David Zweig Explains Misleading Covid Hospitalization Data

CLAY: We had David Zweig on — I don’t know — probably a month ago or so to talk about masks on children and whether or not they made sense based on the statistical data. And, by the way, his article said that did not make sense. He’s now got a really interesting piece up that I’m encouraging you guys to all go read, looking at the data.

And the data in particular is surrounding hospitalizations and whether or not it’s a reliable metric. And the title of the piece at The Atlantic is, “Our Most Reliable Pandemic Number Losing Meaning,” and the subheading is:

“A new study suggests almost half of those hospitalized with covid have mild or asymptomatic cases.” You hear about this hospitalization number all the time. David Zweig joins us now. David, how could that be true, and what did you uncover as you reported this article that we just referenced?

ZWEIG: Yeah. It’s an interesting thing. During the pandemic, we generally have used three different metrics for kind of tracking what’s happening. There’s cases, but the problem with using cases is it really depends on who is tested and when, and that can greatly affect what we’re seeing — and also, cases doesn’t mean that someone is sick.

It just means that they’re infected. On the flip side, you have tracking mortality, which is finite. (chuckles) But it’s also a really lagging indicator, and it doesn’t count people who perhaps were really ill but recovered ultimately. So hospitalization numbers have been looked at as a really kind of good Goldilocks metric that’s definitive.

Either you’re in the hospital or not, and it tracks people who ostensibly are really ill. However, what I found in my reporting — and I write about this study, which is in preprint, which means it isn’t yet peer reviewed. But the authors of the study have a really strong track record with all their publications, and what they found is that in this year, in 2021, roughly half of all covid hospitalizations are for mild or even incidental or asymptomatic cases.

CLAY: And that’s a really interesting data point. So, roughly… I’m looking at the New York Times board right now. They say that 90,000 people right now are hospitalized with covid. Based on what your data suggests from that preprint article, roughly 45,000 of those people would have legitimate covid cases; the other 45,000 may be incidental, they might be relatively insignificant in terms of how severe they actually are?

ZWEIG: That is correct. That is what these authors found. So your listeners may be wondering, “Well, how is that possible? (chuckles) Why would people be in the hospital if they’re not really sick?” And there’s actually a good reason for it. The first is is that hospitals… It’s not necessarily a nefarious reason. Hospitals have to report to the federal government every patient who tests positive.

And most hospitals in the country are doing a universal screening on anybody who is admitted to the hospital. You are tested. So if you have someone from a car accident, someone with a broken foot or appendicitis, they’re in the hospital for something totally unrelated to covid, but in order to be admitted, they have to be tested.

The test comes back positive? Boom. That gets checked off as a, quote, “covid hospitalization,” and then another category of patients are people who showed up with covid symptoms, but they’re relatively mild and perhaps an emergency room doctor said, “Well, let’s admit this person anyway for observation because they have some underlying conditions.”

Or the patient’s complaining of a subjective feeling of shortness of breath even though their blood oxygen level is very high and seems normal, subjectively. And then their illness never progressed into anything beyond mild illness. So, you have these two large categories, and it’s pretty astonishing, but they found that they comprised approximately half of all covid hospitalizations.

And what’s interesting about this study, I had written a number of months ago about two separate studies on pediatric hospitalizations. And they used a totally different methodology. And they came to very similar conclusions with their findings. I think it was roughly 40 to 45% of the pediatric admissions were also deemed incidental or largely unrelated to covid.

BUCK: David, it’s Buck, and I was about to ask you about the pediatric covid hospitalizations, because we just mentioned this before you came on. There’s a lot of news coverage of that in August, right before the opening of schools and all the talk about what kind of mitigation measures should be in place and should it all be in-person learning. All those things, all these news stories — kind of breathless news stories — about an all-time peak in covid hospitalizations.

And I’m looking at one here from CNBC from just last month where they were saying that they were looking at an absolute peak for the week ending August 22nd when they thought this was the highest — I think it still is the highest — they had seen, of 300 admissions a day. So from what you’re telling us, it really could have been more actually like 150 a day nationwide who are hospitalized with covid as the primary reason for admission?

ZWEIG: Correct, that based on… So we now have, that I’m aware of, three separate studies. The pediatric studies, they did a method where the doctors did what’s called a retrospective chart review. They actually meticulously pored through each chart of the patients one by one. And this is in two different hospital systems in California.

Whereas the researchers on this current study, which was for an adult population, they used sort of data analysis, and they simply looked at the metric of anybody who is a covid hospital admission, and then they cross-referenced that: With those patients, how many of them had a blood oxygen level below 94%, which is NIH’s definition of severe covid.

And then they also included did any of these people get supplemental oxygen. And they deemed that sort of a moderate case of covid. And what they found is that roughly half of these people who are covid hospitalizations did not get supplemental oxygen and did not have an oxygen level below 94%. What they’re saying is, “If you don’t have either of those criteria, you don’t have severe or even moderate covid.”

BUCK: We’re speaking to David Zweig. He’s got a piece in The Atlantic: “Our Most Reliable Pandemic Number Losing Meaning.” I gotta ask David, are you gonna dig into the death data on covid as well? Because from the very beginning there has been a concern, a theory, however you want to frame it that there’s a portion — and I do not pretend to know what it is, but a portion — of the covid mortality that would be people who died “with covid” instead of “from covid.”

Because when we’re talking about hospitalizations, you’ve already found that there’s some major studies that are looking like maybe half of people this year who have been hospitalized with covid, they were hospitalized with, not from, right? Are you gonna look into that same distinction? Have you seen any data about that distinction when it comes to death from covid as it’s coded at hospitals across the country?

ZWEIG: Yeah, obviously that’s a really good question. I’ve talked with a number of researchers about this. And the reason why no one has really come out with something on this is that death certificates are apparently a really kind of noisy, messy, sloppy thing to look at, and obviously… So hospital admissions things are electronic records, there’s charts.

But the way death certificates are written and put together, there’s like an enormous amount of kind of discretion on who’s actually filling the thing out. There’s some people who have looked at this, and there is a pretty wide range of how the certificates are filled out. So it’s something that would be very challenging to come up with a definitive answer in the way that these hospital admissions studies have come up with a more definitive answer.

With that said, of course, this certainly begs the question (chuckles) that if you, you know, are finding this in the hospital admissions, one would assume there’s possibly at least some correlation with what are deemed covid deaths. It’s just that for whatever reason or reasons, people have not wanted to invest the resources for actually really digging into it because, from my understanding, it’s not something that can be done just using a simple sort of data analysis the way these researchers did on this hospitalization study.

CLAY: David, we talked with you about masking school children. If anything, this has just continued to be such a tempestuous and high-temperature-related debate. Have you found any more evidence that supports the idea of masking children in schools since you last talked to us and said, “Hey, there’s no real scientific evidence that this makes sense”?

ZWEIG: I have not seen anything since then that I find particularly compelling. As I probably mentioned with you guys last time, again, what I would urge everyone to do is to look toward Europe and where you see we have basically a real-world experiment taking place. For me, what I find most persuasive, more so than a study that’s done that has all sorts of noise kind of built into it and limitations. Observational evidence is pretty powerful.

And of course, countries are not apples to apples in the same way that, you know, different states or communities aren’t. But when you have a variety of countries that each of them have different vaccination rates, different case rates presently, and different mortality rates per capita — it’s all over the place — and yet there are unifying factors that none of them are masking little kids, to me, that’s compelling.

CLAY: And I just want to reinforce that because what you’re saying I feel like doesn’t get enough attention. Basically, all of the western democracies in Europe — in other words, peer countries of ours — are making the decision that masking kids in schools doesn’t make any sense?

ZWEIG: That’s correct. There are different age cutoffs. Some of them are no one’s wearing a mask, basically, all the way through the end of high school except perhaps in hallways or other certain circumstances. There’s another handful of countries where it’s up to age 12. And then there are a bunch of countries where it’s under age 6.

And that corresponds with the World Health Organization, which what they have repeatedly said, which is, “We do not want kids under age 6 wearing masks ever, period,” and they have stuck by this guidance again and again. So they haven’t wavered from it. So the United States really is an outlier in that. And if I may, it’s good that you brought this up because I think both of these things dovetail with each other.

The idea of risk to children, that some hospitalization numbers, for example, if we’re looking at them specifically for kids, that that drives a lot of the public conversation and ultimately policy. So if we have a misunderstanding as the public about risk to kids, well, that misunderstanding then, to me — you just follow the bouncing ball — then goes towards different mitigation measures that people may feel are necessary.

As a side point of whether they work or not, but just people say, “Hey, we better do anything we can.” So this isn’t to say that there aren’t some masks that do work in some circumstances. But the evidence, to me, is very, very thin that a mask mandate of little kids provides much benefit, if any, at all. And that’s what that large study talked about. So I think that these two issues are very much intertwined when we look at these sort of risk metrics and how they then lead to different mitigation policies.

BUCK: We agree — and, David, we appreciate the data and fact-based analysis and presentation for all of us. David Zweig, folks, writing a piece in The Atlantic on hospitalization metrics, and David please come back. Take a look at the death data if you would, please, and come back and talk to us about it.

ZWEIG: I’ll try.

BUCK: All right. Thanks.

ZWEIG: Thanks for having me. I appreciate it.

BUCK: Thanks so much. We appreciate it.