They did not have the bravery to control for race. You might as well throw it in the trash, it is already politically colored junk.
And no, controlling for socio-economic attributes does not fulfill the role "approximated via race."
Statistically, racial differences tend to include:
- Cultural differences, including preferences for field-of-work, family structure, socialization patterns, etc., all of which affects R0
- Economic differences (e.g. Does one need to take a Tylenol and show up to work even when sick?)
- In some cases, medical differences (e.g. vitamin D, which appears to play a major role in COVID19)
- Geographic differences (communities tend to segregate by race, and outbreaks are geographic in nature)
- Trust of institutions (which impacts vaccination, testing, and treatment rates)
... and so on.
This is about group-level statistics, and isn't helpful with individuals.
But if you're doing a population study, yes, this is an important control.
Sickle cell disease is an example of this, as it mostly appears in black people ("Sub-saharan descent").
- We adapt skin color based on latitude to maintain vitamin D levels and prevent sunburns at different levels of sun exposure.
- At this point, we have incredible strong correlations between vitamin D and COVID19 outcomes.