Memo to the media: When it comes to reporting on nutrition and health studies, it’s important to walk on eggshells.
A recent story published by CNN exaggerated a study with questionable results. The news story claimed that eating more than two eggs a week would lead to an increase in cardiovascular diseases and a faster death. Since almost everyone eats eggs, the headline alone would cause concern.
But one sharp-eyed observer tweeted this:
.@CNN has egg on its face today. Told America eggs are killing us, when the study shows eggs have an insignificant effect, controlling for other foods. Hire reporters/editors who understand science, unless you want that #FakeNews title to stick. https://t.co/hcBPdMvjyc pic.twitter.com/HfmpzqXL9g
— Julia Pollak (@juliaonjobs) March 18, 2019
She makes a great point, which is why it’s important to read the full study (which normal people generally won’t, because they are behind paywalls). Here’s the thing: A hazard ratio below 1 indicates something is less likely to happen. A hazard ratio greater than 1 indicates something is more likely to happen. So what happens with the hazard ratio confidence interval is both above and below 1? Hint: You shouldn’t conclude anything.
That’s not the only issue we can see. Eggs seem to have the least amount of effect compared to other foods tested. From the headline of the study and the article CNN published it suggests that eggs are the ultimate evil.
Additionally the study is based off of food frequency questionnaires, which require the study’s participants to accurately recall and report their diet. It is hard to believe that 29,615 people accurately reported what they ate everyday over the course of the 31 years of data this study uses. Do you even remember what you ate for breakfast last Tuesday?
The TODAY Show had a good piece expressing caution at some of the reports. Frankly, it should be more aggressive. Food studies have a long history of making broad claims (positive or negative) based on problematic data sources. Often times, it seems as though the conclusion is written first and the study second.
Given the study’s vast range of claims, we’d say it’s past the expiration date.