Right of Way

305 Broadway, Room 402

New York, NY 10007


Drinking and Bicycling - A Reality Check

Recently, efforts to promote bicycling as a healthful activity compatible with ordinary Americans' life-styles were set back by a spate of news stories claiming that imbibing just one drink increased a cyclist's risk of a serious-injury crash more than five-fold.

The stories, which ran in U.S. News & World Report, Reuters, and The New York Times, among others, were prompted by a Feb. 21 article in the Journal of the American Medical Association (JAMA). That article has now been rebutted by Right Of Way member Charles Komanoff in a letter published in the May 16, 2001 issue. The authors of the original article have accepted Komanoff's criticisms and, for all practical purposes, have withdrawn their original inflammatory conclusion.

In his letter, Komanoff, a statistician and policy analyst, reveals that the JAMA authors conflated the injury rates for cyclists with a wide range of blood alcohol levels, "ranging from a single drink to extreme intoxication." They then used this lumped-together data set to derive a five-fold increased risk of injury, and misleadingly implied that even one drink induces this level of increased risk. It was that "finding" that provoked the flurry of media attention and bolstered the received wisdom of blaming cyclists for being hit by cars.

As Komanoff suggests, the proper approach would have been to exhibit the curve of increased risk with increased consumption (technically, the "dosage-response") - rather than lumping together everyone from the mildly buzzed to the totally hammered and computing a meaningless composite risk for these very different cases.

Responding to Komanoff, the JAMA authors graciously concede the validity of Komanoff's criticism and point out that their sample was too small to derive a reliable dosage response. They acknowledge, however, that as far as their data go, the increased risk of "moderate" blood alcohol (in the range 0.02-0.07, representing one to three drinks) is only 40% - a far cry from the fivefold increased risk which made their original article so newsworthy.

Here's a point-by-point comparison of media scaremongering vs. the facts.

Michael Smith

May 17, 2001



Cars Suck! -- research

Right Of Way home page