To the Editor.
—The article by Yesalis et al1 is a timely, descriptive report of AAS use among young adults in the United States. The cross-sectional study design estimates prevalence of AAS use, which is simply the number of persons who ever used AASs divided by the number of persons sampled. The sampling method was complex; however, that is not an issue. Technically, the number derived is a proportion because the numerator is a subset of the denominator.2,3 The authors report in Table 1 of their article the rate of AAS use as a percentage.The use of rate in this context is troubling on three counts: (1) other researchers, wishing to compare prevalence of AAS use from similar surveys, may decide this comparison is inappropriate because of the use of "rate" in the table and the text; (2) rate implies person-time in the denominator, and it would
Sigurdson A, Hoque A. Use of Anabolic-Androgenic Steroids: We Are Talking Prevalence Rates. JAMA. 1994;271(5):347-348. doi:10.1001/jama.1994.03510290029027