How much of societal diet-related scientific illiteracy can be blamed on the publication decisions of medical journals around food studies?
That was the question I pondered when reading “Association between kimchi consumption and obesity based on BMI and abdominal obesity in Korean adults: a cross-sectional analysis of the Health Examinees study,” recently published in BMJ Open. Although I will get to the study particulars momentarily, that it’s 2024 and journals are still publishing cross-sectional studies of the impact of a single food’s subjectively reported consumption on health outcomes is mind boggling.
You might wonder why I wasn’t mind boggled by the authors rather than the journal — but the authors’ interest in publishing a study on kimchi’s supposed impact on obesity is an easy thing to explain, in that the study was funded by the World Institute of Kimchi, where two of its four authors are employed.
You might also wonder why I wasn’t mind boggled by media running with this story — but the media’s job is to capture eyeballs, and who doesn’t love a good magic food story, doubly so for one involving obesity and one with a study backing it up?
Back to this World Institute of Kimchi project looking at kimchi intake on obesity rates. No doubt if I worked for the World Institute of Kimchi, I would want kimchi to be shown to be somehow magically protective against weight gain. So how might I go about exploring that?
Well, I could look to the data from the Health Examinees (HEXA) Study. The HEXA study was a cross-sectional look at South Koreans; included in their data collection was a 106-item food frequency questionnaire (FFQ).
That questionnaire looked at 106 food items — yep, you guessed it, explicitly including kimchi. Not included in this FFQ, though, were prepared foods, meaning that it was unable to measure seasonings, spices, or cooking oils. Also perhaps problematic is that no doubt most of us consume more than 106 total food items in our diets. Perhaps this is why the validation study of HEXA’s food item–based FFQ found that it had “relatively low validity” when compared against 12-day food diaries and why its creators themselves report it to be in their study’s conclusion only “reasonably acceptable” to apply to a population. But yes, kimchi!
So for the sake of this exercise, though, let’s assume that instead of only a reasonably acceptable FFQ with low validity, the FFQ was fantastic and its data robust. How great then is kimchi at preventing obesity? Certainly, the media report it’s pretty darn good. Here’s a smattering from the literal dozens of headlines of stories covering this paper:
Eating kimchi every day could help stave off weight gain, new study says — NBC News
Eating kimchi every day may prevent weight gain, research suggests — Sky News
Want to avoid piling on the pounds? Try kimchi for breakfast — The Telegraph
But when we turn to the paper itself, suddenly things aren’t so clear.
According to the paper, men who reported eating two to three servings of kimchi per day were found to have lower rates of obesity, whereas men who reported eating three to five servings of kimchi per day were not. But these are overlapping groups! Also found was that men consuming more than five servings of kimchi per day have higher rates of obesity. When taken together, these findings do not demonstrate a statistically significant trend of kimchi intake on obesity in men. Whereas in women, things are worse in that the more kimchi reportedly consumed, the more obesity, in a trend that did (just) reach statistical significance.
So even if we pretend the FFQs were robust enough to make conclusions about a single food’s impact on obesity, and we pretend there was a well-described, plausible mechanistic reason to believe same (there isn’t), and we pretend that this particular FFQ had better than “relatively low validity,” there is no conclusion here to be drawn about kimchi’s impact on obesity.
What we can conclude is that when it comes to publishing papers purporting to find the impact of single foods on obesity, journals will still happily publish them and their publication will lead to hyperbolic headlines and stories, which in turn reinforce the scientifically illiterate notion that the highly complex multifactorial problem of obesity boils down to simple food choices, which in turn keeps weight loss grifters everywhere in business while fueling societal weight bias.
Dr. Freedhoff is Associate Professor, Department of Family Medicine, University of Ottawa; Medical Director, Bariatric Medical Institute, Ottawa, Ontario, Canada. He disclosed ties with Bariatric Medical Institute, Constant Health, and Novo Nordisk.
A version of this article appeared on Medscape.com.