What is the real risk of death from COVID-19? Is it actually worse than the flu?

Biology/Immunity Infection and Spread Uncertainty and Misinformation

A: It’s complicated…. but YES!

As a demographer and epidemiologist, mortality rates are in my wheelhouse so to speak, but getting a handle on COVID-19 has challenged us all.

First some definitions:

Case Fatality Rate (CFR): % of confirmed cases who die.

In the U.S., the estimated CFR is currently 4.33%
In the U.K., the estimated CFR is currently 15.51%

(Data source: Our World in Data)

Since this number only includes CONFIRMED cases in the denominator, it is affected by the amount of testing and by asymptomatic or mild cases that never get tested. If only the most severe cases are tested, the CFR will be an overestimate of the real risk of dying while infected.

Infection Fatality Rate (IFR): % of ALL COVID-19 infections who die.

This is the number WE WANT TO KNOW. If infected, how likely is it that someone will die?

Obviously, the catch here is that to calculate the IFR you need accurate counts of ALL CASES and ALL DEATHS. Getting that DENOMINATOR right is the huge challenge right now. How many people have actually been infected?

One strategy that can help us get at that are “serology” tests—testing the population for antibodies or evidence of previous COVID-19 infection. As the Nerdy Girls have previously discussed, antibody tests are currently not great for *individual* knowledge of previous infection (lots of false positives), but at a population level it can give us a decent range of the % of people infected (Read more in The Atlantic).

So far, not many countries have what we call “population representative” seroprevalence studies to estimate the % of the population exposed. We have seen a variety of “volunteer” based seroprevalence or convenience samples such as blood donors. Population random sampling is a Nerdy Girl lesson for another day, but the bottom line is that people who donate blood or are otherwise not chosen randomly may be different from the general population, and thus it is not clear that the % infected in those populations reflects the entire country.

Spain has done the best *nationally representative* seroprevalence study thus far, released this week in the Lancet. They tested over 60,000 individuals with TWO DIFFERENT types of antibody tests to minimize false positives.

They estimated that around 5% of the Spanish population has been infected with SARS-CoV-2, with higher levels (up to 14%) in hot spot areas.

So what does this mean for the Infection Fatality Rate (IFR)?

The population of Spain is around 47 million, so 5% would be 2.35 million infections.

With 28,396 confirmed deaths, this gives in IFR of : 28,396/2,350,000….


Using the “excess mortality” number from Spain (the number of people who died during this period above and beyond the averages from past years), of 48399, the IFR is *2.1%.*

Looking at New York City, where one estimate is about 20% of the population infected (but from a non-random sample), gives a similar IFR of *1.1%*

This is more than 10X WORSE that a bad seasonal flu season (0.1% estimated IFR).

Moreover, only 8% of the population typically gets the flu each year, because of vaccination and protection from previous exposures. COVID-19 has the potential to reach a MUCH larger fraction of the population.

Finally, it is important to remember that for many infections but especially COVID-19, a single IFR is not that informative. This is because there are HUGE differences in the CFR by age (and also underlying health conditions).

In the Spain example for instance, the IFR ranged from *12%* for those 80+ to 0.20% for those 50-59.

In NYC, the estimates (assuming 20% seroprevalence) are:

8.3% for ages 75+
3.3% for ages 65-74
1% for ages 45-65
0.01% for ages 18-44.

This is one reason why single CFRs and IFRs can also change over time, depending on the ages of people getting infected.

BOTTOM LINE: While there are still many “known unknowns” in measuring COVID-19 mortality, our best estimates suggest this is WAY worse than the flu…like at least 10-20X worse.

AND as we continue to emphasize, death is not the only metric we should care about, due to the potential long-term health effects for those infected and hospitalized.

We KNOW all of this disruption and sacrifice totally SUCKS, but be assured that your efforts are not for nothing, or to avoid a simple cold or flu.

Your everyday actions are saving lives, and we can continue to do this TOGETHER!!

#MaskUp #StaySMART


Useful mortality links:

Live Science

The Economist

Our World in Data

NYC Health

Link to original FB post