Why Covid-19 death rates have fallen, and why they could rise again

Doctors can now save many more lives from Covid-19 — until hospitals reach capacity.

When Theodore Iwashyna cared for his first Covid-19 patient in March, he felt a wave of terror wash over him. “I was convinced I was going to either die or get my family sick,” the pulmonologist and professor of critical care medicine at the University of Michigan recalls.

Little was known about the novel coronavirus then, and Iwashyna couldn’t stop thinking back to the last pandemic coronavirus, SARS, when more than half of all deaths in a nearby hotspot, Toronto, were hospital-acquired infections.

Nowadays, Iwashyna is less worried about himself, his family, or even his very ill patients. “I don’t expect my Covid patients to die,” he says, “I expect most of them to get through.”

By several measures, the coronavirus mortality rate has fallen over the course of the pandemic — even for those at the highest risk of death. People who are diagnosed today appear to have better odds of surviving the disease compared to the first wave of patients.

“It’s a silver lining in a hard time,” said Bilal Mateen, a clinical-academic at Kings College Hospital and an author on a new study focused on mortality in UK critical care units, “but it’s by no means an excuse to take your foot off the pedal.”

That’s because even with the advances in Covid-19 treatment, there’s still a tipping point at which the mortality rate could spike — when there are far more patients than hospital beds, staff, and ventilators. “If hospitals get overwhelmed, obviously it’s difficult to provide optimal care,” said Tom Frieden, the CEO of the health nonprofit Resolve to Save Lives.

And hospitals are becoming overwhelmed again, on both sides of the Atlantic.

France is reporting the highest number of Covid-19 hospitalizations since April — with more than half of all ICU beds full in the country. The Czech Republic — home to the fastest-growing epidemic in Europe — had to mobilize the army to build a Covid-19 field hospital and call its foreign health workers home to help with the emergency. In Belgium, even doctors diagnosed with the disease have been asked to continue working while non-urgent surgeries have been called off.

The stateside outlook is equally grim. Covid-19 hospitalizations have increased by 46 percent from a month ago, surpassing 44,000 on October 27. As Vox’s Dylan Scott reports, this has forced radical measures across the country: Wisconsin and Texas are building field hospitals, while Idaho is planning to transfer patients out of state, and Utah is ready to ration care.

So we may be approaching the tipping point at which the progress in saving lives can’t be sustained. “Each hospital’s overwhelmed point is different now than it was in April but there is a point that’s too much for any hospital,” Iwashyna says. “There are only so many hands. You can only be in so many rooms.” Here’s what we know about the falling Covid-19 mortality rate — and why it might rise again.

From spring to summer, the Covid-19 death rate dropped in Europe and the US

It’s been seven months since the World Health Organization (WHO) declared the coronavirus a pandemic, and in that period, the mortality rate among people with Covid-19 improved in countries around the world.

At first, it wasn’t clear whether this trend could be explained mostly by shifting demographics of the disease (people getting sick tended to be younger and healthier over time), more testing, or changes in medical treatments. Researchers have been trying to unpack the data — and they’ve found, even after controlling for differences in patient populations, that mortality improvements appear to hold. Just take a look at these recent papers:

• In this study, recently published in the journal Critical Care, researchers (including Mateen) in England looked at the 30-day mortality rate for 21,000 severely sick Covid-19 patients who were admitted to critical care units across England between March and June 2020. They found the odds of survival improved across the country — and the trend held even after adjusting for a patient’s underlying risks (such as age or other illnesses). In late March, 72 percent of patients who had been admitted to the HDU — a type of critical care ward where patients can be looked after more intensively — were still alive at 30 days. By the end of June, that number climbed to 93 percent. For ICU patients, the number also shot up from 58 percent to 80 percent over that period.

Critical Care

• A second paper, published in the Journal of Hospital Medicine, focused on data from a 3-hospital health system: New York University Langone in New York. This time, they looked at the March to August period and all Covid-19 hospital patients (so not just the sickest patients who wind up in critical care). They also adjusted the data to account for changes in demographics and severity of illness over time. Here, too, they found the mortality rate among patients declined — 26 percent in March to 8 percent in August.

Journal of Hospital Medicine

• In a third paper, just published in The Lancet, researchers (including Yang) took a slightly different approach: They used modeling to gauge the infection-fatality risk — or risk of death among all infected people, not just those hospitalized — in New York City between March and June. The overall risk of death decreased over time, especially for people in older age groups. For those ages 65 to 74, the estimated infection-fatality risk was 7 percent in April, and 4 percent by the end of May. For people 75 years and older, the infection-fatality risk nearly halved — from 19 percent to 11 percent.

Lancet Infectious Diseases

So, by several different measures, mortality from Covid-19 has declined throughout the pandemic.

Deaths have fallen because doctors’ understanding of Covid-19 and treating it has improved

Now the important question: What’s driving this trend? This is where things get more complicated.

Let’s start with the low-hanging fruit. In every study, researchers point to the most obvious explanation: At the start of the pandemic, we barely understood this disease. Doctors thought they were dealing with a viral pneumonia that mainly affected the lungs.

“We didn’t think about it as a whole body viral assault— and yet that’s what we have,” said Lewis Kaplan, president of the Society of Critical Care Medicine. “There’s virtually no organ system — from skin to blood vessels to brain to nose — that is not implicated in one way or another.”

With this shift in understanding has come better medical care.

In the spring, it was common to rush to use invasive mechanical ventilation — placing a tube in a patient’s throat and hooking them up to a ventilator to force air in and out the lungs, essentially doing the job of breathing on their behalf.

There were a couple of reasons for this approach: It was thought to be less risky for health care workers than other forms of oxygen therapy, because there’s a lower risk of aerosol spread. Doctors also suspected patients needed help breathing as fast as possible, explained Leora Horwitz, an NYU medical professor and author on the NYU Langone paper. “The patients oxygen levels were so low, they’d get bad within hours, and we didn’t want to risk someone dropping so much they stopped breathing before we had them on ventilators.”

Chesnot/Getty Images

But it turns out that unlike SARS, the novel coronavirus doesn’t easily spread in hospitals through aerosols when personal protective equipment — like N95 masks and face shields — is properly used. So ventilation didn’t actually reduce the risk of hospital-acquired infections. And less invasive forms of breathing support — like high flow nasal oxygen — were actually less damaging for the lungs and led to better outcomes. “We now understand we do not need to rush to put people on ventilators,” Horwitz added. Instead, doctors now try to avoid it all together.

There’s another important shift in coronavirus understanding. At the beginning of the pandemic, many doctors thought “there was something uniquely pro-inflammatory about Covid-19,” said Theodore Iwashyna, a professor of internal medicine at the University of Michigan. They pointed to cytokine storm — patients’ immune systems going into overdrive to fight the virus — as a leading cause of death, and thought drugs that suppress inflammation (such as IL-6 inhibitors) would help.

Those drugs didn’t help, and researchers learned Covid-19 “as a critical illness is a lot like other critical illnesses,” Iwashyna said. Patients with severe disease often experience acute respiratory distress syndrome (ARDS) — a deadly form of sepsis, where inflammation leads the lungs to fill with cells and fluid, eventually causing other organs to fail. Critical care doctors have long had a playbook for handling ARDS patients in ICU. “The very specific pathophysiology [of Covid-19] at a molecular level will prove to be a bit different, but our current excellent supportive care is helpful to patients without operating at that specific molecular level,” Iwashyna added.

Now, there’s strong evidence that common steroids like dexamethasone can reduce the risk of mortality in severely sick in-patients. Putting patients to rest on their stomachs instead of their backs (a practice known as proning) also seems to help.

Though there’s still a lot of progress to be made, the treatment approach has become more standardized over time, said Jen Manne-Goehler, an infectious disease doctor at Brigham and Women’s and Massachusetts General hospitals. When she started treating Covid-19 patients in the spring, it felt like practice was changing every few days. Now, it’s more streamlined — and that’s undoubtedly helping with survival, too.

“Whenever we standardize management [of patients] around a set of interventions — and we have at least some evidence [of what works],” Manne-Goehler said, “we have the potential to do a better job of managing any condition.”

The most important factor in the falling mortality rate

Better treatment approaches, however, are far from the whole story. As much as doctors have uncovered things that help Covid-19 patients, they also tried a lot of things that— at best — failed to help, and at worst, harmed patients, King College’s Mateen pointed out. (Think of early ventilation, or hydroxychloroquine.)

“The size of the change we’re seeing [in mortality] can’t be explained by any one thing,” he continued. “The mortality improvement we saw with dexamethasone doesn’t explain this 50 percent drop we see from the peak of the pandemic.”

So what else could account for the falling death rate? Perhaps the studies can’t fully control for the shift in the demographics of the disease. “Maybe those who are at greatest risk early in the course of the virus succumbed at a very high rate and now those same patients are not as widespread and represented in all our communities,” said Kaplan.

Another hypothesis is that, as masks became more common, people are exposed to less virus when they get sick. “Perhaps you get infected with fewer virus particles and maybe having a lower dose of the virus causes less severe disease,” Horwitz said.

Because of more widespread testing, of course, people are being diagnosed sooner than they were at the start of the pandemic, when there was essentially no testing, so doctors can intervene earlier in the course of the disease.

But the most important factor driving the falling mortality rate is likely one that has nothing to do with medicines, viral exposure, or improved diagnosis, the critical care doctors told me: hospitals have simply become less overwhelmed over time.

“We just can’t operate as well under real strain,” said intensive care doctor Lakshman Swamy, who works with the Cambridge Health Alliance, “and overall strain has dropped considerably over the course of the surge.”

In Boston, for example, Swamy saw his whole hospital pivot toward Covid care in the spring, with multiple new ICUs staffed by volunteer physicians. Back then, doctors, nurses and respiratory therapists were stretched every day. “ICU patients just didn’t get the same attention as they can get now, with the numbers so far down,” he said.

“This has consequences,” he added. ICU care is nuanced, labor intensive. Without adequate supplies and staffing, “Changes in condition are noticed a bit later than they would have been, interventions take a little longer to institute, and the ventilator — which requires round-the-clock attention — has longer and longer intervals before some who knows what they are doing can give it the attention it needs.”

The ominous reality of a winter surge

You can see a hint of the effect of overburdened hospitals in Mateen’s UK study. He and his coauthors chose to use the week of March 29th as the reference point for how mortality changed over time because it was the peak of hospital admissions in the UK. But if they’d gone back to the first two weeks of March instead, pre-surge in the UK, the study would have found no improvement in mortality. Mateen and his colleagues decided to start at peak occupancy precisely because of the role they figured it played in mortality rates — a finding in another related study, currently in pre-print.

The ominous implication of the winter spike in cases and hospitalizations, Mateen said, is that “we risk seeing what we saw at the beginning of the first wave, with a creeping up mortality rate.”

Horwitz also has little doubt increased hospital strain will impact care for the worse. “I think it’s very possible for mortality rates to come back up,” she warned. “If you overwhelm the health system, you overwhelm the health system.”

As the new coronavirus wave gets larger, patients in small hospitals may fare particularly poorly, a recent JAMA Internal Medicine study suggests. The researchers uncovered a large variation among the mortality rates across US hospitals, and that hospital capacity prior to Covid-19 was a major predictor of outcomes. Hospitals with the fewest ICU beds had more than three times the mortality rate compared to large-volume hospitals, even after accounting for differences in patient populations.

Large or small, hospitals that didn’t use the summer to prepare for the crush of patients already arriving in their wards also “face real risks,” Iwashyna warned.

But overwhelming hospitals with Covid-19 patients is not an inevitability, said Wan Yang, an infectious disease modeler at Columbia University and author of the Lancet coronavirus mortality study. Stopping the spread of the coronavirus requires collective action, she said. “It really depends on how people behave because it’s a disease that thrives in people. So if we reduce contacts and take preventive measures like wearing masks, we’ll be able to keep the infection rate in the population at a low level.” If we fail to do that, Yang added, she doesn’t want to think about what might happen next.

Help keep Vox free for all

Millions turn to Vox each month to understand what’s happening in the news, from the coronavirus crisis to a racial reckoning to what is, quite possibly, the most consequential presidential election of our lifetimes. Our mission has never been more vital than it is in this moment: to empower you through understanding. But our distinctive brand of explanatory journalism takes resources. Even when the economy and the news advertising market recovers, your support will be a critical part of sustaining our resource-intensive work. If you have already contributed, thank you. If you haven’t, please consider helping everyone make sense of an increasingly chaotic world: Contribute today from as little as $3.