Abstract
The studies by Small et al. and Alland et al. (June 16 issue)1,2 substantially advance our understanding of the recent resurgence of tuberculosis in urban areas of the United States.3 However, both groups present odds ratios that appear to show an association between certain characteristics of patients with tuberculosis and the probability that their isolates were part of a cluster (a group of isolates from cases caused by identical strains of Mycobacterium tuberculosis). Both reports repeatedly refer to the patients' characteristics as risk factors for clustering or recent infection. This terminology is misleading because these characteristics do not necessarily indicate any increased or decreased risk.