SETTING: Urban community and jail. OBJECTIVES/DESIGN: Evaluate outcome and process of an extensive tuberculosis contact investigation, including completion of treatment of latent TB infection (TLTBI). RESULTS: Between April 2000 and September 2001, 18 epidemiologically-linked tuberculosis cases were identified; 15 were culture-confirmed, all with a matching 14-band DNA fingerprint pattern. The source case had cavitary pulmonary disease and had been incarcerated 4 months prior to diagnosis. Sixty-six of 67 (99%) community contacts and 221/344 (64%) jail contacts were evaluated. The presumed new infection rate was 56% for community contacts (11 cases, 25 tuberculin skin test [TST] positive) and 20% for jail contacts (6 cases, 32 TST converters). Screening results for 113 (33%) jail contacts were obtained in the jail TST registry upon rearrest. All identified cases completed treatment. Of 22 community contacts initiating TLTBI, 11 completed (44% of infected, 50% of initiators). Of 32 infected jail contacts, 12 initiated TLTBI (all who remained incarcerated), and 10 completed (31% of infected, 83% of initiators). None of 20 additional infected jail contacts, all of whose TST conversions were identified with re-arrest data, were subsequently located. Two additional related cases have been identified as of October 2003. CONCLUSIONS: Close health department/corrections collaboration facilitated this extensive contact investigation, which identified high Mycobacterium tuberculosis transmission rates and controlled the outbreak. Numerous contacts were identified and screened, but rates of treatment completion for infected contacts were low. Novel strategies are needed to maximize the number of infected contacts who are not only identified and evaluated, but completely treated.