Interesting article, important issues, but it seems to conflate at least a couple of different things? Overall the theme is bad norms, but one issue is insufficient diversity in normative samples, and the other is the difference between "multivariate norms" and "univariate norms".

I'm not sure the problem is so much a myth of a universal patient as much as it is inadequately modeled variance. Maybe those are the same problem though at some level though.

>The failures of race-based medicine aren’t an argument for ignoring physiological diversity. Pretending that differences don’t exist doesn’t make them disappear; it only drives practitioners to rely on flawed intuitions. Familiar racial categories do a poor job of tracking ancestry and genetic variation. Yoruba people, in Nigeria, and Bench people, in Ethiopia, both qualify as Black, yet genetically they are further apart than an English person is from a Tamil. Instead of clinging to dubious classifications that obscure variation, we would be better served by developing methods that account for people’s distinctive ancestry and lived environment.

This just means we need more categories. Not that categorization is inherently bad.

https://archive.ph/maWBN