Why The True Fatality Rate Of COVID-19 Is Hard To Estimate NPR's Mary Louise Kelly talks with Natalie Dean, an assistant professor of biostatistics at the University of Florida, about the real fatality rates of COVID-19 — and why estimates vary.
NPR

# Why The True Fatality Rate Of COVID-19 Is Hard To Estimate

#### Why The True Fatality Rate Of COVID-19 Is Hard To Estimate

• `<iframe src="https://www.npr.org/player/embed/844562935/844562936" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">`
• Transcript

NPR's Mary Louise Kelly talks with Natalie Dean, an assistant professor of biostatistics at the University of Florida, about the real fatality rates of COVID-19 — and why estimates vary.

MARY LOUISE KELLY, HOST:

The numbers are frankly overwhelming. Every day the coronavirus death count rises. At first, based on early numbers in late February, epidemiologists and experts estimated about 1% of infected people would die from COVID-19, but it turns out it is hard to come up with one number that accurately captures how deadly this virus is. It is killing people in different states and different countries at different rates. According to Johns Hopkins, the mortality rate here in the U.S. is 5.7% - 5.7. In Italy, it's over 13%. In China, it's 5.5% - which made us wonder. Why is the true death rate so all over the place and so hard to pin down? Well, Natalie Dean, assistant professor of biostatistics at the University of Florida, is here to help us try to answer that question.

Natalie Dean, welcome.

NATALIE DEAN: Thanks for having me.

KELLY: So how do you calculate - how do you define what a true death rate looks like with regards to the coronavirus?

DEAN: So the numbers that you just reported, we call those the observed or crude case fatality rates. And those are calculated as the number of deaths divided by the number of cases that have actually been confirmed. So that's very important because there are a lot of cases that are not confirmed. People who have mild illness might be turned away or not be able to access testing. And we also know that there are a lot of people who are infected but don't develop symptoms, so that means that there's actually a much bigger denominator than what is reflected in this observed case fatality rate.

KELLY: So you're saying testing plays a crucial role in understanding how many people have been sickened and - have are now infected with COVID-19, and without knowing those numbers, it's really hard to get what the death rate is.

DEAN: Right. We know that testing tends to focus on people who are most severely ill. And so it's missing a lot of those mild cases, and it's definitely missing anyone who doesn't have symptoms. So we know that those numbers are overestimates, and they also vary a lot by country because there'll be different availability of testing in different countries.

KELLY: Where is the death rate highest? Where is it lowest so far?

DEAN: So looking at the statistics from Johns Hopkins University, you can see Belgium right now seems to be reporting the highest crude case fatality rate at a little over 15% - similarly high in France, the United Kingdom and Italy. We know that it's quite low in some Asian countries like Singapore and South Korea. And certainly, it's reflecting who's being tested, but then it also reflects things like quality of care in those countries.

KELLY: You've done research on Ebola, on Zika. Has - how does the challenge of identifying the true death rate for the coronavirus compare to - with others?

DEAN: So the coronavirus is definitely different from Ebola and Zika. Ebola is a very severe disease, and so we have pretty few of these asymptomatic or mild infections. That makes it easier to detect people who are sick and then track them to see who does. Zika is the other end of the spectrum, where most infections are pretty mild or asymptomatic. So then it becomes really hard to establish severity. Coronavirus is somewhere in the middle. Somewhere likely below 1% percent of people who are infected die, but we're trying to nail down exactly where that number lies.

KELLY: And why is this number important? Why do scientists need to devote resources to figuring it out?

DEAN: The number impacts how we determine our response strategies. We're very interested in understanding risk factors for severe illness - so whether that differs by sex or the presence of other diseases, race, ethnicity - and then that will allow us to identify groups that we want to pay closest attention to.

KELLY: Natalie Dean - she's assistant professor of biostatistics at the University of Florida.

Thank you very much.

DEAN: Thanks. I really appreciate it.