For most medical issues random control trials, involving sometimes thousands of people, are usual. But most research on food is only observational, using unreliable food diaries and then tracking the health of participants. About a million nutrition research papers have been published, but a fraction of these are good quality, randomized studies. Most are very short in timespan. The fact is that the effects of diet change take years to become obvious.
It turns out that you can take vitamin supplements and eat whole grains, fruit, vegetables and fish, and there is no reliable proof that they have any affect on mortality rates, heart attacks. cancer or anything else.
The problem is partly explained by the fact that people with low incomes tend to have unhealthy lives generally. On the other hand those who are prosperous and who eat, say, blueberries, have fewer heart attacks because they have more effective healthcare and other privileges of money. How do you judge whether health outcomes are due to the poverty or the prosperity? Do the blueberries matter?
What has happened over the years is that researchers have been able, without huge projects or even a lot of work, to find “evidence” that their favorite health food is the answer to long healthy life. Moreover the media boosts interesting findings that make headlines, without looking at the scientific evidence.
The answer lies in Epicurean moderation: eat a wide variety of foods but moderately. Then you will be reasonably (moderately?) fit.
(Written after reading an article on food by by Clare Wilson, New Scientist, July 13, 2019)