Even with more than 500,000 dead worldwide, scientists are struggling to learn how often the virus kills. Here's why.
More than six months into the pandemic, the coronavirus has infected more than 11 million people worldwide, killing more than 525,000. But despite the increasing toll, scientists still do not have a definitive answer to one of the most fundamental questions about the virus: How deadly is it?
A firm estimate could help governments predict how many deaths would ensue if the virus spread out of control. The figure, usually called the infection fatality rate, could tell health officials what to expect as the pandemic spreads to densely populated nations like Brazil, Nigeria and India.
In even poorer countries, where lethal threats like measles and malaria are constant and where hard budget choices are routine, the number could help officials decide whether to spend more on oxygen concentrators or ventilators, or on measles shots and mosquito nets.
The question became even more complex last month, when the Centers for Disease Control and Prevention released data suggesting that for every documented infection in the United States, there were 10 other cases on average that had gone unrecorded, probably because they were very mild or asymptomatic.
If there are many more asymptomatic infections than once thought, then the virus may be less deadly than it has appeared. But even that calculation is a difficult one.
On Thursday, after the World Health Organisation held a two-day online meeting of 1,300 scientists from around the world, the agency's chief scientist, Dr. Soumya Swaminathan, said the consensus for now was that the IFR is about 0.6 per cent — which means that the risk of death is less than 1 per cent.
Although she did not note this, 0.6 per cent of the world's population is 47 million people, and 0.6 per cent of the US population is 2 million people. The virus remains a major threat.
At present, countries have very different case fatality rates, or CFRs, which measure deaths among patients known to have had Covid-19. In most cases, that number is highest in countries that have had the virus the longest.
According to data gathered by The New York Times, China had reported 90,294 cases as of Friday and 4,634 deaths, which is a CFR of 5 per cent. The United States was very close to that mark. It has had 2,811,447 cases and 129,403 deaths, about 4.6 per cent.
Those percentages are far higher rates than the 2.5 per cent death rate often ascribed to the 1918 flu pandemic. Still, it is difficult to measure fatality rates during pandemics, especially at the beginning.
In the chaos that ensues when a new virus hits a city hard, thousands of people may die and be buried without ever being tested, and certainly without them all being autopsied.
It is never entirely clear how many died of the virus and how many died of heart attacks, strokes or other ills. That has happened in both New York City and in Wuhan, China, where the outbreak began.
Normally, once the chaos has subsided, more testing is done and more mild cases are found — and because the denominator of the fraction rises, fatality rates fall. But the results are not always consistent or predictable.
Ten sizable countries, most of them in Western Europe, have tested bigger percentages of their populations than has the United States, according to Worldometer, which gathers statistics. They are Iceland, Denmark, Spain, Portugal, Belgium, Ireland, Italy, Britain, Israel and New Zealand.
But their case fatality rates vary wildly: Iceland's is less than 1 per cent, New Zealand's and Israel's are below 2 per cent. Belgium, by comparison, is at 16 per cent, and Italy and Britain at 14 per cent.
Both figures — the infection fatality rate and the case fatality rate — can differ quite a bit by country.
So far, in most countries, about 20 per cent of all confirmed Covid-19 patients become ill enough to need supplemental oxygen or even more advanced hospital care, said Dr. Janet Diaz, head of clinical care for the WHO's emergencies program.
Whether those patients survive depends on a host of factors, including age, underlying illnesses and the level of medical care available.
Death rates are expected to be lower in countries with younger populations and less obesity, which are often the poorest countries. Conversely, the figures should be higher in countries that lack oxygen tanks, ventilators and dialysis machines, and where many people live far from hospitals. Those are also often the poorest countries.
The WHO and various charities are scrambling to purchase oxygen equipment for poor and middle-income nations in which the coronavirus is spreading.
And now, new factors are being introduced into the equation. For example, new evidence that people with Type A blood are more likely to fall deathly ill could change risk calculations. Type A blood is relatively rare in West Africa and South Asia, and very rare among the Indigenous peoples of South America.
Before this past week's meeting, the WHO had no official IFR estimate, Oliver Morgan, the agency's director of health emergency information and risk assessment, said in an interview in early June.
Instead, it had relied on a mix of data sent in by member countries and by academic groups, and on a meta-analysis done in May by scientists at the University of Wollongong and James Cook University in Australia.
Those researchers looked at 267 studies in more than a dozen countries, and then chose the 25 they considered the most accurate, weighting them for accuracy and averaged the data. They concluded that the global IFR was 0.64 per cent.
The CDC relies on a "symptomatic case fatality ratio" that "is not necessarily equivalent to the number of reported deaths per reported cases." The best estimate for the United States is 0.4 per cent, according to a set of planning scenarios released in late May. The agency did not respond to requests to explain how it arrived at that figure, or why it was so much lower than the WHO's estimate. By comparison, 0.4 per cent of the United States population is 1.3 million people.
The 25 studies that the Australian researchers considered the most accurate relied on very different methodologies. One report, for example, was based on diagnostic PCR tests of all passengers and crew aboard the Diamond Princess, the cruise ship that docked in Japan after it was overcome by the coronavirus. Another study drew data from an antibody survey of 38,000 Spaniards, while another included only 1,104 Swedes.
The current WHO estimate is based on later, larger studies of how many people have antibodies in their blood; future studies may further refine the figure, Swaminathan said.
But there is "a lot of uncertainty" about how many silent and untested carriers there are, Morgan of the WHO said.
To arrive at the CDC's new estimate, researchers tested samples from 11,933 people for antibodies to the coronavirus in six US regions. New York City reported 53,803 cases by April 1, but the actual number of infections was 12 times higher — nearly 642,000, the agency estimated.
New York City's prevalence of 7 per cent in the CDC study was well below the 21 per cent estimated in a state survey in April. But that number was based on people recruited at supermarkets, and so the results may have been biased toward people out shopping during a pandemic — often the young, who have been less affected.
The global fatality rates could still change. With one or two exceptions, like Iran and Ecuador, the pandemic first struck wealthier countries in Asia, Western Europe and North America where advanced medical care was available.
Now it is spreading widely in India, Brazil, Mexico, Nigeria and other countries where millions are crowded into slums, lockdowns have been relatively brief and hospitals have few resources.
But the death rates may also shift in wealthier northern countries as winter approaches. Most of the spread of the virus in Europe and North America has taken place during mild or warm weather in the spring and summer.
Many experts fear that infections and deaths will shoot up in the fall as colder weather forces people indoors, where they are more likely to infect one another. Discipline about wearing masks and avoiding breathing on one another will be even more important then.
In each of the eight influenza pandemics to hit the United States since 1763, a relatively mild first wave — no matter what time of year it arrived — was followed by a larger, much more lethal wave a few months later, noted Michael T. Osterholm, director of the Center for Infectious Disease Research and Policy at the University of Minnesota.
More than a third of all the people killed by the Spanish flu, which lasted from March 1918 to late 1920, died in the stretch between September and December 1918 — about six months after a first, relatively mild version of what may have been the same virus broke out in western Kansas.
"We will go much higher in the next 12 to 18 months," Osterholm said. Because this is a coronavirus, not influenza, it may not follow the same pattern, but it is "a much more efficient transmitter than influenza."
Written by: Donald G. McNeil Jr.
Photographs by: victor Moriyama
© 2020 THE NEW YORK TIMES