If you want you can skip the explanation and skip straight to the bottom, where I say exactly what I am looking for.
I was talking to my step-dad today and mentioned the recent Oxford study that shows people who do not eat meat have a 14% lower chance of getting cancer.
He suggested it could be because the vegetarians are a certain type of people, who are also less likely to smoke, more likely to exercise, etc. (A fair point, even the study itself mentions this limitation.)
I mentioned that there are many Randomised Controlled Trial studies (which should take care of his previous criticism of vegetarians being generally healthier) where they show that a plant based diet can reverse cardiovascular problems etc.
He said that even in this case, if the plant based diet in the RCT study is simply competing against an average omni diet, then it does not take much to come out ahead, because an average omni diet is full of super unhealthy food like bacon, red meat, cheese, etc. In that case, he argued, the study simply proves that going on a healthy diet is better than eating like an average person, but that you could achieve the same by going on a "healthy" omni diet with lots of fish, poultry, eggs, vegetables, fruits, greens, whole grains, etc.
What I am looking for:
So, both him and I agree, that the absolute best study would be one that satisfies the following criteria:
Randomized controlled trial study
A test group on a plant-based diet
A control group on a "healthy" omnivore diet. By "healthy" I mean the kind of diet than mainstream non-plant-based nutritionists and dieticians recommend. For example, based on whole grains, fish, vegetables, fruits leafy greens, and white meat. It does not need to be exactly that, it just need to be what most people would associate with "healthy".
The study could look at anything related to health, but preferably to avoiding or reversing serious illnesses or risk factors. I realize an RCT looking at prevention of illnesses is unlikely to exist, but I am hoping at least one on reversing may exist.
Does a study satisfying these criteria exist?
If not, what is the most credible research I could point to? Maybe like a comprehensive review, meta analysis of many different studies or something.