Why “exercise lowers cancer risk” headlines mislead readers

The following post is by Kevin Lomangino, Managing Editor of HealthNewsReview.org, who stresses that he has nothing against exercise and believes that it is a healthy activity that should be encouraged. This post is about inaccurate news media descriptions of an exercise study and how those descriptions could be improved.

---

The big health story of the day is about a study on exercise and cancer risk. Multiple national outlets are carrying the story, with headlines claiming that exercise lowers risk for 13 different cancers.

NBC: Exercise Lowers Risk for These 13 Cancer Types

LA Times: Exercising drives down risk for 13 cancers, research shows

HealthDay: Exercise may cut risk of 13 cancers, study suggests

TIME: Exercise Can Lower Risk of Some Cancers By 20%

These headlines all loosely follow the framing of a National Institutes of Health news release about the study, “Increased physical activity associated with lower risk of 13 types of cancer.”

And yet there’s a crucial difference in the wording of the appropriately cautious news release and the more definitive-sounding news coverage. The NIH news release states merely that exercise was “associated with” lower cancer risk, whereas the news story headlines all claim that exercise “lowers/drives down/may cut” cancer risk.

It may seem like a small distinction, but it’s a critical signal as to what this observational study can and cannot tell us about the relationship between exercise and cancer risk.

STAT’s Sharon Begley, whom we profiled in a podcast last week, delved into this point in rich detail in her coverage of the study:

An observational study like this — ask volunteers how much they exercised, then see who develops cancer — can’t show that exercise caused a lower risk of cancer. It can only show that the two go together. Proving causality requires assigning some volunteers to exercise, others to be sedentary, and then comparing their cancer incidence. Only that type of study can make sure some third factor is not at work.

She noted that smoking and obesity, for example, are both linked to exercise frequency as well as to cancer risk. Could those factors be skewing the numbers to make exercise look like a bigger player than it really is?

And what about the factors that the researchers didn’t account for and we may not even know about? Notes Begley:

The study did not include genetics, which might influence both exercise and cancer risk, said statistician Rebecca Goldin, a professor of mathematics at George Mason University who was not involved in the research. “Other health factors might influence both, too. Imagine you have something simmering before cancer has reared its ugly head, which might cause you to both not exercise and develop cancer.” That would show up as the association the study found, undermining the idea that exercise helps prevent cancer.

Bottom line: The headlines claiming that exercise “lowers cancer risk” aren’t accurate, because they suggest that exercise directly impacts cancer risk in a causal way. They impart a certainty to the findings that the underlying science simply doesn’t support. (Read more about this issue in our Toolkit explainer: Observational Studies – Does The Language Fit The Evidence? – Association Versus Causation.)

Bear in mind that some stories did appropriately call attention to limitations of the study in the body of their text. For example, HealthDay helpfully noted that “the study only found an association between exercise and reduced cancer risk; it did not prove a cause-and-effect relationship.”

NBC also mentioned that “it’s always possible that people who are able to exercise more are healthier in other ways and less likely to develop cancer for some other reason.” But then it immediately shot down that caveat with this claim: “cancer experts say the evidence is very convincing that exercise directly affects the growth of tumors.”

Are all cancer experts similarly convinced? Even the researchers themselves noted in their paper that they “cannot fully exclude the possibility that diet, smoking, and other factors may affect the results.”

I understand that many reporters have no say over what goes into their headlines and that’s unfortunate. But someone needs to take responsibility for setting the appropriate tone on health news.

Even the most thorough and careful list of caveats in the body of a story is no match for an overstated headline — which may be the only part of a story that most readers see.

In fact, headlines on health stories are so important that we’ve written some tips on how to write them more accurately.

Tip #1 in that piece? “Learn the difference between association and causality.”