Author Affiliation: Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland (email@example.com).
To the Editor: Dr Freedman and colleagues1 reported that the relative risks for tobacco smoking in association with incident bladder cancer increased in the last 20 years and that the population attributable risk (PAR) was stable to increasing despite lower prevalence of smoking. They postulated that this finding may reflect a change in composition of cigarettes. However, they presented PARs for ever (rather than current) smoking, and former smoking accounted for 39% (in women) to 57% (in men) of person-time. Therefore, former smoking accounted for a majority of the ever-smoking PAR in both sexes. The PAR is best interpreted as the proportion of disease that could be averted if all individuals were changed from one risk category to another; the use of this measure to characterize a risk category that cannot be changed (ie, former smoking) is questionable.2
Dowdy D. Tobacco Smoking and Bladder Cancer. JAMA. 2011;306(20):2216-2217. doi:10.1001/jama.2011.1717