the deadliness of COVID I.

Mean, I think the deadliness of COVID depends on not just the virus itself, but who it infects. So probably the most important thing about it, about the deadlines of COVID is this steep age gradient in the mortality rate. So according to these seroprevalent studies that have been done now, hundreds of them, mostly from before vaccination, because vaccination also reduces the mortality risk of COVID The seroprevalence studies suggest that the risk of death if you're, say, over the age of 70 is very high. You know, 5% if you get Covid, if you're under the age of 70, it's lower, 0.05, but there's not a single sharp cutoff. It's more like I have a rule of thumb that I use. So if you're 50, say the infection fatality rate from COVID is 0.2% according to the seroprevalence data. That means 99.8% survival if you're 50. And for every seven years of age above that, double it. Every seven years of age below that, have it. So a 57 year old have a 0.4% mortality. A 64 year old would have a 0.8% and so on. And if you have a severe chronic disease like diabetes or if you're morbidly, morbidly obese, it's like adding seven, seven years to your life.

And this is for unvaccinated folks.

This is unvaccinated before Delta.

Also, are there a lot of people that would be listening to this with PhDs at the end of their name that would Disagree with the 99.8, would you say so?

I think there's some disagreement over this and the disagreement is about the quality of the seroprevalence studies that were conducted. So as I said earlier, I was a senior investigator in three different seroprevalence studies very early in the epidemic. I view them as very high quality studies. In Santa Clara county. What we did is we used a test kit that we obtained from someone who works in major league Baseball. Actually he ordered these test kits very early in March 2020 that measures very accurately measures antibody levels, antibodies in the bloodstream. These test kits were eventually were approved by, had an EUA by the emergency use authorization by the fda sort of shortly after we did this. And it had a very low false positive rate. False positive means if you don't have these Covid antibodies in your bloodstream, the kit shows up positive. Anyways, that turns out to happen about 0.5% of the time. And based on studies, a very large number of studies looking at blood from 2018, you try it against this kit and 0.5% of the time, 2018, there shouldn't be antibodies there to Covid. So if it turns positive, it's a false positive, it's 0.5% of the time. And then like a false negative rate, about 10%, 12%, something like that. I don't remember the exact number, but the false positive rate is the important thing there, right? So you have a population in March 2020 or April 2020 with fraction of patients have been exposed to Covid. You don't know how much, but low, even a small false positive rate could end up biasing your study quite a bit. But there's a formula to adjust for that. You can adjust for the false positive rate, false negative rate. We did that adjustment and those studies found in a community population, so leaving aside people in nursing homes who have a higher death rate from COVID that the Death rate was 0.2% in Santa Clara county and in LA county, across.

All age groups in a community Community meeting just like regular folks.

Yeah, so that's actually a real important question too. So the Santa Clara study, we did this Facebook sampling scheme which is, I mean not the ideal thing, but it's like it was very difficult to get a random sample during lockdown where we put out an ad on Facebook soliciting people to volunteer for the study. Randomly selected set of. We were hoping to get a random selection of people from Santa Clara county, but the people who tend to volunteer were from the richer parts of the county. Like I had Stanford professors writing begging to be in the study because they wanted to know their antibody levels. So we did some adjustment for that in LA County. We hired a firm that had a pre existing representative sample of LA county, but it didn't include nursing homes, it didn't include people in jail, things like that. Didn't include the homeless populations. So it's representative of a community dwelling population both of those. And there we found that both in LA county and Santa Clara county in April 2020, something like 40 to 50 times more infections than cases in both places. So for every case that had been reported to the public health authorities, we found 40 or 50 other infections, people with antibodies in their blood that suggested that they'd had Covid and recovered.

So people were not reporting or severe, at least in those days under reporting?

Yeah, I mean there was, you know, there's testing problems. I mean there weren't so many tests available. People didn't know a lot of a lot of them. We asked a set of questions about the symptoms they'd faced and most of them said they faced no symptoms or the most 30, 40% of them said face their no symptoms.

I mean even these days how many people report that they get Covid when they get Covid. Okay, have those numbers that 0.2% has that approximately held up over time, that is.

So if Professor John Ioannidis, who's a colleague of mine at Stanford, is a world expert in meta analysis, probably the most cited scientist on earth, I think, at least living. He did a meta analysis of now 100 or more of these seroprevalence studies. And what he found was that that 0.2% is roughly the worldwide number. In fact, I think he cites this lower number, 0.15% as the median infection fatality rate worldwide. So we did these studies and it generated an enormous amount of blowback by people who thought that the infection fatality rate is much higher. And there's some controversy over the quality of some of the other studies that are Done. And so there are some people who look at this same literature and say, well, the lower quality studies tend to have lower IFRs. The higher quality studies, IFR. Oh, infection fatality. Right, I apologize. I do this in lectures too, and.

I'm going to rudely interrupt you and ask for the basics sometimes, if it's okay.

No, of course. So these higher quality studies, they say, tend to produce higher ir, but the problem is that if you want a global infection fatality rate, you need to get seroprevalent studies from everywhere, even places that don't necessarily have the infrastructure set up to produce very, very high quality studies. And in poor places in the world that like places like Africa, the infection fatality rate is incredibly low. And in some richer places like New York City, the infection fatality rate is much higher. There's a range of ifrs, not a single number. This sometimes surprises people because they think, well, it's a virus, it should have the same properties no matter where it goes. But the virus kills or infects or hurts in interaction with the host. And the properties of both the host and the virus combine to produce the outcome.

But you also mentioned the environment too.

Well, I'm thinking mainly just about the person. Like, if I'm going to think about it, the most simplest way to think about it is age. Age is the single most important risk factor. So older places are going to have a higher IFR than younger places. Africa. 3% of Africa is over 65. So in some sense it's not surprising that they have a low infection fatality rate.

So that's one way you would explain the difference between Africa and New York.

City in terms of the fatality rate. Is the age, the average age.

Yeah. And especially in the early days of the epidemic in New York City, the older populations living in nursing homes were differentially infected based on because of policies that were adopted to send COVID infected patients back to nursing homes to keep hospitals empty.

What do you mean by differentially infected?

The policy that you adopt determines who is most exposed.

Right?

Okay, so that's what it means.

It's the policy, it's the person that matters.

I mean, it's not like the virus.

Just kind of doesn't care. I mean, the policy determines the nature of the interaction. And there's also, I mean, there is.

Some contribution from the environment. Different regions have different proximity, maybe of people interacting or the dynamics of the way they interact.

Like, if you have situations where there's lots of intergenerational interactions, then you have a very different risk profile than if you have societies where generations are more separate from one another. Okay, so let me just finish real fast about this. So you have, in New York, you have a population that was infected in the early days that was very likely going to die, had a much higher likelihood of dying if infected. And so New York City had a higher ifr, especially in the early days, than Africa has had. The other thing is treatment. So the treatments that we adopted in the early days of the epidemic, I think actually may have exacerbated the risk of death.

Which treatments?

Using ventilators. The over reliance on ventilators is what I'm primarily thinking of. But I can think of other things. But that also we've learned over time how better to manage patients with the disease. So you have all those things combined. So that's where the controversy over this number is.

I mean, New York City also is a central hub for those who tweet and those who write powerful stories and.

Narratives in article form.

And I remember there's quite dramatic stories about sort of doctors in the hospitals and these kinds of things. I mean, there's very serious, very dramatic, very tragic deaths going on always in hospitals. Those stories a lot. Loved ones losing each other on a deathbed. That's always tragic. And you can always write a hell of a good story about that and you should, about the loss of loved ones. But they were doing it pretty, pretty well, I would say, over this kind of dramatic deaths. And so in response to that, it's very unpleasant to hear, even to consider the possibility that the death rate is not as high as you might otherwise as you might feel.

Yeah, I was surprised by the reaction both by regular people and also the scientific community in response to those studies, those early studies in April of 2020. To me they were studies. I mean, they're the kinds of, not exactly the kinds of work I've worked on all my life, but kind of like the kind of you write a paper and you get responses from your fellow scientists and you change the paper to improve it. You hopefully learn something from it.

Well, but to push back is just a study. But there are some studies, and this is kind of interesting because I've received similar pushback on other topics. There's some studies that, if wrong, might have a wide ranging detrimental effects on society. So that's the way they would perceive the studies. If you say the death rate is lower and you end up, as you often do in science, realizing that, nope, that that was. There was a flaw in the way the study was conducted or we just not representative of a broader population. And then you realize the death rate is much higher. That might be very damaging in people's view. So. So that's probably where the scientific community sort of just steel man. The kind of response is that's where they felt like, you know, there's some findings where you better be damn sure before you kind of report them.

Yeah, I mean we were pretty sure we were right and it turns out we were right. So like when we. So we released the Santa Clara study via this open science process and this server called medRxiv, it's designed for releasing studies that have not yet been peer reviewed in order to garner comment from scientists before peer review. The LA county study, we went through this traditional peer review process and got it published in the Journal of American Medical association sometime in like July, I think, I forget the date of 2020. The Santa Clara study released in April of 2020 in this sort of working paper archive. The reason was that we felt we had an obligation, we had a result that we thought was quite important and we wanted to tell the scientific community about it and also tell the world about it. And we wanted to get feedback. I mean that's part of the purpose of sending it to these kinds of places. I think a lot of is that when people think about published science they think of it as automatically true. And if it goes through peer review, it's automatically true. If it hasn't gone through peer review, it's not automatically true. And especially in medicine we're not used to having this access to pre peer reviewed work. I mean in economics actually that's quite normal. It takes years to get something published. So there's a very active debate over or discussion about papers before they're peer reviewed in this sort of working paper way, much less normal or much newer in medicine. And so I think part of that, the perception about what process happens in open science when you release a study that got people confused and you're right, it was a very important result because we had just locked the world down in middle of March with I think catastrophic results. And if that study was right, if our study was right, that meant we'd made a mistake and not because the death rate was low. That's actually not the key thing there. The key thing is that we had adopted these policies, these test and trace policies, these policies, these lockdown policies aimed at suppressing the virus level to close to zero. That was essentially the idea. If we can just get the virus to go away, we won't have to Ever worry about it again. The main problem with our result as far as that strategy was concerned, wasn't the death rate. It was the 40 to 50 times more infections than cases. It was the 2.5% or 3% or 4% prevalence rate that we identified of the antibodies in the population. If that number is right, it's too late. The virus is not going to go to zero. And no matter how much we test and trace and isolate, we're not going to get the viral level down to zero.

So we're going to have to let the virus go through the entire population in some way.

We can talk about that in a bit. That's the Great Barrington Declaration. You don't have to let the virus go through the population. You can shield preferentially. The policy we chose was to shield preferentially the laptop class, the set of people who could work from home without losing their job. And we did a very good job at protecting them.

Well, let me take a small tangent. We're going to jump around in time, which I think will be the best way to tell this story. So that was the beginning.

Yeah. Okay. Actually, can I go back one more thing for that? Because that's really important and I should have started with this. What led me to do those studies was a paper that I had remembered seeing from the H1N1 flu epidemic in 2009. I'd been much less active in writing about that. I had written a paper or two about that in 2009. There. There was actually this same debate over the mortality rate, except it unfolded over the course of three years. Two or three years. The early studies of the mortality rate in H1N1 counted the number of cases in the denominator, counted the number of deaths in the numerator cases, meaning people identified as having H1N1 showing up to doctor tested to have it. And the earliest estimates of the H1N1 mortality were like 4%, 3%, really, really high. Over the course of a couple of more years, a whole bunch of seroprevalence studies, seroprevalence studies of H1N1 flu came out. And it turned out that there were 100 or more times of people infected per case. And so the mortality rate was actually something like 0.02% for H1N1, not the 3, 3, like a hundred fold difference.

So this made you think, okay, it took us a couple of two to three years.