American Excess Deaths from COVID
We did much better in 2020 and worse since 2021 than we thought.
It takes a while for David Wallace-Wells to get to his point in “What a Single Metric Tells Us About the Pandemic.” He begins with a longish setup about the changing narratives, including this roundup:
Some are familiar: The disease wasn’t spread through the air, then it was; masks weren’t worth it, early on, then became not just essential but badges of personal vigilance, then only useful if they were KN95s. Some narrative shifts were more obscure: Omicron was said to be “mild,” though it is roughly as severe as the original strain in immunologically naïve populations. Others have been somewhat memory-holed, as when much of the public-health Establishment spent the fall of 2020 suggesting that herd immunity would be reached when 60 or 70 percent of the country was infected or vaccinated, a threshold we have now long since surpassed with nothing like herd immunity in sight; or when it spent the summer of 2021 insisting that breakthrough cases were exceedingly rare and breakthrough deaths essentially nonexistent, when in fact probably a quarter of all American deaths since Delta have been among the vaccinated. Some reversals were technical, as when rapid tests were first considered imprecise, became indispensable during Omicron, then had their efficacy in preventing transmission called into question. Some had to do with policy: School closures were once part of a first-response wave of restrictions, but a growing understanding of the relatively low risk to kids and real costs of keeping them home has meant schools are now broadly viewed as among the most important places to remain open. And some had to do with personal behavior, as when many of the same people who spent 2020 yelling at Thanksgiving travelers and arguing that responsibility to protect others should dominate one’s personal behavior spent 2021 reasoning that vaccines had absolved us all of that responsibility. Many of those who once reacted in horror to “Let it rip” proponents began wondering if anything at all could have stopped the early spread in its tracks.
But he concludes this doesn’t mean what so many think it means:
Our experience of the pandemic has been littered with bad-faith argumentation and instigation, but most of these narrative reversals are not that, or even signs of what Harvard’s William Hanage has called the “motivated reasoning” of the pandemic. One narrative replacing another is one description of the scientific method, and among the many astonishing features of this pandemic is how quickly science was able to process and respond — perhaps without adequate speed, but at least fast enough for vaccines to be designed within two days, manufactured within two months, and rolled out to the vast majority of the world within two years. But the unsteady narratives of COVID-19 are reminders that, as sure as we might have been about how to interpret our experience of it, the stories we told ourselves about what we were dealing with and what we should be doing to protect ourselves were often incomplete, clouded by much more uncertainty and ignorance, wishful thinking and reflexive panic, than we were ever comfortable acknowledging.
Ultimately, then, he gets to the titular single metric: excess mortality. It’s hardly unfamiliar to anyone reading this. Indeed, the front-pagers here have written about it many times since the pandemic started. But Wallace argues that we’re at the point where this metric can help us cut through the noise.
The idea is simple: You look at the recent past to find an average for how many people die in a given country in a typical year, count the number of people who died during the pandemic years, and subtract one from the other. The basic math yields some striking results, as shown by a recent paper in The Lancet finding that 18.2 million people may have died globally from COVID, three times the official total. As skeptical epidemiologists were quick to point out, the paper employed some strange methodology — modeling excess deaths even for countries that offered actual excess-death data and often distorting what we knew to be true as a result. A remarkable excess-mortality database maintained by The Economist does not have this problem, and, like the Lancet paper, the Economist database estimates global excess mortality; it puts the figure above 20 million.
This, incidentally, seems to undercut a canard I’ve heard many times from intelligent people who aren’t generally conspiracy theorists: that the US and other governments were counting as “COVID-related” pretty much any death. Did in a car accident after testing positive? COVID-related. Apparently, not so much.
As a measure of pandemic brutality, excess mortality has its limitations — but probably fewer than the conventional data we’ve used for the last two years. That’s because it isn’t biased by testing levels — in places like the U.S. and the U.K., a much higher percentage of COVID deaths were identified as such than in places like Belarus or Djibouti, making our pandemics appear considerably worse by comparison. By measuring against a baseline of expected death, excess mortality helps account for huge differences in the age structures of different countries, some of which may have many times more mortality risk than others because their populations are much older. And to the extent that the ultimate impact of the pandemic isn’t just a story about COVID-19 but also one about our responses to it — lockdowns and unemployment, suspended medical care and higher rates of alcoholism and automobile accidents — excess mortality accounts for all that, too. In some places, like the U.S., excess-mortality figures are close to the official COVID data — among other things, a tribute to our medical surveillance systems. In other places, the numbers are so different that accounting for them entirely changes the picture of not just the experience of individual nations but the whole world, scrambling everything we think we know about who did best and who did worst, which countries were hit hardest and which managed to evade catastrophe. If you had to pick a single metric by which to measure the ultimate impact of the pandemic, excess mortality is as good as we’re probably going to get.
This reinforces a longstanding hunch of mine: that one reason that the US fares as poorly as it does in so many international comparisons is that we’re really, really good at collecting and reporting data. While the Framers got a lot of things wrong, one thing they got right, because most of them were genuine intellectuals, is instituting a Census as part of our founding document. George Washington was sworn in as President in 1789 and the first, detailed count of the population was the very next year. And we’ve done it, like clockwork—with more metrics added over time—every decade since.
So what does it say? A year ago, it seemed easy enough to divide pandemic outcomes into three groups — with Europe and the Americas performing far worse than East Asia, which appeared to have outmaneuvered the virus through public-health measures, and much of the Global South, especially sub-Saharan Africa, which looked to have been spared mostly by its relatively young population. Today, a crude count of official deaths, not excess mortality, suggests the same grouping: North America and Europe have almost identical death counts with official per capita totals eight times as high as Asia, as a whole, and 12 times as high as Africa. South America’s death toll is higher still — ten times as high as Asia and 15 times as high as Africa.
The excess-mortality data tells a different story. There is still a clear continent-by-continent pattern, but the gaps between them are much smaller, making the experiences of different parts of the world much less distinct and telling a more universal story about the devastation wrought by this once-in-a-century contagion. According to The Economist, Europe, Latin America, and North America have all registered excess deaths ranging from 270 to 370 per 100,000 inhabitants; excess mortality in Asia is estimated between 130 to 330; in Africa, the range is 79 to 220. These numbers are not identical, but, all things considered, they are remarkably close together. The highest of the low-end estimates is barely three times the lowest; the highest of the high-end estimates is not even twice as high as the lowest.
Considering that the Western populations are much older and much more vaccinated than those in Africa, that’ actually rather remarkable.
If you adjust for age, as the Economist database does separately, the differences among continents grow more dramatic — suggesting a reversal of outcomes, rather than a convergence. Outside of Oceania, Europe and North America were among the best in the world at preventing deaths among the old, and they were several times better at protecting their elderly, of whom they had many more, than Africa and South Asia. East Asia performed better, but only slightly: Canada is in line with China, Germany just marginally worse than South Korea, Iceland in the range of Japan. By almost any metric, Oceania remains an outlier: The Economist estimates zero excess deaths among the elderly in New Zealand, for instance, and gives the whole region an excess-mortality range of negative 31 to positive 37 per 100,000 residents, meaning it’s possible fewer people died there than would’ve had we never even heard of SARS-CoV-2.
This is even more startling. It was clear from the outset that President Trump contributed greatly to politicizing the pandemic, resisting a lot of early measures that could have helped contain its spread and making states with Republican governors incredibly defiant on public health matters. And yet we did roughly as well as our European counterparts when adjusted for demographics?
In the country-by-country data, the divergences grow even bigger. Perhaps most striking, given both self-flagellating American narratives about the pandemic and current events elsewhere on the globe, is that the worst-hit large country in the world was not the U.S., which registered the most official deaths of any country but ranks 47th in per capita excess mortality, or Britain, which ranks 85th, or even India, which ranks 36th. It is Russia, which has lost, The Economist estimates, between 1.2 million and 1.3 million citizens over the course of the pandemic, a mortality rate more than twice as high as the American one.
That Russia vastly under-reported its deaths and did little to mitigate them is not the least bit surprising. But that the US and UK, both of which were led by clowns during the crucial pre-vaccine period, fared reasonably well certainly is.
Russia is not an outlier. While we have heard again and again in the U.S. about the experience of the pandemic in western Europe — sometimes in admiration, sometimes to mock — it has been eastern Europe that, of any region in the world, has the ugliest excess-mortality data. This, then, is where the pandemic hit hardest — in the countries of the old Warsaw Pact and formerly of the Soviet bloc. In fact, of the ten worst-performing countries, only one is outside eastern Europe. The world’s worst pandemic, according to the data, has been in Bulgaria, followed by Serbia, North Macedonia, and Russia, then Lithuania, Bosnia, Belarus, Georgia, Romania, and Sudan. (Have you read much about pandemic policy in any of these countries?) Peru, which had what is often described as the most brutal pandemic in the world, ranks 11th — with the smallest gap, among those countries with the most devastating pandemics, between the official COVID data and the estimated excess mortality. (You probably haven’t read much about Peru, either, but its lockdowns were severe — for months, only one member of each household was allowed out once a week. At one point, an exemption was extended allowing for children under the age of 14 to leave their homes for 30 minutes of exercise per day, so long as it was conducted less than 500 meters away.)
I must admit that I was indeed unaware of the draconian measures undertaken by the Peruvians.
But there’s a plot twist!
Because The Economist allows you to explore how excess mortality evolved over time, country by country, the data also clearly showcases the pandemic as a tale of two years — a mitigation year, 2020, and a vaccination year, 2021. Early in the vaccine-distribution phase, with the U.K. and U.S. moving most quickly, it was striking how so few of the countries that had done well in preventing spread in 2020 were doing well in providing vaccines quickly. Over the course of 2021, many of those gaps disappeared, with countries across East Asia and Oceania eventually accelerating their vaccine distribution and parts of Europe that were slow at the outset starting to catch up too. But the U.S. took the opposite course. In 2020, the U.S. had done a bit worse than average among its OECD peers. In 2021, when pandemic outcomes were often determined by the relative uptake of American-made vaccines, the U.S. did much, much worse than that. In country after country in Europe, the pandemic killed a fraction as many last year as it had the year before. In the U.S., it killed more. A year ago, it was possible to defend the American record as merely below average — worse than it should have been but not, judging globally, cataclysmically bad. Today, it is cataclysmically bad, which is both outrageous and ironic, given that it is largely American vaccine innovation that has changed the pandemic landscape for the rest of the world — the rest of the rich world, at least.
So, we’re doing worse under Biden than under Trump? Yes.
On February 1, 2021, just after the inauguration of Joe Biden, the U.S. had registered, according to The Economist, 178 excess deaths per 100,000 inhabitants, quite close to Britain’s 166, Belgium’s 162, and Portugal’s 201. Fast-forward a year and those gaps have exploded. The U.S. has now registered 330 excess deaths per 100,000 — meaning our total has roughly doubled. In Britain, the excess mortality grew only 30 percent; in Portugal, it was 17 percent.
The gaps between deaths in the U.S. and countries that had done better in the first year of the pandemic, like Germany or Iceland, have gotten even bigger. If the U.S. had the same cumulative excess mortality of Germany, it would have had 600,000 fewer deaths. If it had the excess mortality of Iceland, it would have had a million fewer deaths — and would have only lost about 100,000 Americans in total.
But that’s not the right framing.
How did this happen? The answer is screamingly obvious, if also, in its way, confusing: The U.S. drove an unprecedented vaccine-innovation campaign in 2020, which empowered much of the world to turn the page on the pandemic’s deadliest phases, then, in 2021, utterly failed to take advantage of its power itself. But what is perhaps even more striking is that American vaccination coverage isn’t just bad, by the standards of its peers, but getting worse. About two-thirds of Americans have received two shots of vaccine, a level that is in line with Israel and not far off from the U.K., though below many other wealthy countries. (And even in the U.K., vaccination was more effectively directed toward the old.) But over the last six months, the country has had an opportunity to make up that gap with boosters and has simply not taken it. Only 29 percent of Americans have had a booster shot of vaccine, which puts us behind Slovenia, Slovakia, and Poland and means that less than half of those people happy to be vaccinated a year ago have chosen to get a third shot through Delta and Omicron. Booster campaigns seem like an obvious opportunity for easy public-health gains, yet remarkably few Americans seem to think it’s worth the trouble. Why? For everything we think we know about the pandemic and how people have responded to it, that one remains a maddening mystery.
The obvious answer, given my priors, is that this is a lag effect of the politicization of the vaccine. Even though Trump himself deserves significant credit for launching Operation Warp Speed, he and his minions undermined public confidence in the vaccine and underplayed the virus itself. But, of course, that doesn’t explain why we’re roughly in line with other developed countries in being “fully vaccinated” (under the original definition) but wildly lagging in getting boosted.