practices in three hospitals in high-incidence counties in California (Sutton et al., 1998, 2000). The investigators used questionnaires and reviews of written documents to assess policies. They then attended tuberculosis control meetings and training sessions, directly observed work practices for patient isolation, and measured the ventilation performances of isolation rooms. The first report indicated that of 67 rooms equipped with continuous airflow monitoring devices, devices in 8 rooms did not accurately reflect the direction of airflow. Of 62 workers observed using a respirator, 65 percent did not put it on properly. In the second report from the study, investigators found that only one hospital followed the CDC’s recommendations for respiratory protection. Of the isolation rooms tested, 28 percent (7 of 25) were under positive pressure. In most of the rooms tested (26 of 27), air moved toward rather than away from workers. None of the three facilities regularly checked the performances of the isolation rooms.

In 1994, researchers associated with a midwestern hospital system combined a written survey with record reviews and on-site testing of isolation rooms in seven rural and six urban hospitals (Woeltje et al., 1997). All hospitals reported having tuberculosis control plans and performing annual tuberculin skin testing. Eleven of 13 hospitals had negative-pressure isolation rooms. The researchers found that the median percentage of rooms with effective negative pressure was 95 percent (with one institution reporting a median of only 44 percent). Three hospitals provided high-efficiency particulate air masks, and eight provided dust-mist or dust-mist-fume masks. This inspection occurred before NIOSH had certified the use of N95 respirators. Actual worker use of the masks was not observed.

Although reports of facility inspections cover relatively few institutions, the results may still provide some insights into the match between institutional policies and routine, day-to-day practices. In general, they suggest that departures from recommended tuberculosis control measures occur at both the institutional level (e.g., provision of appropriate respirators) and the individual level (e.g., appropriate use of respirators). Implementation is probably most complete for administrative controls including written plans and procedures. For engineering controls, implementation is likely better for the installation of isolation rooms than for their day-to-day operation in accordance with guidelines.

Implementation Lapses Mentioned in Outbreak and Other Case Reports

Rather than systematically describing the implementation of tuberculosis control measures, outbreak reports typically focus on factors that might have contributed to the outbreak, including the failure to implement specific controls. As discussed later in this chapter, most reports



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement