tuberculosis to 3 of 4 household members, 16 of 24 classmates, 10 of 32 school-bus riders, and 9 of 61 day-care contacts (Curtis et al., 1999). This case was also unusual because tuberculosis in children is not usually transmitted to others.
Although national CDC data show approximately 400 to 600 cases of tuberculosis in health care workers per year in recent years, case investigations of hospital or other outbreaks indicate that outbreaks reports account for only a small number of all cases of tuberculosis (Dooley and Tapper, 1997; Garrett et al., 1999, CDC, 2000b). For example, the 19 outbreak reports (1965 to 1995) summarized by Dooley and Tapper (1997) covering the period 1965 to 1995 account for fewer than 50 cases of active disease among health care workers. The 28 outbreak reports (one covering more than one institution) (1962 to 1996) reviewed by Garrett and colleagues (1999) account for fewer than 90 cases, of which 27 were associated with one 1962 to 1964 outbreak. The outbreak investigations have linked some cases of active disease to workplace exposure, but other cases were not explicitly linked to either a workplace or a community source. The reviews by Dooley and Tapper and Garrett and colleagues did not report on deaths associated with the outbreak studies they summarized, although Garrett and colleagues noted 9 deaths among at least 20 workers known to have contracted multidrug-resistant disease. Chapter 6 discusses what these investigations suggest about the association between outbreaks and the implementation of tuberculosis control measures.
As a source of information on the occupational risk of tuberculosis, outbreak reports have a number of limitations. Most are retrospective or observational. Skin testing procedures are often poorly described, as are data about test skin conversions and cases of tuberculosis. Important information about the facilities and their employees is often missing, inadequately described, or inconsistently measured across outbreak studies. For example, many reports do not include information about variables such as employee age, length of employment, job category, work location within a facility, country of origin, race, and past vaccination with bacille Calmette-Guérin (BCG).
In addition to the limitations of individual reports, no comprehensive, systematic national system exists for the reporting and publishing of information on workplace outbreaks of tuberculosis. Published reports do not represent the universe of outbreaks (even those investigated by CDC staff), and they may appear years after the first investigation of an outbreak. Moreover, published reports on outbreaks probably over represent unusual circumstances (e.g., an unusual location). Indeed, as the very