This post was brought to you by Erin Kissane, Jessica Malaty Rivera, Peter Walker, and Nadia Zonis.
As the third surge of the COVID-19 pandemic continues to accelerate, we are highlighting some of the patterns in the data we have compiled from US states and territories for the past seven months. Yesterday, we wrote about the relationship between cases and hospitalizations. In this post, we’ll take a look at the relationship between hospitalizations and COVID-19 deaths.
Throughout the course of the pandemic in the United States, rises in COVID-19 cases have been followed closely by rises in COVID-19 hospitalizations. The time-lag between case surges and increases in reported deaths has been longer and less consistent—but when cases go up, deaths eventually do too. This summer, we learned how long the lag between rising cases and reported deaths can be.
After the early outbreaks centered in the Northeast this spring were brought under control, deaths began a long decline from their peak of more than 2,000 reported fatalities per day in April. Cases began to rise again in June, driven by major outbreaks in Arizona, California, Florida, and Texas. Deaths, however, continued their downward trend for weeks, and media outlets puzzled over the decline. Then, on July 6, 2020, falling death counts from the Northeast were eclipsed by rising death counts from the Sunbelt, and nationally deaths began to rise, reaching a peak of more than 1,000 reported fatalities per day in early August.
We are now a few weeks into the third surge of new cases in the United States, and deaths have begun rising once again, driven largely by the intense outbreaks underway in North Dakota, South Dakota, and Wisconsin. As cases rise across the country, we expect to see widespread increases in reported deaths as well. The higher the case numbers go, the greater number of deaths we expect to be reported in the ensuing weeks and months. As US fatality counts rise, it’s worth reiterating that the deaths reported by states and territories are almost certainly a significant undercount.
Deaths will rise, but ‘another NYC’ is far from inevitable
In the past week, US states and territories have reported all-time highs in new cases of COVID-19. The surge in cases is rightly alarming, given what we know about subsequent hospitalizations—and eventually, deaths—and given what we’ve begun to learn about the long-term health effects of the disease. Even so, some narratives about the pandemic’s progress appear to exaggerate the likeliest near-term outcomes. In particular, based on the patterns we’ve seen so far, illustrated in the above charts, we should not expect a spike in deaths that mimics the devastating cases-to-deaths ratio we saw in the spring.
The reason for this is well-known, but worth revisiting: The spring outbreaks were dramatically larger than official case counts suggest, because very little testing was performed in the pandemic’s early weeks. An estimated two million New Yorkers were infected with the virus by the end of March, while New York state reported only 75,795 confirmed cases by March 31. Because actual infections were much higher than reported case counts, the fatality rate appeared artificially high. By the second surge in cases, when US testing capacity had expanded, we saw far fewer deaths per case. It remains to be seen whether the outbreaks currently blooming across the country will conform to the patterns of the second surge. But since testing rates have remained high, we would expect this third surge to look more like the summer’s experience than the spring’s.
Hospitalizations and fatalities
A close look at the relationship between hospitalizations and fatalities can offer another clue about what we might expect to see this fall and winter.
Due to the strong day-of-week effects1 and backlogs common to reported death numbers, even the seven-day average of this metric is ragged, but it’s easy to see that hospitalizations and deaths have been strongly correlated in the pandemic so far. The data does, however, show that far fewer people who were hospitalized with COVID-19 in the summer and fall have subsequently died than was the case in the spring. The gap that opens up between the hospitalization and fatality curves illustrates one of the most encouraging pieces of news about the pandemic in the United States.
This pattern in the data we compile is borne out by a recent finding that in one large New York healthcare system, patients hospitalized with COVID-19 had a greater than 25 percent chance of dying at the beginning of the pandemic. By August, COVID-19 patients admitted to the same hospital system had a less than eight percent chance of dying. The study’s authors suggest that the improvements in fatality rates among hospitalized patients are due to several factors, including the following:
COVID-19 hospitalizations are up among younger and healthier people, more of whom are likely to survive the disease.
Clinicians and researchers have learned more about the disease and how to keep those who contract it alive, using both pharmacological and non-pharmacological treatments that modestly reduce fatalities.
Hospitals are less overtaxed than they were in the early months of the pandemic.
These improvements are worth celebrating, but it’s important to reiterate that if current (or future) US outbreaks get significantly worse, hospitals could be overwhelmed, resulting in worse outcomes for patients. We are now seeing reports from Texas and Utah that hospitals are near or over capacity, and in Wisconsin, the state has opened a field hospital to handle a surge in COVID-19 hospitalizations. Nevertheless, based on the data we have seen to date, we have reason to hope that the United States is unlikely to return to the death rates of New York City’s catastrophic spring.
1 Death reporting drops on weekends and holidays in most US jurisdictions, and “catches up” over the week, though the exact patterns differ across states and territories.
The COVID Tracking Project is a volunteer organization launched from The Atlantic and dedicated to collecting and publishing the data required to understand the COVID-19 outbreak in the United States.
More “Hospitalization and Death Data” posts
20,000 Hours of Data Entry: Why We Didn’t Automate Our Data Collection
Looking back on a year of collecting COVID-19 data, here’s a summary of the tools we automated to make our data entry smoother and why we ultimately relied on manual data collection.
A Wrap-Up: The Five Major Metrics of COVID-19 Data
As The COVID Tracking Project comes to a close, here’s a summary of how states reported data on the five major COVID-19 metrics we tracked—tests, cases, deaths, hospitalizations, and recoveries—and how reporting complexities shaped the data.
How Lagging Death Counts Muddled Our View of the COVID-19 Pandemic
During the worst parts of the COVID-19 pandemic, the United States struggled to keep up with COVID-19 death counts.