conducted a risk assessment using some of the data cited above. For its estimation of the annual risk of infection in general populations, OSHA developed a model based on relating risk of infection to incidence of disease. Estimates of the prevalence of tuberculous infection in the United States were also provided to OSHA by Dr. Christopher Murray of Harvard University. OSHA concluded that the overall annual rate of infection in the general population of the United States varied by state, from a low of 0.0194 percent/year to 0.3542 percent/year, and chose a population size-weighted average of 0.146 percent/year for the country as a whole.
As its database for estimating the annual rate of tuberculous infection in health care workers, OSHA used information published and obtained directly from the state of Washington (29, 45), the state of North Carolina (45, 78), and Jackson Memorial Hospital in Miami, Florida (39, 45, 79). Using these data, OSHA estimated that the occupational risk in Washington was 1.5 times, that in North Carolina was 5 times, and that at Jackson Memorial Hospital 9 times that for the general population of the surrounding state, region, or community. Similar estimates were made for workers in nonhospital settings. For Washington State the occupational risk for employees of long-term-care facilities was judged to be 11 times that for the general population and for home health care workers it was 2 times that for the general population. OSHA’s risk estimate for the population of the United States as a whole in 1994 is about three times that of 0.05%/year considered by the author to be his best estimate of the national rate.
The most careful attempts to assess the occupational tuberculosis risk of health care workers in relation to their workplaces and the communities in which they reside are studies conducted by the Division of Tuberculosis Elimination of CDC. These investigations include some of the only prospective studies of the problem. They are also notable because they all included initial two-step tuberculin skin testing to minimize confounding booster effects. Some of them have not yet been published, but abstracts were kindly made available to me by their authors, who gave me permission to cite them.
Panlilio and Burwen followed 1,961 initially tuberculin-negative health care workers in Boston and New York City at 6-month intervals beginning in April 1994 and reporting their results in abstract form in May 1996 (80). Overall, 30 (1.5 percent) conversions were documented. Conversion was correlated with foreign birth, Asian race, and recent entry into the United States. The authors concluded that it was difficult to determine the source of infections in their subjects.