Tuesday, August 12, 2008

Running can ward off death, says a news study focused on the effect of exercise in the elderly. Yay running!

[Of course, as a professional researcher I must note that I doubt the study showed any such thing. Instead, it probably found a correlation between time spent running and decreased morbidity. Maybe the people predisposed to spend time running when they get older are also predisposed to enjoy good health longer than average -- even if for some reason they don't actually run.]

[The popular press frequently conflates correlation and causation when reporting on medical studies. I suppose it's a misguided attempt to make the "conclusions" of the article seem easier to understand.]

[I couldn't find study itself, after a brief search. If anyone knows where it is, please give a pointer in the comments -- then we'll be able to test my claims above.]
Posted by at
Labels:

6 comments:

tentrillion said...

I found the original study. In short, as usual, I think that popular reports of the study grossly exaggerate the findings.

The study tracked two populations over a 20 year period -- one of runners, and one of non-runners. After the study period the researchers checked for statistical differences in common measures of health between the two groups. They found several interesting things, including that a higher percnetage of the non-running group had died, and that the running group as a whole enjoyed a better (lower) average Health Assessment Questionnaire Disability Index.

What's the problem with the popular reports then? Well, first, the study identified its running population by sending mailers to running clubs. Maybe the regular social contact of being in a running club is responsible for some of the health benefits. If so, then popular headlines might read "Having friends slows the effect of aging" instead of "Running slows the effect of aging".

Of course, both versions of the headlines are probably misleading. The verb "slows" strongly implies causation. But the study used only statistical inference. Without a experimentally validated biophysical model of how running lessens morbidity and increases life span, any claims as to causation are totally unfounded. See how hard it is to use strong verbs?

You might think that my concern here is overblown, but in general I think that these kinds of popular reports distort the public perception of how well we understand our bodies and our health.

Let me point out two dangers arising from inaccurate public perceptions of medical uncertainty. First, misplaced public confidence in our medical "knowledge" may diminish public support for further research, and also drive patients into a blind-faith acceptance of their doctors' medical advice.

Second, over the long term, new studies often overturn the findings of earlier, more preliminary studies. When contradictory findings are publicized, it unjustifiably erodes the credibility in the medical science enterprise as a whole. If the popular reports of the initial studies had emphasized the caveats and uncertainties, we could much better appreciate the new findings as a development instead of a reversal.

Anyway, I still like running. In fact I think it's highly likely to be extremely beneficial to one's health. It's just that this study didn't prove it like all the news reports suggest that it did.

8:17 AM
Jennifer said...

wow. this post is like main topic, ps, pps, pps, then in the comment area, PPPS.

5:35 PM
tentrillion said...

You mean you read the whole thing?

8:36 PM
Jennifer said...

Isn't reading the whole thing the point of procrastination-by-visiting-blogs-of-friends?

8:39 AM
tentrillion said...

I guess so. So, were you convinced? Posting your own thoughts is an entirely other avenue of procrastination that I urge you to enjoy.

1:45 PM
phyllis said...

dude! this post is awesome! i love ripping apart popular reports of scientific studies. :D feel free to dissect other studies related to exercise.

and yeah, i totally agree re: the post and re: procrastination by reading blogs of friends. it's a testament to how busy i've been that i haven't procrastinate-read this until now.

4:44 PM

Post a Comment

Subscribe to: Post Comments (Atom)