tions that are at negligible risk of occupational transmission of M. tuberculosis, the standard is unlikely to benefit workers at the same time that it would impose significant costs and administrative burdens on covered organizations and absorb institutional resources that could be applied to other, potentially more beneficial uses.

In addition to the broad concerns about whether an OSHA standard would allow sufficient flexibility for organizations to match tuberculosis control measures to the risk facing workers, a narrower question is whether a standard would allow organizations reasonable flexibility to adjust tuberculin skin testing programs to reflect the changing epidemiology of tuberculosis and, possibly, changing CDC recommendations. As described in Chapter 2 and Appendix B, when the prior probability of tuberculosis infection is low because of low prevalence, Bayes’ Theorem shows that the probability of false positive test results increases. When prevalence decreases to very low levels, the majority of those with positive tests will not in fact be infected. This, in turn, increases the potential for workers to be treated unnecessarily for latent tuberculosis infection. The most serious possible harm of unnecessary treatment involve a very small risk of liver damage, although this risk is lower than previously thought (Nolan et al., 1999, as discussed in Chapter 2). Less serious potential harms include rashes, gastrointestinal upsets, fever, and joint pain. In addition, some people may suffer needless anxiety or fear related to a false-positive test result or subsequent treatment. Furthermore, the less an individual’s social contacts understand about the meaning of a positive test result, the greater the potential for a person’s social relationships to be compromised by such a result. Excessive workplace testing and treatment efforts would also waste resources that could be constructively used to support other aspects of a workplace tuberculosis control program.

That the tuberculin skin test has some limitations, especially in low-prevalence environments, does not mean that it is a poor test. It has been a valuable element in tuberculosis control programs, including in outbreak situations and as part of surveillance programs in health care and other facilities in high prevalence areas. Nonetheless, as recommended by the IOM report Ending Neglect: The Elimination of Tuberculosis in the United States (IOM, 2000), better diagnostic tests for both infection and active disease are needed.

Recognizing that the circumstances that prompted the 1994 guidelines have changed, the CDC’s Advisory Council for the Elimination of Tuberculosis recently recommended that CDC review and, if appropriate, revise the guidelines, including the recommendations for tuberculin skin testing. Such a review is now under way, and the committee under-