Contact investigations in congregate settings, New York City.

SETTING Large urban tuberculosis control program. OBJECTIVE To evaluate results of procedures implemented for systematic investigation of tuberculosis exposures in congregate settings. DESIGN Between October 1995 and December 2000, a unit consisting of epidemiologists, health educators and tuberculin screening staff investigated exposures in sites with >15 persons. Transmission at the site was defined as likely, possible, unlikely or unknown. RESULTS Among 100 investigations, 12 were tuberculosis case clusters, five were source case investigations, and 83 were exposures to single infectious cases. Transmission was likely in 24 (21%), possible in eight (8%), unlikely in 62 (62%), and could not be assessed in four (4%). Among the 83 exposures to single infectious cases, 2740 contacts were tested; 502 (18%) were infected. Among 1202 close contacts, 996 (82%) were tested, 197 (20%) were infected and started treatment of latent tuberculosis infection (LTBI) and 102/197 (52%) completed treatment. Sites with likely transmission had index patients with longer duration of cough (13 vs. 6 weeks, P = 0.01) and cavitary lesions (84% vs. 44%, P = 0.01) compared to sites with unlikely transmission. CONCLUSION A systematic approach for conducting contact investigations in congregate settings is useful for assessing transmission. As such investigations are resource intensive and transmission is not common, performing tuberculin skin testing after most persons would have converted should be considered in low-risk groups. Additional efforts are needed to increase completion of treatment for LTBI in contacts identified in these settings.