respiratory protection standard for other hazards. Again, the outbreak studies suggest that the administrative controls adopted by hospitals played the major role in ending the outbreaks and that the kinds of respiratory protections they implemented added little.
It is also illustrative to examine two reports from nonoutbreak hospitals that had relatively high admission rates for patients with active tuberculosis and had adopted tuberculosis control measures to reduce their potential for outbreaks. As summarized below, these reports also suggest a limited role for respiratory protections. Again, all control measures were adopted prior to the 1994 CDC guidelines and the NIOSH certification of N95 respirators.
In May 1992 Columbia-Presbyterian Hospital in New York City revised its infection control program to be consistent with the 1990 CDC guidelines (Bangsberg et al., 1997). The facility had not experienced an outbreak of tuberculosis, but administrators were concerned about the potential for an outbreak based on reports from other city institutions. Columbia first (May 1992) instituted extensive administrative controls that emphasized stricter respiratory isolation policies; shortly thereafter (July 1992), it installed two tuberculosis isolation rooms in the emergency department. In July 1993, the hospital began to require that medical house staff don a 3M disposable respirator to enter respiratory isolation rooms; they provided surgical masks prior to that. House staff were fit tested and instructed in the use of the new devices. The tuberculin skin test conversions among house staff dropped from 10 percent preimplementation to 0 percent to 2 percent for time intervals after implementation of the administrative controls and engineering controls but before the provision of new respirators. The authors felt that administrative controls were the main reason for the improvements observed.
St. Clare’s Hospital in New York City implemented the 1990 CDC guidelines in 1991 (Fella et al., 1995). This hospital focused first on administrative controls (especially, early recognition and isolation of patients with active tuberculosis) and then on engineering controls (including installation of 44 negative-pressure isolation rooms in a 2-year period and installation of ultraviolet [UV] germicidal irradiation lights in patient rooms and general use-areas). The institutions made a series of changes in the respiratory protection devices provided employees (switching in January 1992 from the Technol shield to a particulate respirator, then in January 1993 to a dust-mist-fume respirator with fit testing beginning in June 1993, and, finally to HEPA respirators in 1994 after the study period ended). From 1991 to 1993, the tuberculin skin test conversion rate among health care staff fell from 20.7 percent in the first 6-month testing interval to a range of 3.2 to 6.2 percent over subsequent 6-month intervals. Changes in conversion rates were not associated with changes in personal respiratory protection.