While long term low dose impacts are very difficult to track resulting in this surveying challenge, there are scientific alternatives, like measuring blood glucose or insulin.

This is how we know that sugar and refined carbs are bad. The author does a huge disservice by claiming everything is a correlation study