To the Editor.
—In the articles on bicycle helmets and head injuries by Ms Thompson and colleagues,1,2 the authors state: "Our study design requires that both case and control groups have equal probability of striking their heads; that is, all cyclists must have experienced a crash." However, doesn't the probability of striking the head depend on crash circumstances? For example, while the authors believe it unlikely that there could be a 4-fold difference in risk-taking behavior between bicyclists, there could be at least a 4-fold difference in risk of head injury between riding slowly on a side street with little traffic vs, for example, riding at high speed in heavy traffic. General risk-taking behavior would also be important to consider. Furthermore, since cases were involved in more serious crashes than controls, the study does not have a comparable control group. To correct for this, the authors state they adjusted for "motor vehicle involvement."
Goldman D. Do Bicycle Helmets Protect, and Should They Be Mandatory?. JAMA. 1997;277(11):883-884. doi:10.1001/jama.1997.03540350033024