ACSH would like to induct a recent study seeking to link increased red meat consumption to a higher risk of stroke in women into our very own Data Dredging Hall of Fame/Shame. In an analysis of 34,670 Swedish women between the ages of 39 and 73, researchers found that those in the top tenth for red meat consumption — eating about 3.6 ounces daily — had a 42 percent higher chance of suffering from an ischemic stroke (from a blocked artery) than women who ate just under an ounce of red meat daily.
If you’re just as confused by all of these numbers as we are, then fear not because ACSH's Dr. Gilbert Ross has a simple explanation: “This study is an observational study that relies on dietary recall, which is a notoriously unreliable parameter. Further, the ‘42 percent increased risk,’ while making for a good headline zealously exploited in simplistic reporting, makes me wish yet again that epidemiologists would be a little more honest so that unless they observe a relative risk in such a study of at least double, perhaps they should not bother reporting it. As long as data dredging like this is published by allegedly respectable journals, we can’t blame reporters for jumping on it.”
The study, like so many lacking proper controls, presents an array of problems with regard to confounding variables and the problem of separating cause and effect.
The media has capitalized on the research, however, because “it fits in with the popular wisdom that red meat is bad, and the less of it you eat, the healthier you are,” says ACSH’s Dr. Elizabeth Whelan. “Other retrospective analyses have demonstrated that red meat actually has no effect on stroke risk, yet they get thrown down the epidemiology garbage chute because it doesn’t create nearly as big of a media splash.”
