[function_seven]

The first paragraph says:

> The researchers had estimated the rates of daily current smokers among COVID-19-infected patients and compared them to the rates of daily current smokers within the general French population, after controlling the data for sex and age.

So, assuming they correctly controlled for age, then no, it’s not “anti-survivorship” bias. (“Casualty bias”?)

[ipsum2]

There could still be survivorship bias, e.g. the people who don't die from smoking have stronger lungs, and aren't affected as much by covid.

[kolinko]

This would be true if the outcome of smoking was binary and quick (i.e. unless you have strong lungs you die from smoking soon, otherwise your lungs remain stronger than the rest of population until natural death).

That’s not a pattern with smoking.

[brianjunyinchan]

Ah, missed that, thanks. Even more interesting then!