The Death Rate Is Only an Estimate
The mortality rate is an awfully squishy number that’s being reported as if it’s a stone-cold fact. On Tuesday, a number of headlines trumpeted that the World Health Organization was saying the death rate was 3.4%. Some hand-wringing ensued over how this number was higher than the previous estimate of 2%.
Here’s what WHO Director General Tedros Adhanom Ghebreyesus said: “Globally, about 3.4% of reported COVID-19 cases have died.” Let’s zoom in on the word “reported.” The WHO puts out a daily situation report that you can find here. It defines confirmed as “a person with laboratory confirmation of COVID-19 infection.” As of Tuesday, the total number of deaths reported globally (3,112) as a fraction of the total number of confirmed cases reported globally (90,869) was 3.4%.
Here’s the problem, though. That denominator is laboratory-confirmed cases. As we know, in the U.S., it’s pretty hard to get tested right now. In fact, based on this definition, as of Wednesday night, the U.S. mortality rate based on CDC numbers — 9 reported deaths and 80 laboratory-confirmed cases — was 11%. You know that’s bogus. You know that’s because there’s not enough data, the denominator is pitifully small and we need to be testing a whole lot more people.
Over the last few weeks, many more countries have realized that the coronavirus has hit their shores. Some, like South Korea, are doing tons of testing and generating lots of data. Others, like the U.S., aren’t, as ProPublica has reported. The rate will also depend, country by country, on demographics (this virus is more deadly to the elderly) and resources (like ventilators). It’s not surprising that the global mortality rate based on confirmed cases might fluctuate for a while.
When most people talk about fatality rates, they’re thinking: If I get this, will I die? The only way to actually answer that question is to know how many people have been infected, and for now, that’s nearly impossible. As Marc Lipsitch, an infectious disease epidemiologist at Harvard’s T.H. Chan School of Public Health points out, deaths are the most obvious and easy thing to catch, whereas infected people who stay at home and those with no symptoms are incredibly hard to account for. That tends to skew the fatality rate higher, especially earlier on in an epidemic.
What we do know for now is that it’s more deadly than the seasonal flu, which generally kills far fewer than 1% of those infected, and less deadly than a disease like SARS, which killed about 10% of those infected during the outbreak in 2002-3.
When I write about the mortality rate, I try to use caveats like “estimated” or “scientists understand it to be around” so readers understand it’s not fixed in stone.
Instead of saying: The mortality rate is X%.
Say this: Scientists estimate the mortality rate is X%, based on the information they have.
Be Careful with Projections
Another slippery number out there is what’s known as the basic reproduction number, R~0~ (pronounced R-naught). It’s a measure of contagion, the average number of people who will catch the disease from a single infected person. For similar reasons as above, this number is currently a moving target, as more data is gathered from around the world. So far, estimates have largely been in the range of 2 to 3.
What this means for reporters is that if someone tries to say something like, there’s going to be X number of cases by a certain date, that can’t be a hard and fast number. I’d want to know what assumptions were used to calculate that forecast. What was the R~0~ presumed? How about the serial interval, the duration between the onset of symptoms between one case and its secondary cases? Tweaking either of those numbers by just a bit can result in very different forecasts, which you can see by playing around with this interactive tool by the University of Toronto. Generally, I shy away from putting a projection in a headline, where any hope of nuance might be lost, but if I have to, a range is safer than a single number that readers might interpret as somehow immutable.
Furthermore, as of early March, there are many fundamental questions about the novel coronavirus that scientists still don’t fully understand. For example, while it’s clear that the primary method of transmission is via droplets, drops of fluid from the mouth or nose emitted when an infected person coughs or sneezes, it’s not clear if it can transmit as an aerosol, meaning it is airborne and floats around (this is considered to be unlikely). It’s also not conclusive if the virus can be spread by infected people before they present any symptoms.
Instead of asking: How many cases will there be at X point in time?
Ask this: What assumptions were used to calculate your prediction? What’s the upper and lower range of your projection?
Information Is Changing Quickly and May Soon Be Out of Date
One last thing I’d like to add: Even more so than usual, things are moving quickly. I’ve been on interviews where the information I was given was outdated — as in just plain wrong — by the time I filed my draft 12 hours later. This is, of course, terrifying as a reporter. So I’m trying my best to put information like “as of Wednesday morning” alongside facts and figures in my stories, and I’m encouraging my sources to update me as often as they can.
OK, but How Do I Protect Myself?
Over the last two days, I’ve gotten numerous DMs over Twitter from concerned members of the public, asking me what they should do to be safe. Honestly, this breaks my heart and speaks to a failure of local health officials to educate them. I’m having the same conversations over and over again, so I thought I’d share some of my thoughts here. I’m not a medical professional, so this is not medical advice.
Start by knowing yourself. Are you elderly or immunocompromised? Young and healthy? Your risk varies depending on your personal profile. If you’re concerned about your health, I encourage you to talk through your fears with your doctor. I’m 29; I know there’s little chance that this virus would kill me given the information I’ve seen. (In data published last month by the Chinese CDC, out of more than 72,000 diagnosed cases, 8.1% were 20-somethings, and the fatality rate in that age bracket was 0.2%.) That said, given my personal medical history and tendency to get bronchitis, I would really prefer not to get infected.
So how does that translate into action? Here have been my personal choices so far. I’m still flying; I just got off a plane to attend a reporting conference in New Orleans. (I would not attend a conference in the Seattle area, however, given how signs are pointing to widespread community transmission.) I don’t see how being on a plane increases my personal risk any more than being on the New York City subway. That said, I am not shaking any hands at this conference, and I’m ramping up my hygiene game: washing my hands more frequently and encouraging my colleagues to do so as well.
I’m aware of the possibility that I may need to work from home in the near future, if I or my husband get sick, or if there’s an explosion of cases in New York City and social distancing measures are encouraged. So we are slowly but methodically picking up a little bit of extra food with every grocery run (for our two cats as well!), just so that we’d have enough at home if we need to be indoors for a few weeks. I’m not panicked, nor should you be. I’d encourage you to check on your neighbors — especially the older ones, or those with young children, and see if you can pick up some additional groceries for them.
Even if we have to stand a little ****her apart from one another, the best way to get through this is with a bit of extra compassion to bridge the gap.
If you have expertise or tips you’d like to share with me and members of my reporting team, please fill out this form or email us at
coronavirus@propublica.org.