If you joke that whatever product is being touted as a miracle for you this week will be killing you next year, you are not alone. It's become such a recurring meme that the public has very little confidence in nutritional claims.
A generation ago, bacon was bad for you because of saturated fats, then saturated fats were okay, now an International Agency for Research on Cancer (IARC) committee has declared bacon a carcinogen again, as hazardous as smoking.
Well, of course it isn't as hazardous as smoking, people now just shake their heads at Miracles Cures and Scare Journalism - unless they are Dr. Oz viewers or buying supplements from that Joe Mercola guy.
The credibility problem is that most nutrition papers are observational, notes Dr. Ross Pomeroy at Real Clear Science. The authors simply look at people who eat Food X and compare their health to those who don't eat X or eat it less. He cites American Council on Science and Health adviser Stan Young, who with Alan Karr teamed up to analyze twelve randomized clinical trials that tried to replicate the results of 52 observational studies. Most of the observational studies showed various vitamin supplements produced positive health outcomes but the robust clinical trials disagreed; 100% of the observational trials failed to replicate.
In business they say a product can be Fast, Good, and Cheap - but you only get to pick two. So it goes with nutrition papers, which end up being undermined by confounders like that they use self-reported data. Randomized, controlled trials don't fare much better because they are too short and it is hard to control for how different people react to different foods.
When studies conflict, the door is wide open for supplement hucksters, which is exactly what has happened.
Is there a solution? Not really, there is too much money capitalizing on food crazes to have anyone want to spend the money on long-term trials that will only be completed after the fad is over.