A paper published in the Journal of Clinical Oncology, “Coffee Intake, Recurrence, and Mortality in Stage III Colon Cancer,” is receiving a lot of news media attention, and much of it is incomplete on key points. The researchers concluded: “Higher coffee intake may be associated with significantly reduced cancer recurrence and death in patients with stage III colon cancer.”
But association ≠ causation, a point made clearly by the researchers – and a point made repeatedly by us when we see misleading news coverage about observational studies.
The paper includes several paragraphs of discussion about what the findings may or may not mean and the researchers wrote:
“our observational study does not offer conclusive evidence for causality.”
That fact was missed or misunderstood by many journalists, who reported such things as the following:
The New York Times Well blog’s headline was, “Daily Coffee May Boost Colon Cancer Survival.” Yes, it may, but, because of the lack of conclusive evidence for cause-and-effect, it also may not. But at least the Times, in its second paragraph, included this:
“But, the researchers caution, cancer patients should not start ordering extra tall coffees. The study, the first to report such findings, does not prove a cause-and-effect relationship between coffee drinking and a lower risk of colon cancer recurrence. As other experts note, there may be differences between heavy coffee drinkers and abstainers that the research was not able to account for.”
But that was only after putting in the first paragraph that “significant benefits” started “at two to three cups a day.”
What is a reader supposed to make of this? A specific statement about benefits in paragraph one, down to the level of how many cups render that benefit. Followed by a clear statement about no cause-and-effect proof in paragraph two. Overall, it wasn’t a bad article. But it would be understandable if readers’ heads were spinning after just the first two paragraphs. And that was unnecessary. So even when a news source starts to get it right, they may get it wrong over an inaccurate word here or there.
But the words – all of them – matter.
Again, we refer readers to our primer, “Observational Studies – Does The Language Fit The Evidence? – Association Versus Causation.”