1/A quick note about "true" death rates (IFR) from coronavirus. Disclaimer: I am not an epidemiologist, virologist, or any other type of expert.
In fact, I think there might not be a "true" IFR, since who dies might depend on stuff like age and how people get infected.
In fact, I think there might not be a "true" IFR, since who dies might depend on stuff like age and how people get infected.
2/We know that coronavirus mortality depends on things like age, diabetes, etc.
We can adjust for those things. But there are undoubtedly OTHER factors we don't know about. And because we don't know about them, we can't adjust for them!
washingtonpost.com/health/as-offi…
We can adjust for those things. But there are undoubtedly OTHER factors we don't know about. And because we don't know about them, we can't adjust for them!
washingtonpost.com/health/as-offi…
3/Suppose that coronavirus mortality depends on some genetic factor, or on diet, or whatever. And we just don't know that yet.
No matter how many people we test, we'll see different death rates in different places because of those factors, and we won't know why.
No matter how many people we test, we'll see different death rates in different places because of those factors, and we won't know why.
4/Next, there's the possibility of dose-response effects.
In other words, if you huff a huge cloud of coronavirus, you might be a lot more likely to die than if you merely get a light whiff.
cebm.net/covid-19/sars-…
In other words, if you huff a huge cloud of coronavirus, you might be a lot more likely to die than if you merely get a light whiff.
cebm.net/covid-19/sars-…
5/If dose-response effects are big, it means that death rates will always depend on HOW people get infected.
For example, a country where mostly doctors, nurses, transit workers, and eldercare workers are getting infected would probably have a much higher death rate.
For example, a country where mostly doctors, nurses, transit workers, and eldercare workers are getting infected would probably have a much higher death rate.
6/Dose-response effects could mean that the true death rate of coronavirus depends a lot on things like:
A) population density
B) transit use and ventilation of transit
C) mask-wearing
D) multigenerational households
E) nursing home systems
...and lots of other factors!!
A) population density
B) transit use and ventilation of transit
C) mask-wearing
D) multigenerational households
E) nursing home systems
...and lots of other factors!!
7/Thus, there is probably no "true" death rate for coronavirus. Even if we test everyone, there will probably be substantial variation in death rates from place to place, due to unobservable individual characteristics and (possibly) to dose-response effects.
(end)
(end)
