Comments section, what do you think? http://www.foxnews.com/health/2017/03/08/bacon-soda-too-few-nuts-tied-to-big-portion-us-deaths.html
I'ts a so-called "meta-analysis" paper, which means they parse data from other papers and do statistics on the data. Generally speaking meta papers are trash, exercises in futility made for publishing.
Basically, researchers have to put results out frequently to keep their funders happy, but the actual breaktroughs don't happen as nearly as often as they portray it... So it means if they do a test 20 times, and 1 time they find even a breadcrump which says something "shocking", they will blow it out of proportions just to publish something to keep their funders happy