You probably heard the news or read the headlines linking coffee consumption to a lower risk of prostate cancer. It’s a good study from a good group appearing in a good journal. So is it conclusive? We asked Eric Jacobs, Ph.D., American Cancer Society strategic director of pharmacoepidemiology. Here’s what he told us.
“This is a large well-designed study, but it is the only study to show a link between coffee consumption and lower risk of fatal prostate cancer. It is premature to conclude that drinking coffee might help prevent fatal prostate cancer. More studies are needed to replicate this finding. We do, however, know that both smoking and obesity are associated with higher risk of fatal prostate cancer, as well as death from many other diseases. So it is fine to enjoy a nice cup of coffee, but avoiding smoking and maintaining a healthy weight are among the surest ways to stay healthy.”
It is interesting to hear that this is the first large study to see this effect. Still, nice to know one of my favorite vices might be doing more than lifting my spirits every morning.
Meanwhile, Gary Schwitzer on his HealthNewsReview Blog gives a fascinating rundown of the study’s media coverage, highlighting the hits and misses. Gary focuses on the concept of association and causation, giving less favor to news stories that fail to point out the difference:
“We simply don’t know why more news organizations can’t do an adequate job of explaining the limitations of observational studies – most notably, that they can’t prove cause and effect.
“Yes, they can show strong associations. But they can’t prove cause and effect.”
Well now you’ve done it: you’ve got us thinking. There’s no doubt some news stories gloss over a study’s weaknesses to avoid readers asking why the study is being covered at all. And Schwitzer’s point is a critical one: association does not equal causation. The gold standard is of course a controlled trial, where half a group is given an intervention and the other half a placebo. The coffee study on the other hand, like so many others that get reported on, was observational, where researchers look at health and behavior data from a large group of people to find patterns and associations. Are those studies worth covering, or even more important, basing health recommendations on? Again, I turned to Eric Jacobs.
“It’s often said that observational studies can’t prove cause and effect. However, in the real world, observational studies alone can sometimes provide sufficient evidence to justify recommendations, particularly if an association is strong (e.g. has a large relative risk), is biologically plausible, and is consistently observed in many different studies. Some factors to consider when deciding if observational evidence justifies action were laid out by Sir Austin Bradford Hill in a classic paper in 1965.”
Dr. Jacobs points to a striking example.
“No randomized trial has ever been done to show that smoking cigarettes causes lung cancer, but strong consistent evidence from observational studies is more than enough to justify a strong recommendation against smoking.”
The bottom line: observational studies, while not perfect, have an important role in science and in public health. And Schwitzer’s absolutely right: Consumers of news deserve to know what kind of study was done, so they can better understand how strong the evidence is.