Thread on deducing the death rate of 2019-nCoV: OK, I think I get it now, sorry that I have been so dense. China is relating the number of confirmed cases instead of the best estimates, which relates more to the testing capacity in the crisis region than to the incidence.
-
-
Journalists reporting on these numbers without noticing that can be explained as basic arithmetic illiteracy (NYT) and domesticated monkey behavior (Buzzfeed), but how can the WHO blunder like that?
Show this thread -
Which brings me to the next inconsistency: the ratio of dead to confirmed diagnosed has been uncannily steadily at 2.1+/–.1% for many days in a row now. If you really go and fully fake the number of deaths, why make it so obvious?!
Show this thread -
It is clear that China, the WHO and the CDC must have statistically sound estimates of the number of estimated cases and the mortality rate (which you can probably determine by tracking a cohort of early infected cases). Why don’t they say so? They may all have good reasons.
Show this thread -
China (as perhaps our own governments as well) does not consider it prudent to report the full extent of China‘s present and our future humanitarian crisis. But at least they are letting us know the mortality estimate. It’s the pattern encoded into two meaningless numbers: 2.1%.
Show this thread -
If the disease is mainly killing people with diminished lung capacity, bear in mind that China has more smokers and severe air pollution, so the rate may be much lower in the US. All of the above are speculative thoughts. Don’t believe anything I say.
Show this thread
End of conversation
New conversation -
-
-
Where did the nine days come from? Or is it just hypothetical?
-
The latter, to give a sense of a terrible upper bound. But if 25% of all cases are in serious condition, I would not think that more than half of those (12.5%) die when receiving proper care.
- 1 more reply
New conversation -
-
-
This ought to be obvious! One other thing to look at might be the ratio of deaths to confirmed recoveries (though this has been gradually improving from 1:1 to almost 1:4 now) https://gisanddata.maps.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6 …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I had been thinking something very similar about the confirmed diagnosed deaths to confirmed diagnosed unwell ratio with a delay depending on how long it takes until someone dies, but you’ve expressed it far more clearly than my half formed idea
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.