that have a policy of transferring such individuals, workers are likely to see no or very marginal additional protection from an extensive respiratory protection program. In a high-risk setting with many tuberculosis admissions, questionable administrative and engineering controls question, and, especially, cases of multidrug-resistant disease, a rigorous respiratory protection program may be beneficial.

The 1997 proposed OSHA rule allows little flexibility for organizations to adopt respiratory protection programs that reflect the variability in the level of risk facing workers. For low-risk institutions, a proportionately modest program might include the availability of N95 respirators and the training of key individuals in their appropriate use. In high-risk facilities, a program might include a spectrum of respiratory protection devices including N95 respirators for most situations and a more protective respirator for selected high-hazard procedures such as bronchoscopy and autopsy. The education and fit-testing elements of a respiratory protection program would then be tailored to the risk facing different employees.

In both high- and low-risk institutions, the highest priority would still be administrative controls that promote prompt identification and isolation of those with signs and symptoms suspicious for tuberculosis. For institutions that may admit those with tuberculosis, engineering controls are also important.

Given the variability of masks from different manufacturers that was noted earlier, it may be appropriate for policymakers to focus more attention on manufacturers so that generally poor-fitting respirator models are not marketed. In addition, further research and analysis may be useful to examine fit-testing criteria and methods in laboratory versus operational settings and to determine levels of respiratory protection that will reasonably reduce risk in environments posing different degrees of risk to workers.

Overall, the committee concludes that an OSHA tuberculosis standard can have a positive effect if it meets three basic conditions: (1) it is consistent with tuberculosis control measures that appear to be effective, (2) it sustains or increases the level of compliance with those measures, and (3) it allows employers appropriate flexibility to adopt control measures that are matched to the level of risk facing their workers. The committee expects that a standard will meet the first two conditions by sustaining or increasing the rate of use of tuberculosis control measures that appear to be effective. The committee is, however, concerned that if an OSHA standard follows the 1997 proposed rule, it will not meet the third condition of allowing organizations reasonable flexibility to adopt measures appropriate to the level of risk facing their workers.