That's world, not USA. And the obesity inflection point was in the mid to late 70s. Total USA meat consumption since 1980 is up about 9%, with an interesting shift in distribution (as per the ourworldindata link): -11kg beef/year, -4kg pork, and +31kg poultry. It also says nothing about protein. Those numbers should suggest an increase in protein, but that's not entirely clear due to changes in the way food is consumed and the way it's measured. For instance a 10 piece set of McDonalds chicken nuggets weighs around 163 grams, but contains only 23g of protein. By contrast 163 grams of chicken breast contains just about exactly 50 grams of protein. The more processed something is, the more it tends to drift from its nominal nutritional value, and people are certainly eating more processed junk than ever before.

Observational studies on diet are completely irrelevant because they invariably end up comparing some group of people opting into some diet or another, versus the population at large. An apples to apples comparison would be something closer like comparing those on a long-term vegetarian diet to those on a long-term carnivore or keto diet. The idea is to just make sure you control for the "I care about my diet and have sustained dietary self control" variable.