In part, according to one of the head researchers, “compared to prior studies that relied solely on epidemiological data, we integrated multiple layers of information, including epidemiological data, conventional metabolic biomarkers, and cutting-edge metabolomics”.

So in addition to finding the metadata, they also dug into the biomarkers present, which the molecular pathways are already known in detail. On top of this, metabolomics interpreted the chemical processes involving metabolites, the small molecule substrates, intermediates, and products of cell metabolism. Together, these aspects helped demonstrate the molecular consequences of ingesting animal products. The epidemiological layer assessed the incidence rate of developing T2D and then investigated how many involved the consumption of animal products for the individuals.

This being the case, it appears to be more than just correlation, as once the metadata was found, the team investigated the causation of the metadata's findings.

You don't think the Department of Nutrition at Harvard would've thought to try and control for this?

Not on sci-hub yet but here's a similar one:

https://jn.nutrition.org/article/S0022-3166(22)02864-4/fulltext#secsect0020

This is another reason to favor Beyond over Impossible! Additionally, Beyond has always been gluten-free while Impossible made their recipe gluten-free years down the road.

Would imagine that their main goal is to make it easier for the average current meat eater consumer to actually try non-meat options. And is their version of heme the same level of problem as actual meat based heme?

Sounds like an obscure garage tool to me.