Drinking five cups of coffee a day could help women live longer.
This television segment, which reports on findings about the health effects of coffee, is flawed at the core.
The published study itself in the Annals of Internal Medicine concludes that coffee does not appear related to increased risk of death due to heart disease, cancer or any cause. It suggests that higher levels of consumption by healthy people may have cardioprotective benefits–but that this study does not have the power to determine that.
The study also reports that these links appear largely independent of caffeine consumption–that is, something in the coffee other than the caffeine may be at work.
Yet the broadcast frames the news as if the results say five cups of daily coffee per day may help women live longer. Much of the discussion focuses on caffeine, rather than coffee–in direct contradiction to the study’s findings.
This report is an excellent example of how not to report on an observational study. The interpretations made are both incorrect and incomplete.
This is a powerful example of how excessive inference–driven largely by a desire to make the news broadcast itself more compelling–can mislead. It also exasperates them with the institutions of medical research and health journalism, since this dramatic, apparently conclusive report will inevitably be overturned by the next dramatic, apparently conclusive report.
While the broadcast does follow some best practices which create context and provide caveats–including the shopworn incantation of "moderation, moderation, moderation–none of it can overcome the exaggerated, inaccurate framing of the story.
n.b. The Annals of Internal Medicine publishes a plain-language "Summary for Patients" for this study, which does an excellent job of translating the results for lay audience.
Given the context, it’s not necessary to report the price of coffee.
Having said that, an imaginative reporter or host might have mentioned, as an aside, that people who drink 4-6 cups of coffee per day are spending $1 to $9 per day–a good reason not to take up coffee drinking for medicinal purposes.
The segment leads with an aggressive overinterpretation of a small finding of the study–the observation that women who drink five cups per day could have reduced risk of death by heart disease. It uses a very specific figure–a reduction of risk of 25 percent in women–without providing any context. Twenty-five percent of what underlying risk? What is the absolute risk?
The key finding of the study, however, is that coffee appears not to increase the risk of death. The research also suggests that coffee may have some protective benefits for heart disease, but this needs to be explored with further research. The story incorrectly suggests a causal relationship that has not been demonstrated.
The segment responsibly points out that excessive caffeine consumption is linked to elevated blood pressure, anxiety, shakes, even heart arrythmias.
It gets extra points for mentioning the differences in risk between those who metabolize caffeine slowly or quickly.
The report itself is an observational study based on self-reported data gathered over a long period of time. The segment properly discusses, at some length, the limitations of this kind of study.
Having said that, there’s a legitimate question about whether the study itself, due to its many limitations, provides evidence of sufficient quality to justify such a high-visibility report.
Even if the answer to that question is yes, the conclusions offered at the top of the piece are not justified by the evidence in question.
Further, the study shows a potential inverse relationship between coffee and cardiovascular death. But the discussion repeatedly focuses on caffeine, though the study suggests the coffee itself, not the caffeine, may be responsible for the possible link.
Again, the inital comments in the story are incorrect: Drinking five cups of coffee may not help women live longer. It’s another case of a journalist treating causation and correlation as synonymous terms. See a past Publisher’s Note on this topic.
So: Despite the segment’s responsible discussion of the study’s limits, it earns an "Unsatisfactory"rating under this criterion.
The host’s overstatement of the results at the top of the segment–that "drinking five cups a day could help women live longer"–generates both false hope and false anxiety.
Its overplay of the effect on women exaggerates a small point to dramatize the results and (we suspect) attempt to make this story more compelling to the program’s core audience.
The only source used is an internist at the New York University School of Medicine who came in to the studio "on short notice." There’s no evidence he is an expert on coffee or this area of research.
Since the segment’s claim is a lower risk of death, especially from heart disease, it should have at least glancingly referred to the tried-and-true treatments for risk reduction such as weight control, a healthy diet, moderate physical activity and drugs such as statins and aspirin if indicated.
A best practice would indicate how much impact coffee is likely to have–even with the most generous interpretation of these findings–relative to other methods of reducing heart risk and death.
The availability of coffee is not in question here.
Coffee’s ubiquity is implied and its novelty is not an issue.
The story does not draw excessively on the journal’s press release.
But maybe it should have. The press release parses the conclusions and data far more carefullly than this broadcast segment does.