American Review of Respiratory Disease

An outbreak of tuberculosis at a shelter for homeless men was studied in detail to further the understanding of the epidemiology of tuberculosis in this setting. The shelter provides evening accommodations for men aged 50 yr and older. The capacity is approximately 200 clients, and the client pool is approximately 1,000 men/yr. During a 6-wk period in December 1986 and January 1987, seven cases of tuberculosis were diagnosed in shelter clients. Nine cases were reported in clients during the preceding 12 months, and four cases in the year previous to that. The majority of outbreak cases were pulmonary tuberculosis, sputum smear positive. Drug resistance was rare. Phage typing of 15 Mycobacterium tuberculosis isolates revealed one predominant type and four other types. The goals of the control plan (and the steps taken to achieve them) were to render known infectious cases noninfectious (directly observed therapy); to find undiagnosed infectious cases (repetitive mass screenings); to protect exposed clients (repetitive tuberculin skin testing and isoniazid preventive therapy); and to make the shelter environment safe (exclude infectious, noncompliant clients and improve the shelter's ventilation system). Implementation of this plan rapidly terminated the outbreak; following the first mass screening in January 1987, at which six asymptomatic cases were detected, only five additional cases occurred in shelter clients during a 2-yr period of follow-up. The investigation suggested that the outbreak evolved during 1986 as a result of the presence at the shelter of an increasing number of men with undiagnosed infectious pulmonary tuberculosis. Factors favoring disease transmission at the shelter were a relatively stable client pool, of which many were susceptible to tuberculosis; close proximity of bedding; and a closed ventilation system that provided extensive air recirculation during the winter months. Clients with prior tuberculosis infection and disease were not protected from involvement in this outbreak. The study suggests that at emergency shelters with a pattern of regular client usage, a program of ongoing tuberculosis control should be implemented to detect endemic cases, to limit the pool of clients with undiagnosed, infectious tuberculosis, and to prevent such outbreaks.


No related items