As I've been harping in every repeat topic - the way the news outlets are calculating the fatality rate is wrong. It is NOT [deaths]/[cases], which the article linked uses to reach it's conclusion. It's [deaths]/[deaths + recoveries]. You cannot include "indeterminant" cases when doing this math. Otherwise you severely undershoot the severity by operating on the implicit assumption all incomplete cases will resolve with survival. The CDC has a vested interest in reporting the naive number in order to stop people from panicking despite this number being wildly inaccurate for an infection in-progress.
If you use the
actual math and
this:
South Korea
--------------------
Total deaths = 40
Total recoveries = 135
40/[135+40] = ~23% of RESOLVED cases resulting in death so far.
This overshoots deaths marginally (a better number would include more specifics about the cases that haven't resolved yet) and the sample size is too small at the moment. This number isn't that useful.
The most accurate numbers we probably have are china:
Mainland China
---------------------
Total Deaths: 3042
Total Recoveries: 53738
3042/[3042+53738] = ~5% of resolved cases resulting in deaths so far.
To note: Using the naive estimate the Spanish Flu had a naive CFR of
>2.5%. Given these statistically significant samples we can be almost certain we are staring another pandemic in the face. It only depends how long before the CDC declares it and panics the world.
----------------------------------------------
Here is a paper:
https://academic.oup.com/aje/article/162/5/479/82647
To Quote: