In the spring, during the first COVID-19 surge in the United States, the rising death toll reached a sobering peak in April—a seven-day average of 2,116 daily deaths. This past weekend, the seven-day average of U.S. deaths from COVID-19 broke that record twice, at 2,123 on Saturday and 2,171 yesterday, according to the COVID Tracking Project at The Atlantic.
Yesterday, the seven-day averages for all four of the primary metrics that the COVID Tracking Project follows—tests, cases, hospitalizations, and deaths—were at record highs. But deaths offer the clearest comparison with the spring surge, because in those early weeks many more cases were going uncounted while testing was slow to ramp up. If the seven-day average of deaths remains above the spring record in the weeks to come, it will soon be inarguable that the pandemic winter is worse than the novel coronavirus’s first surge.
And every indication is that this surge will continue to worsen for some time, because of the other milestones the U.S. has passed in recent days: 100,000 hospitalizations for the first time, the first consecutive days of more than 2,500 deaths (three, in fact), the first day of more than 200,000 new cases (which was followed by two more days above this threshold).
Another foreboding sign is how bad conditions are across the country. From the beginning of November through yesterday, there were more than 100 COVID-19 deaths per 1 million people in the Northeast, the South, and the Midwest (which, at 267 deaths per million, had the highest rate in that period), according to COVID Tracking Project data. The West had 94 deaths per million people. In April, during the first surge, only two regions, the Northeast (602 deaths per million) and the Midwest (138 deaths per million), were above that 100-deaths-per-million line; the West was at just 50 deaths per million. In July, only the South exceeded 100 deaths per million.
In this surge, then, the share of deaths is more evenly distributed across the country. In the same period, starting in November, the Midwest represented 37 percent of U.S. deaths; the South, 35 percent; the West, 15 percent; the Northeast, 14 percent. In April, 60 percent of deaths occurred in the Northeast; no other region accounted for more than 17 percent. Since then, no region has seen anything like the conditions once seen in the Northeast, but seven months later, the picture in the country as a whole could be worse.
The seven-day average of deaths is crucially important because the daily death toll is a noisy number. We know from experience that the daily toll falls, substantially, after weekends and holidays because the people counting get a merciful break. Any high daily death toll is worrisome—no matter what, it represents a lot of recent deaths—but the average gives a better indication of the rise and fall of the toll across the country.
While this new record is chilling, and perhaps boosted by numbers coming in after the Thanksgiving delay, it’s not unexpected. The unprecedented hospitalization numbers, which doubled over the course of November, made it all but inevitable. The proportion of cases that end in deaths—the case-fatality rate—is much lower now than it was in the spring, but progress on lowering that number stalled in August, as the epidemiologist Trevor Bedford found last month.
Worse yet, the country could lose what progress it has made on the case-fatality rate since the spring and see it increase, which would mean that people could die who, under other circumstances, might not have. Improvements in the outcomes of COVID-19 cases came in part from doctors and nurses learning how better to identify and treat the disease’s most serious symptoms, which can vary widely from patient to patient. Health-care workers cannot provide that level of care when hospitals are overwhelmed by the most serious COVID-19 cases.
As Robinson Meyer and Alexis Madrigal have reported, hospitals have now hit a breaking point at which they no longer have the capacity to treat COVID-19 patients that, not long ago, might have been hospitalized. One indication, observed by Ashish Jha, the dean of the Brown University School of Public Health, is that in recent weeks, about 3.5 percent of cases translated into hospitalizations a week later. But that number has started falling. The University of Nebraska Medical Center—which started preparing for such a scenario in 2003, making it perhaps the nation’s best-prepared hospital—was nearing its breaking point at the end of November, Ed Yong found.
The last three weeks of the year will tell us the magnitude of the situation. Thanksgiving caused a data lag, and the seven-day average of deaths could likely increase in the next few days, in part because medical examiners will finish catching up on their work. At the same time, we may start seeing the effect of Thanksgiving travel and gatherings on case counts, the surge many have been fearing. By Christmas and into the New Year, we’ll have a sense of how new cases from Thanksgiving will translate into deaths.
Around that same time, vaccines will start to roll out—in small numbers at first, but hopefully making a substantial difference in long-term care facilities such as nursing homes, which have borne the brunt of the pandemic—as we see the worst of this catastrophe and the beginning of its end simultaneously.