SETTINGFive state tuberculosis (TB) control programs in the United States.OBJECTIVESTo determine the prevalence of and treatment for latent TB infection (LTBI) among contacts of active TB cases identified in the workplace, and to describe TB control program policies for the initiation and conduct of workplace investigations.DESIGNRetrospective review of health department records for all culture-positive pulmonary tuberculosis cases aged > or = 15 years reported in 1996, and their contacts.RESULTSThere were 349 cases of active TB, of whom 134 (38%) were employed. Workplace contact investigations were conducted for 42 cases, resulting in the identification of 724 contacts. The rate of LTBI was 29% overall, varying by worksite from a low of 16% to a high of 51%. LTBI estimates were higher for fully-screened contacts of smear-positive rather than of smear-negative index cases. However, fully-screened contacts of index patients with cavitary disease had lower LTBI estimates than those of index patients without cavitation. Treatment for LTBI was initiated in 45% of infected contacts. The five programs had somewhat variable policies regarding workplace contact investigations. Data on HIV co-infection and place of birth of contacts were largely missing.CONCLUSIONFactors contributing to LTBI among workplace contacts may include the presence of persons with pre-existing LTBI or a positive skin test as a result of BCG vaccination, clinical characteristics of the index case, and workplace environmental characteristics conducive to transmission. Standard guidelines for workplace investigations, written workplace investigation policies, and standard data collection practices are needed to better apportion the causes of observed infection rates in the workplace.