Seems like a good chance to gather data (perhaps yearly) on a few hundred or more people who generally stick to this 'fad' (I'm kidding) diet. Even though macronutrient ratios are argued about, most people generally steer clear of gluten, legumes, milk, and refined sugars.
Perhaps a basic height/weight might be a good place to start but maybe include lipid panel numbers or something?
I really think there should be a lipid panel database where you'd enter values for various tests and it would have checkboxes that would elucidate which particular ancestral diet permutation you're following. Someone could then run some statistical analysis on it and see what sort of correlations exist. It wouldn't be "scientific" but it would likely yield a good deal of data.