Hacker News new | ask | show | jobs
by dragonwriter 2043 days ago
> Consider this: median age of COVID death is currently 78 years old. US life expectancy as of 2018 is 78.9 years. That is, approximately half the people who are dying of this would die anyway that same year.

You are confusing life expectancy at birth of 78.9 for life expectancy at age 78 of 0.9, that's not correct. Life expectancy at age 78 is 9.43 years for males, 10.98 years for females.
1 comments
m0zg 2043 days ago
Yes, and a lot of those 78.9 year old COVID patients will survive, although of course their probability of death is much higher. The point is not that. The point is it's dumb to pretend that the CFR is going to be uniform, and that just because it's 2.2% in the US at the moment, it can be linearly scaled up to the entire population (which is what would need to happen if we are to compare it to the flu, which easily 60% of people get in any given year), given the underlying demographics of fatalities. Downright dishonest, if you ask me.

Again, it is more dangerous than the flu. But nobody knows by how much, and nobody is even trying to find out as far as I can tell. Seems like a crucial question that needs to be answered, no?

link
nl 2043 days ago
> The point is it's dumb to pretend that the CFR is going to be uniform, and that just because it's 2.2% in the US at the moment, it can be linearly scaled up to the entire population

This argument makes a lot more sense than your previous one.

> nobody is even trying to find out as far as I can tell.

As someone who has friends who have been working on this for months, I can assure you this is absolutely not true. There must be tens of thousands working on it, since the people I know are members of a group approaching 100.

Two of the big problems I hear about are:

Different interventions make modelling the effective reproduction number difficult, since this gets altered so much by the interventions

Different tests have different false positive (and false negative) rates and it is really hard to find out what test is used in which jurisdiction (and even harder to find that out historically).

It's not some big conspiracy. It's more that everyone Excel files are in different formats... sigh.

link