>This assumes that social disparities can’t be studied empirically, which is not true.

I agree with this, but the explanation that racism did this is the scientific equivalent of god did it. It's a solution that gets you nowhere. A much better study would attempt to pinpoint the exact causes (air pollution, smoking rates, higher risk jobs, etc) which actually gets us somewhere. You can recognize that all these might be second order effects of racism, but we still need to know the exact cause and effect relationships in order to be able to translate that into real world actions. For example, segmenting people by both race and city/rural groups lets us study the effect of air pollution on outcomes...if that turns out to be relevant then that changes the way we think about the relationship between highway construction and high-density housing which can actually lead to an improvement.