I mean everyone has some anecdotal opinion. But there is cite-able research. To bring this full circle back to HN community, I wouldn't go start a company on my opinion alone, I would do customer development, user research, etc. to make sure my assumptions are valid.
Did it say how they measured BMI? Was it the lazy height/weight way or the real way? I can't see the full study due to a pay wall and it doesn't say in the abstract. I would also be interested to see if they had any selection criteria based on lifestyle (exercise and healthy food selections) and family history since it only mentions adjustments for age, sex, and race.
Also, I would like to see the BMI by subgroup. Based on their breakdown they are comparing omnivores to vegans but leave out the vegetarian and pesectarian results.
Here's one that compares a healthy Mediterranean diet with a healthy vegetarian one. The conclusion was that they are comparable.
Does it actually help avoid obesity? Many of the vegetarians I know are overweight or even obese.