[stupac2]

South Korea flattened the curve by mass testing and texting people with x distance of a confirmed case every time it happened. Japan is FOS. They aren’t testing hardly at all and trying to keep its numbers down because it doesn’t want the Olympics cancelled. 



Singapore, Hong Kong, and Taiwan have done very well to stop spread in part because of complete transparency about data and also based on being prepared by learning from SARS outbreak.

Why do people believe these bogus CFR numbers?   The calculation they are using has a terrible flaw.
If Santa Clara County finds 1000 cases tomorrow, the CFR calculated this way would be 0.1%.  That's great, right!
If Santa Clara County finds 100,000 cases tomorrow, the CFR calculated this way would be 0.001%.  That's better!  Get more cases and your CFR goes down!  It doesn't matter how many eventually die!

This calculation makes the invalid assumption that all who have the disease will survive.

South Korea's death rate for cases that that have either died or discharged is greater than 30%, with no assumptions made about the outcome of those who are still fighting the disease.  Not that such a number is terribly meaningful, but the CFR numbers they are using, and being quoted, are BOGUS.

For Wuhan,  died/(died or discharged) is 7% now and slowly decreasing.  A weekly marginal value is also decreasing, approaching 1.4%.

---

Why are people fixated on the CFR number anyway?  If I told you a disease killed all and only left-handers with hands wider than 9", would you really care what the CFR was?   Look at that graph from Italy (again, badly calculated).)   If you're under a certain age, you have almost no risk of death.  If you're over a certain age and you get the disease, you have a VERY high rate of death, possibly approaching 100%.

For Italy, only 12.9% of cases have gotten to an outcome of discharge or death.  But over 13% of those 80+ that have gotten the disease have died.  If only 12.9% of their cases have gotten to an outcome, then essentially all outcomes for those over 80+ have been death.  It seems it might be pretty darn near 100% death rate for 80+, so far.

The numbers for the 70-79yo in Italy(6%) and China (8%) have suffered the same error.  Those numbers assume everyone with the disease that hasn't died will survive.   I wish it were true, but  it won't be.  Those numbers will climb.   For China, the widely-circulated, miscalculated CFR at the time was 2.3%.  Now, if you do the same miscalculation, you come up with 4.8%.  Does that suggest the 8% CFR for 70-79 should be about 16%?  I really don't have the numbers.

If the early (both China & Italy) demographics holds and the majority (China: 512 out of 1023 as of 11 Feb) of deaths occur in the 70+ age group (7% of the Chinese population), it is probably going to be really bad for seniors.  It doesn't matter how many kids get sick.

I have not found demographic numbers that give the deaths & discharges per age group for any where.

One thing this misses is that the vast majority of cases are extremely mild, and those people might never seek care (or even know they're ill), so it's not clear how they factor into the "recovered" numbers. It seems like that number is just people who have been discharged by the hospital, in which case your denominator would be too low by about a factor of 5 (going with ~20% require hospitalization).

Of course if you're just trying to calculate a CFR for the cases that require hospitalization then that would be valid, but I'm not sure why you'd want to do that when the data with age is more informative.

[stupac2]

South Korea flattened the curve by mass testing and texting people with x distance of a confirmed case every time it happened. Japan is FOS. They aren’t testing hardly at all and trying to keep its numbers down because it doesn’t want the Olympics cancelled. 



Singapore, Hong Kong, and Taiwan have done very well to stop spread in part because of complete transparency about data and also based on being prepared by learning from SARS outbreak.

Why do people believe these bogus CFR numbers?   The calculation they are using has a terrible flaw.
If Santa Clara County finds 1000 cases tomorrow, the CFR calculated this way would be 0.1%.  That's great, right!
If Santa Clara County finds 100,000 cases tomorrow, the CFR calculated this way would be 0.001%.  That's better!  Get more cases and your CFR goes down!  It doesn't matter how many eventually die!

This calculation makes the invalid assumption that all who have the disease will survive.

South Korea's death rate for cases that that have either died or discharged is greater than 30%, with no assumptions made about the outcome of those who are still fighting the disease.  Not that such a number is terribly meaningful, but the CFR numbers they are using, and being quoted, are BOGUS.

For Wuhan,  died/(died or discharged) is 7% now and slowly decreasing.  A weekly marginal value is also decreasing, approaching 1.4%.

---

Why are people fixated on the CFR number anyway?  If I told you a disease killed all and only left-handers with hands wider than 9", would you really care what the CFR was?   Look at that graph from Italy (again, badly calculated).)   If you're under a certain age, you have almost no risk of death.  If you're over a certain age and you get the disease, you have a VERY high rate of death, possibly approaching 100%.

For Italy, only 12.9% of cases have gotten to an outcome of discharge or death.  But over 13% of those 80+ that have gotten the disease have died.  If only 12.9% of their cases have gotten to an outcome, then essentially all outcomes for those over 80+ have been death.  It seems it might be pretty darn near 100% death rate for 80+, so far.

The numbers for the 70-79yo in Italy(6%) and China (8%) have suffered the same error.  Those numbers assume everyone with the disease that hasn't died will survive.   I wish it were true, but  it won't be.  Those numbers will climb.   For China, the widely-circulated, miscalculated CFR at the time was 2.3%.  Now, if you do the same miscalculation, you come up with 4.8%.  Does that suggest the 8% CFR for 70-79 should be about 16%?  I really don't have the numbers.

If the early (both China & Italy) demographics holds and the majority (China: 512 out of 1023 as of 11 Feb) of deaths occur in the 70+ age group (7% of the Chinese population), it is probably going to be really bad for seniors.  It doesn't matter how many kids get sick.

I have not found demographic numbers that give the deaths & discharges per age group for any where.

One thing this misses is that the vast majority of cases are extremely mild, and those people might never seek care (or even know they're ill), so it's not clear how they factor into the "recovered" numbers. It seems like that number is just people who have been discharged by the hospital, in which case your denominator would be too low by about a factor of 5 (going with ~20% require hospitalization).

Of course if you're just trying to calculate a CFR for the cases that require hospitalization then that would be valid, but I'm not sure why you'd want to do that when the data with age is more informative.