The number of patients with pulmonary tuberculosis appropriately isolated during January through June 1992 (preintervention) was only 29/71 (38 percent). This increased to 29/45 (64 percent) from January to December 1992. Subsequent isolation rates continued to improve slightly: 60/82 (72 percent) from January to June 1993 and 33/44 (75 percent) from July to December 1993 (p < 0.01 for trend). Results considering only HIV infected patients were similar.
PPD conversion rates among house staff were as follows: June 1992, 10 percent (5.8/100 person-years); December 1992, 3 percent (5.1/100 person-years); June 1993, 0 percent; December 1993, 1 percent (2.3/100 person-years); June 1994, 0 percent. Conversion rates were calculated per 100 person-years of exposure because of varying exposure times possible at the June 1992 testing (12–36 months, depending on the year of the resident).
Because the biggest drop occurred between December 1992 and June 1993, the authors imply that isolation policy and possibly the tuberculosis isolation rooms in the ED were most important in leading to the improvements. Clearly their expanded isolation policy resulted in much better isolation of patients with pulmonary tuberculosis over this time period.
Stroud and colleagues (18) reviewed the effects of control measures at Roosevelt during three 15-month periods: period I, January 1989 to March 1990; period II, April 1990 to June 1991; and period III, July 1991 to September 1992. Period I was essentially a preintervention period, during which there was an outbreak of nosocomial tuberculosis (7). Patients with suspected tuberculosis were admitted to private room (only 1 of 16 with negative pressure), doors were often left open, and isolation was discontinued without negative AFB smears. Surgical masks were used for respiratory protection. Most rooms, however, did exhaust to the outside.
During period II administrative controls were enforced—a lower threshold for initiating isolation was set, more aggressive evaluation for possible tuberculosis was started, and more aggressive treatment regimens were started if there was no response to initial therapy. An effort was made to keep HIV-infected patients off wards with tuberculosis patients.
In period III engineering controls were phased in. From July to December 1991, 11 rooms were fitted with UVGI. From November 1991 through January 1992 seven of these rooms were fitted with exhaust fans for ≥6 ACH and negative pressure. Isolation chambers were used for sputum induction/aerosolized pentamidine administration. Surgical masks (Technol 47080070) were used through all three study periods. With the implementation of administrative controls during periods II and III, patients with pulmonary tuberculosis were more likely to be isolated on admission (44 percent versus 0 percent during period I). The median delay before isolation initiated (2 versus 6 days) also improved.