The evidence against “ultra-processed” foods is weaker than you think

Figuring out the healthiest kinds of foods to eat can feel unexpectedly hard. Is it better to look for organic or non-GMO? Foods with added protein or fiber, or foods that have been least-processed? Labels seem like they should be helping us with all of this, but as Vox senior reporter Marina Bolotnikova finds in her latest piece, at least one of those labels has been adding to the misunderstanding: ultra-processed. 

According to an in-depth analysis of recent studies on ultra-processed foods, it seems the way we’ve been studying them has largely been flawed. Marina dives into the details and nuances of the tricky field of nutrition research and finds that the answers and labels aren't always so clear. And that in some cases even single-ingredient, demonstrably healthful foods — like tofu! — have been mis-categorized as ultra-processed. 

Having written about nutrition science for years myself, I walked away from Marina’s new article with a better understanding of how to think about these food categories — and the high stakes for health nationally as the federal government continues to push for limits on foods that have gotten this dubious label. 

This kind of writing is made possible by Vox Members. If you would like to support Vox’s independent journalism, please consider becoming a Vox Member today.

—Katherine Courage, deputy editor Future Perfect

The evidence against “ultra-processed” foods is weaker than you think

New analysis suggests other factors may explain the harms blamed on “ultra-processing.”

by Marina Bolotnikova

Why Americans are living longer again

by Bryan Walsh

The US death rate just hit a record low, and there’s reason to hope the trend will continue.

Medicare is now offering $50 GLP-1 prescriptions. The US may never be the same.

by Dylan Scott

Three ways GLP-1s are about to change America.

Why air conditioning became a cross-Atlantic culture war

by Bryan Walsh

Air conditioning is not a moral failure.

Read more >>>