Kevin Lomangino is the managing editor of HealthNewsReview.org. He tweets as @KLomangino.
Today’s headlines on ultra-processed foods and cancer offer a good case study in the right way — and the wrong way — to frame the results of an observational study about diet and the risk of disease.
The stories are based on a French study in which more than 100,000 people were asked about their diets. Researchers looked specifically at intake of what they described as “ultra-processed foods” (e.g. sugary snacks, soda, chips, chicken nuggets) and the rates of several common cancers. They found that “a 10% increase in the proportion of ultra-processed foods in the diet was associated with a significant increase of greater than 10% in risks of overall and breast cancer.”
The researchers documented about 2,200 cases of cancer among the 105,000 participants during 5 years — a rate of approximately 2.1%. Put in absolute terms, then, a 10% greater risk might translate to 0.2 percentage point increase in the overall cancer rate — from 2.1% to about 2.3%.
A study such as this is capable of determining a statistical association between two things — but not whether one thing caused the other to happen. The tentative nature of the relationship is appropriately conveyed by headlines suggesting a “link” between certain foods and cancer — or that these foods are “tied” to cancer risk. For example:
But it would be inappropriate to suggest, as these headlines below do, that ultra-processed foods “increase the risk” of cancer or that they “raise” cancer risk.
The framing of these latter headlines suggests that processed foods caused an increased in cancer risk — a leap that goes way beyond what this preliminary study is designed to show.
In fact, if you read further down into these stories, they all will tell you that the results are not proof of cause-and-effect, as NBC does here:
The researchers stress that the study does not show that the highly processed foods caused the cancers. People who eat a lot of junk food have other habits that predispose them to cancer, not the least of which is smoking.
Which raises this question: If the study isn’t capable of showing that highly processed foods cause cancer, why does the story immediately jump to speculation on that point in its headline (“Highly processed foods may raise cancer risk, study finds”)?
It starts the discussion on the wrong foot and addresses an issue that won’t be clarified for many years — if ever.
Meanwhile, just think of all the other habits and behaviors — beyond smoking rates — that might differ between groups of people who eat lots of junk food and those who eat more healthily. This list of confounding factors would include physical activity levels, socioeconomic status, and alcohol use — to name just a few. Attempts to “adjust for” these factors may not completely eliminate their impact on the findings.
And even if we assume that the researchers have identified a cause-and-effect relationship between diet and cancer risk, that relationship may not be as simple as “ultra-processed food causes cancer,” said Yoni Freedhoff, MD, a nutrition and weight management expert at the University of Ottawa and a HealthNewsReview.org contributor. “The question that leaps to mind is whether ultra-processed foods lead to increased cancer, or whether it’s simply more ultra-processed foods leading to a decrease in the consumption of foods that are protective?” he said.
Even nutrition scientists agree that the public doesn’t trust scientific messages about nutrition.
That’s in part because of headlines that overstate the findings of observational studies, which are a mainstay of nutrition research.
Those headlines often emphasize benefits or harms of specific foods or diets, based on evidence that is incapable of proving those benefits or harms.
When new studies contradict the previous findings, a new round of contradictory headlines sows confusion and misunderstanding.
The solution is for everyone — scientists, journalists, and the public — to become more realistic and informed about what these studies can and cannot tell us about nutrition.
While it may seem like a small thing, a shift to more cautious language in news headlines may facilitate a shift to more accurate thinking about these issues.
Here are some more tips on how to write better headlines for diet studies.