Ultramarathon DNF Fallacies

Roy Stevenson in the July/Aug 2011 issue of Marathon & Beyond magazine falls hard, alas, into the correlation-versus-causation pit. The concluding paragraph of his article "Factors Affecting the Probability of Finishing an Ultramarathon" promises:

Here, then, are the latest observations from the research on ultrarunners. Consider what changes you can make to your nutrition, training, and racing habits to reduce your chances of dropping out. It's a numbers game—if you make the changes that are positively associated with increasing your chances of finishing and of running a faster time, the odds of your ultra being successful are greatly increased.

No, No, and No! Some of the factors mentioned in the article (such as being male and under 38 years old, or never having failed to finish a prior race) can't (effectively) be changed by the runner. The reported link to taking in more calories during the event could, as Stevenson notes, be a red herring due to successful finishers going farther and dropouts being sick. People with a naturally-low body fat percentage could do better in ultras than those who diet to reduce their numbers. There's no suggestion in the reported research of causality.

One factor that is mentioned and dismissed, however, is particularly interesting: over 60% of finishers took NSAIDs (ibuprofen, etc.) and under 50% of DNF'ers did. The sample size may be too small for this to be significant, but Stevenson rejects it out of hand as a race strategy, based on potential bad side-effects of drugs. It definitely deserves more study. "Vitamin I, anyone?"

(cf. Big Stick, DidNotFinish, ForbiddenSubstances, ...) - ^z - 2011-07-07