Death Rates Pose a Puzzle

The WHO’s decision to elevate COVID-19’s estimated fatality rate to 3.4% this week fueled more confusion—but it appears to be because there aren’t as many asymptomatic cases as thought, CNBC reports
At this point, it’s just too early to say for the US, Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, told the US Congress yesterday. While the math among official cases in China, it looks like a 2% fatality rate—but that shoots up when you look at different age and risk groups, he explains.

The US doesn’t have clear enough data on infections so far to tell.

Even if the case fatality rate turns out to be 1%, Fauci said, that’s still 10 times more deadly than seasonal flu.

Comments +


scott maroney
March 7, 2020

Are the transmission rates calculated using exponential models? using A=Aoe^-kt it seems like it spreads much faster than Ebola ever did. That mathematically suggests COVID-19 spreads by means other than simply bodily fluids. the math can give clues to how the disease is speradingl. I would like to see more analysis that includes calculating death rates and rates of transmission.

Matthew E
March 7, 2020

From the chart in the lower right of the
"Coronavirus COVID-19 Global Cases by Johns Hopkins CSSE"
web page (JHU) it's clear to me why it's just not known how deadly it is.
As of 3/1, only ~40k of the ~80k then-confirmed cases had recovered; most of the other 40k were neither dead nor recovered. Since ~3.1k are already dead, the fatality rate has to be at least 3.1k/80k = 3.9% - a bit higher than the 3.4% the WHO said - and could be much higher. How high? Since we now have 59.9k recovered (per JHU graph on lower right) or ~57k (per JHU headline #s) of 80.7k cases (per JHU-both places), the fatality rate has to be lower than (80-59.9)k/80.7k = 24.9%. Probably a lot lower.
(I see this as basic math. If you think I've gone wrong, explain/prove it.)

March 10, 2020

I wish the virus tracking map had the break-down by county available again or that someone could send me the code so I could post my own state interactive map broke down to county and city levels for California. One thing is clear, some people looking at the data are not getting better or not getting better quickly at any rate, I suspect there are a lot of critical cases that will eventually die and some with organ damage that long after the infection clears will also bring about death at some later point and that will be the hidden death toll that probably won't make it into the virus statistics. How many surviving the virus that were critical cases don't end up with damage to their liver, heart or kidneys that will eventually cause their early death later? It also seems that the virus is more contagious in some countries than other. Why is it so much more infective in Italy than say the US? Is it because of cultural customs, because of better identification, testing and reporting, or because different areas are hosting different versions of the virus with differing contagion and virulence levels? Why do some countries seem to recover cases much more quickly than others? Why are all the counties not requiring mandatory enforced quarantines for anyone entering their territory from any other county at this point, it's silly to do otherwise or assume any one county less infective than any other at this point with all the asymptomatic cases running around now.

Chris Koffend
March 10, 2020

We have one population that has gone through 100% testing - only one! That is the Diamond Princess Cruise ship of roughly 3,600 people, roughly 700 of which had tested positive. Repatriated passengers (to their home countries) and those remaining in Japan were all placed in quarantines upon departure of the ship. This is a statistically significant quantity to assist in calculating death rates - especially since their entire population of the ship's inhabitants have been tested (at least once). While the ships population falls within the higher and highest risk category groups (by age), this is an extremely important and valuable database! How is this "population" being tracked? For additional infections identified during quarantine periods, for actual deaths of this population stemming from the virus? For calculating this (higher risk) population's death rate? This is the best "population" and/or "case study" that is presently available in the world - is anybody doing proper, scientific, statistical follow-up and tracking on this population? FYI - the death count on the John Hopkins site is wrong as of this morning (3/10/2020) as there are now a reported 7 deaths (not the 6 as shown on the JH tracking site) associated with the infected population of the Diamond Princess population.

Chris Koffend
March 10, 2020

Matthew E., the flaw with your math is the assumption that only those who have been tested, and have been tested as positive, is the population of infected people. What we don't know, since there is only one case in which 100% of a population has been tested (the Diamond Princess Cruise Ship), is how many are truly infected. With regards to the 100% population based testing - the cruise ship - 700 people proved to be infected. At this point (3/10/2020), 7 of those 700 people have died. This is a death rate of 1%. Of course, more may test positive from this group and more may die from this group. So that is just a preliminary number. It should also be noted that this population is largely made up of high risk and very high risk people - and yet 2 weeks into the high infection counts for this population, we are still at the 1% death rate - which is pretty telling for such a high risk group. All people who have left the ship left to go into quarantine - so these numbers should be able to be updated. Hopefully, John Hopkins or another legitimate, reliable and capable organization will continue to statistically track the Diamond Princess Cruise Ship population and results.

Bruce Bradbury
March 11, 2020

Matthew E. The 80k in the denominator of your calculation is 'confirmed cases'. You need to replace this with the number of people infected (not known, but it will be much higher). Calculations from the Diamond Princess (where most people were tested) show an infection fatality rate of around 1%.

Emma Woodrow
March 12, 2020

I am very shocked that the U.K. has decided to abandon community testing and allow the virus to run unchecked through the community with the only advice being ‘stay at home for 7 days and contact someone if you don’t seem to be getting better by then’.
This looks like deliberate sacrifice of all the most vulnerable and a total abandonment of community health.

March 15, 2020

Interestingly the South Korean experience where they’ve been testing a huge segment of the population suggests a mortality rate similar to the Diamond Princess of around 0.9-1%. Thus, in countries such as the US where testing has been minimal the numbers of cases is many factors higher than is being reported.

March 15, 2020

Death rates relative to resolved cases are considerably high for several nations.

WORLD 162687 81002 75620 46.48% 6065 3.73%

CHINA 80995 10788 67004 82.73% 3203 3.95%

ITALY 21157 17750 1966 9.29% 1441 6.81%

IRAN 13938 8624 4590 32.93% 724 5.19%

S KOREA 8162 7577 510 6.25% 75 0.92%
SPAIN 7798 6992 517 6.63% 289 3.71%

US 3244 3170 12 0.37% 62 1.91%


Anthony Murawski
March 16, 2020

Of course, the overall death rate--be it the death rate of people known to be infected, or the death rate of people who are actually infected--is meaningless in-and-of itself in determining the risk of death of death in any particular country. For instance, part of the reason Japan has had a much higher death rate than other countries is that a higher percentage of the population there was at risk to begin with due to advanced age. But that was likely not the only factor. Population density would also be a significant factor. So would the prevalence of extended family structures as opposed to atomic family structures in a society. In a country with a much higher percentage of people who are at risk of death, with a high population density (compared to another country you are comparing it with), and with far more extended families where the likelihood of susceptible people being repeatedly exposed is much higher than the country for which you are trying to estimate the death rate, you have to take those factors and other factors (e.g., what kind of mass transit is used) into account. We shouldn't be talking about a general death rate, because that statistic is meaningless when applied to populations in countries with much different demographics, and where other factors affecting transmission and risk are also considerably different.

That's why I don't understand why the Johns Hopkins Resource Center presents you with the incredibly misleading figures of confirmed infections and deaths without explaining that those two figures alone don't really tell us anything at all concerning the risks in countries with different risk factors.

David Morse
March 17, 2020

What information is known about survival ratesof individuals who need intubation / respiratory assistance ?

Mike Manglitz
March 17, 2020

It is clear from the numbers of dead and recovered per country that the fatality rate for China is a small fraction of what the rest of the world is experiencing. Why would that be? It would seem that they are not telling the truth. I read an early report which proposed that the virus was created in a Chinese lab located very near ground zero. Might they already have a vaccine and might that be why they are not reporting any new cases? Just saying.....

Post a Comment

Restricted HTML

  • Lines and paragraphs break automatically.
  • Web page addresses and email addresses turn into links automatically.
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.
Back to top