Those who do are rare, as they would need to last at least a few decades. IIRC to publish actual research on diet effect on lifespan you need 80 years.

What you do is develop metrics useful to predict lifespan and quality of life and then see the impact of a diet of those metrics. There is nothing you cannot do privately here (using some laboratory for analysis obviously)

I am not saying it is easy and for sure it is error prone, but so is science itself.