The Third Wave of the Pandemic Is Here

A medical worker pushes a stretcher through a hallway at Mount Sinai Hospital in Manhattan on the first day of fall. Photo: Spencer Platt/Getty Images

When Donald Trump checked into Walter Reed medical center more than a week ago, it appears likely to have marked the beginning of the end stage of his presidency. But it was also a milestone for the pandemic, and not just because COVID-19 had infected its most prolific and prominent skeptic and dissembler. In recent weeks, a third wave of the coronavirus has come to the U.S. at almost precisely the time of year scientists warned us about in the spring. But the country has hardly noticed, so paralyzed and preoccupied by the spectacle of the presidential campaign it could barely acknowledge any new cases but Trump’s. There were nearly 50,000 new U.S. infections reported on the day the president was hospitalized, along with 835 new deaths. That’s two 747 crashes’ worth.

When the country passed 100,000 deaths, a spectacularly bleak edition of the New York Times marked the occasion with a six-column headline for a flood of obituaries that ran the full length of the front page (and onto several additional pages). When the toll passed 200,000, it did not even mark the tragic landmark on A1. They are running out of hospital beds in Wisconsin — which used to qualify as a battleground state, incidentally — and in North Dakota, which hasn’t imposed a mask mandate, they are down to 39 open ICU spots. But while the pandemic does indeed appear to be getting worse almost everywhere in the country, it also seems unlikely to return to the center stage of America’s attention until after Election Day — at which point perhaps 25,000 more Americans might have died.

But things won’t really change immediately after November 3, either. The apparent collapse of last-minute stimulus negotiations means that our sclerotic Congress won’t likely extrude any meaningful pandemic relief until January 20. There also won’t be a national testing program erected, or a federal contact-tracing system belatedly instituted, or, probably, a vaccine or novel therapeutics in wide distribution before the next presidential inauguration, either. At which point there might be 100,000 more American deaths than there are today, each a tragedy unfolding amid a considerably uglier humanitarian catastrophe — poverty and hunger, evictions and loss of health insurance, mass joblessness without commensurate federal support — than the pandemic has produced to this point. In other words, the third wave will likely be worse, nationally, than the first; much less buffered by political action and support, at least on the federal level; and, as long as the election eclipses the full attention of the news media, many times less salient. We’ve already tuned it out, and nothing is likely to help anytime soon.

Not very long ago, the pandemic response in Europe appeared to shame the United States. “Cafe society returns to Paris,” the Guardian declared as America’s Sun Belt second wave drove a summer spike from 20,000 new daily cases to 60,000. The day that article was published, the seven-day rolling average of U.S. deaths was 731; in France it was 32. Two months later, in mid-August, America’s seven-day average was at a then-distressing 1,008 cases, while the French figure had fallen to an enviably low 12. France is, yes, a much smaller country. Adjusted for population, the American outbreak, right then, was more than 50 times worse.

But as fortunes turned here — by mid-September, new American cases were down almost half from their peak — they shifted in the other direction over there. In France, where the seven-day rolling average of new cases had been as low as 530 in mid-July, it has grown to 12,000. In Spain, the average went from 250 to 11,000 new cases. In the U.K., where the average grew from 575 to 11,000 new cases, the growth overwhelmed the country’s rickety database, housed on an Excel spreadsheet that literally ran out of rows. And now, with America’s daily caseloads spiking, Europe’s recent experience looks not like an alternate path or even a cautionary tale but a grim forecast for what could transpire here. Already, according to Covid Exit Strategy, 26 American states are currently experiencing “uncontrolled spread.” Another 17 are “trending poorly.” That’s 43 of 50 states. Five of the remaining seven qualify as “caution warranted,” and only two states — Maine and Vermont — qualify as “trending better.” Two states out of 50. However you look at it, a third wave of the pandemic is here.

But “wave” isn’t really the most precise term, and not just because the disease proceeds erratically across the country, with some communities in the West and Mountain West in the grip of a terrifying first encounter with COVID-19 while others are breathing easier and feeling as though they are many months past a peak. It’s because neither of the first two waves ever really crashed, only crested. This makes the third phase even more concerning — recent growth in caseloads, and deaths, comes on top of a distressingly high baseline of spreading sickness — roughly 40,000 new cases and 800 new deaths per day. Three weeks ago, Dr. Anthony Fauci told James Hamblin of The Atlantic that “we must, over the next few weeks, get that baseline of infections down to 10,000 per day, or even much less if we want to maintain control of this outbreak.” Three weeks later, it reached 50,000 — five times the upper end of Fauci’s “safe” range. Lamenting, this week, the maddening lack of a national testing strategy nine months after the coronavirus first arrived in America, immunologist Rick Bright of the NIH wrote, “the country is flying blind into what could be the darkest winter in modern history.”

When the coronavirus first arrived, in the spring, there was much hand-wringing and anxiety about a second wave to come in the fall or winter, as had happened with the 1918 flu — which killed five times more Americans in the winter than it had in the summer. And while experts warned in the spring not to count on a seasonal suppression of the disease in the summertime, their own data often suggested that the disease was in fact probably suppressed somewhat in those months, thanks in part to temperature and humidity effects that are now running in the opposite direction. That many of them are now warning we are underestimating coronavirus seasonality is not among the most conspicuous public-health reversals of the pandemic — reversals on mask wearing, asymptomatic spread, fomite and aerosol transmission, and the safety of the outdoors were all probably more consequential. But having it both ways on seasonality — emphasizing its trivial impact during the summer and its significant impact in the fall and winter in order to produce heightened vigilance in both instances — may well have muddied the public’s understanding of the disease. And possibly, in so doing, made the winter pandemic potentially worse.

“Absolutely, we did hear that a lot,” said Harvard epidemiologist Michael Mina, when I asked him about the summertime admonition to not count on a seasonal decline. “I think that people have become very confused about it all.”

Mina is an assistant professor at Harvard’s Center for Communicable Disease Dynamics and has made a significant name for himself during the pandemic as one of the most clear-sighted advocates of true mass testing, arguing back in the early spring that by far the easiest way to get disease spread under control was to test many millions of Americans very regularly. The fall and winter, he said, may already have gotten away from us. “We still have about 40,000 or 50,000 cases a day right now, and we haven’t really gone too far below that this summer,” he said. “But the fact that transmission can continue and persisted during the summer should not be misconstrued to mean that this is not a seasonal virus. It just means that maybe pretty soon we’re going to have 150,000 cases a day.” In this, Mina is echoing the warnings of other experts. The University of Minnesota’s Michael Osterholm told Meet the Press, “There’s a really hard road ahead of us,” and told STAT News, “I think November, December, January, February are going to be tough months in this country without a vaccine.” Harvard’s Caroline Buckee has compared the approach of winter to dark clouds on the horizon.

“A lot of that comes from just looking at how seasonal viruses like coronavirus generally transmit,” Mina said. “And frankly, they usually go to near zero during the summer months. And I would say that the fact that transmission has continued during those months, despite the fact that this is a seasonal virus; and the fact that normally, you’d expect the virus would go to near zero in the summer to really large numbers, even exponential growth in November, December — that doesn’t bode very well for us.”

By the time the virus was really circulating throughout the country, he said, we had already left the winter for the spring and summer, which means very few places, if any, experienced the pandemic at anything like its natural seasonal peak. In parts of the country where the virus has receded, he said, we’ve had a tendency to attribute that to our behavior — mask-wearing, social distancing, testing — “but I’m not convinced that it really is only from our behaviors. I think that we have probably benefited a tremendous amount just from the natural course of this virus.”

So what’s possible going forward? Mina said it was “very likely” we will exceed the spring peak, when, at one point, 2,500 people were dying from COVID-19 each day. But he also acknowledged there wasn’t yet clear evidence for that outcome and added that in a few places — the Northeast in particular — there remains a concerted focus on mask-wearing and social distancing that could mitigate such a surge, along with some amount of limited herd immunity or community protection from earlier exposure. “But I think many of the parts of the country that aren’t necessarily taking it quite so seriously run a real risk, a very high likelihood, of having substantially more cases.” How substantial? I asked. “Maybe ten times more than they’ve seen so far,” he said. “We’re not sure how bad it’s going to get this winter, but I think there’s a very good likelihood that it will be much worse.”

Warnings like these are, of course, speculative, given that the coronavirus is such a young disease it hasn’t even lived through a single fall yet — which means we can’t really know how it interacts with that weather when it does. (And given that the data on temperature effects of the disease gathered from around the world this year are somewhat muddied, with some studies showing little effect and others showing much more significant ones.) When I recently asked Micaela Martinez, an infectious-disease ecologist among the world’s leading experts on seasonality, whether we were seeing the beginning of such an effect, she cautioned that seasonality may only be playing a small role in the recent bad-news turn. True recurrent seasonality, which returns year after year, only comes when the disease is endemic, she said — that is, not in its initial spread through the population but in subsequent years or eras, once it has fully penetrated a population and can only infect new susceptibles (young kids, new arrivals, those whose immunity has waned). In the meantime, she explained, while there may well be some effect of seasonality on transmission and possibly disease severity, that effect would likely be dwarfed by others: how many susceptible people remained in a given community and our current interventions (masking, social distancing, and closures).

The matter of the remaining number of susceptibles is a tricky one. Early in the pandemic, the conventional rule of thumb was that at least 60 percent of a community would have to be exposed to the disease, and possibly 80 percent, before the community as a whole acquired a kind of collective immune protection against the disease — a phenomenon sometimes called “herd immunity” and sometimes “community protection.” But in the summer, as transmission rates seemed to decline in many places almost independent of policy interventions, a growing number of scientists and modelers began to wonder whether those thresholds were high or perhaps way too high. Many conjectured that caseloads and deaths were improving in part because of better testing and in part because of social behaviors like mask-wearing and social distancing, but also because enough people had gotten the disease, in certain places at least, that the virus was having a harder time finding susceptible victims, naturally slowing its spread. Stronger versions of this argument, often less credited by epidemiologists and virologists, suggested that the pandemic was entirely over in many of these places.

The herd-immunity argument has been renewed, in recent weeks, by the Great Barrington Declaration — a petition of epidemiologists and public-health officials, led by Harvard’s Martin Kuldoroff, Oxford’s Sunetra Gupta, and Stanford’s Jay Bhattacharya, expressing their concern that pandemic shutdowns were unwise and their belief that a less restrictive, more focused approach to managing the disease spread might be preferable. (A similar critique was made in August in The Wall Street Journal by Greg Ip, and even Brown’s Ashish Jha, among the more responsible guardians of coronavirus conventional wisdom, has lately worried that shutdowns were problematic.)

Unfortunately, as the summer has turned to fall, the herd-immunity hypothesis has gotten much less persuasive because in many of those places the disease trajectory has gotten much worse. That is true at the country level, though countries are a crude measure. Many of the European nations hit hardest in the spring have seen dramatic growth in cases in the late summer and fall. One pre-publication paper posted last month, surveying a number of non-European countries whose disease trajectories suggested they had reached herd immunity, calculated likely exposure rates in each ranging from 67 percent in Ethiopia to 80 percent in Madagascar. If the threshold of herd immunity was much lower than those crude 60-to-80 percent estimates offered at the outset of the pandemic, the disease wouldn’t have been able to spread that much before disappearing.

Particular cities offer better case studies, since they are more genuinely single communities, and there, too, the data is discouraging. In Spain, for instance, the places that had been hit hardest in the spring had relatively worse experiences with a second wave than those that had been spared. At the even more local level, some especially hard-hit parts of New York City are among the most concerning of the new fall clusters. And while there are still some signs of enduring community protection — Sweden, long vilified for an implicit herd-immunity approach, is now among the safest places in Europe, with total per capita deaths lower than the U.S. — a bet on community protection now looks much less safe than it might have a few months ago. Through the spring and summer, Belgium was, in terms of per capita deaths, the worst-hit country in Europe; it just registered about 8,000 new cases, the equivalent of 260,000 in the U.S.

Those looking for good news at the outset of the third wave do have something to point to: the lethality of the disease. Thanks to some combination of the age distribution of cases, improved treatments and better understanding of the disease, more vigilance in protecting the country’s most vulnerable, and more widespread mask-wearing, which can reduce the viral load of any exposure and thereby perhaps the risk of infection, the COVID-19 fatality rate appears much lower than it was in the spring. While the real fatality rate is a matter of some dispute and contestation, in August, Youyang Gu, then the pandemic’s most accurate modeler, calculated that it had fallen to 30 percent of its first-wave peak.

When I spoke to Gu in September, shortly before he discontinued his forecast, citing exhaustion and the improved quality of other projections, he was relatively downbeat looking forward into the fall. While the disease’s estimated infection fatality rate remains dramatically lower than it was in the spring, he said he believed that was largely due to the age distribution of cases and, less significantly, breakthroughs in treatment. It was likely just a matter of time, he said, before the rate creeps back up again.

Harvard’s Mina agrees, calling the possibility that the fall and winter could make the coronavirus not just more infectious but more lethal “likely” and citing a variety of possible explanations: that our epithelium is dryer in the winter; that we produce less mucus; that the air both indoors and outdoors is dryer, then, too; and that there may be a seasonal effect on the viral load as well, meaning the body would produce more virus in the winter than in the summer, making it both more easily transmitted and more dangerous. The science of these dynamics in other diseases is not all that well established, he cautioned, but he suggested that it may very well be the case that many diseases we think of as “seasonal” are in fact year-round diseases that only get severe enough that we notice them in wintertime.

For his part, Gu doesn’t believe the American fatality rate will return to its spring peak, when treatment was confused and the disease much less well understood. But he thinks it’s quite likely that, due to seasonal effects, the rate will grow higher this winter than it was this past summer. And it wouldn’t have to get much higher at all to be catastrophically destructive, given the volume of cases we have today (probably about 4 million, nationally and growing). While Gu describes a recent estimate from the University of Washington’s IHME, which used a large seasonal effect to project 400,000 deaths in the U.S. by the end of this year, as “just not possible,” he believes total Americans deaths could reach 300,000 by January, with a few additional months of winter still yet to come.

Along with NYU economist Paul Romer, with whom he has collaborated, Mina is perhaps the most prominent advocate of mass testing as a path to disease suppression. But, given where we are on testing and what the country cases are in the fall, even he is losing faith. “I wish I could say I felt like we were at a point with rapid testing that it would have much of an effect at all,” he said. “But I don’t think it will, unfortunately. I started talking about this back in June, and we haven’t really seen much movement.”

Even the seemingly good recent news about the FDA authorizing the use of rapid testing, he says, has limited significance. “I would like to see the government not just wait for [diagnostics companies like] Abbott or Roche or whoever to come out with a new rapid test but for the government actually to put $50 billion — which is frankly a drop in the bucket when it comes to the economic losses that we’ve seen — into the development and deployment of massive numbers of these rapid tests. But we’ve seen this government largely just stand and sit idly by and wait until a company like Abbott says, ‘Okay, we have, we can make 30 million a month per year.’” He laughed. “That’s 1 million a day for a country of 330 million. That’s not a lot. But the average person doesn’t get that.” Early in the summer, Romer argued that the country needed to be testing 30 million people a day to allow life to return to normal; even more conservative mass-testing estimates put the figure at 30 million a week — four times as many as Abbott can today produce.

“When the president gets up in front of his microphone and says, ‘We’re going to get you 150 million tests,’ most people think, ‘Oh, he’s really doing something,’” Mina said. “But that’s not tomorrow. That’s not every day. That’s 150 million tests between now and probably January or February. When they say things like, ‘This is going to help us open up the schools, and every teacher is going to get a test’ — it can’t just be one test. That’s not the plan. The plan has to be for every teacher to get a test twice a week. One test once does nothing,” he said.

“I just keep thinking about this epidemic — on our soil, in our country. And, like, what would the government do if we had 200,000 people die from bombs being dropped on us? You know, we would not be sitting idly by saying, ‘Oh, I wonder when Lockheed Martin is going to come out with a new bomb, and we’ll buy it whenever they come out with it, it might be next year.’ No, they would be doing everything in their power to push forth new technology to actually build it themselves.”

“We’re seven months into this,” Mina continued, “Why do we still not have a game plan? I mean, it’s just astounding. It’s just remarkable. We are barely in a better position from a testing perspective than we were in May. And we have no surveillance set up for most of the country. What are we doing?”

When I mentioned that I’d written a column in early April called “There Is No Plan for the End of the Coronavirus Crisis” and another, a month later, called “There Is Still No Plan,” Mina responded, “You should just keep writing that. Just take your exact article from April and literally just publish it tomorrow, with the same headline, and not change any words.” He paused. “I just can’t understand what the hell we’re doing.”

One answer: We’re simply waiting until January, and for tens of thousands more Americans to die, before even beginning the project of national pandemic response we should have launched fully a year before.

The Third Wave of the Pandemic Is Here