Background: Generation of proposed policies directed at interrupting the transmission of Mycobacterium tuberculosis resulted in a need for data establishing health care worker (HCW) tuberculosis (TB) transmission risk and control programs that were in place before 1992. Methods: A voluntary sample of 18 Minnesota health care facilities was surveyed retrospectively for the years 1989 to 1991. The survey was designed to establish the frequency of positive tuberculin skin test (TST) results, the rate of positive results during routine screening, the incidence of positive TST results identified through exposure follow up, the type of respiratory protection (mask) used, and ventilation patterns of TB isolation rooms. The TST testing program, including name of product used, method of application, and reading/documentation behaviors, was requested. Results: Data were analyzed from 17 hospitals (one long-term facility was deleted). Three hospitals screened all HCWs annually; the remaining hospitals screened HCWs at varying intervals. Forty-eight positive TST results were identified in routine screening. The rate per 10,000 HCWs screened was 5.8 in 1989; 15.9 in 1990; and 14.8 in 1991. An upward trend was noted in positive TST results only in metropolitan teaching/public hospitals. Intercomparisons of TST screening data demonstrated that employees located in Greater Minnesota were more likely to have positive TST results than employees at facilities in the metropolitan area. Hospitals in Greater Minnesota were more likely than metropolitan area hospitals to use tuberculin skin test material. HCWs screened with tuberculin purified protein derivative were more likely to have positive test results than HCWs screened with tuberculin purified protein derivative. Unisolated patients with communicable TB accounted for 445 patient days of probable risk of exposure for HCWs, without demonstration of a trend by calendar year. There were a total of 33 exposure events. Ten of 1031 HCWs screened after an exposure event were declared to be TST positive, for an overall conversion rate of 9.7 per 1000 employees screened. All hospitals used surgical masks for respiratory protection. Seventy-six percent of the hospitals had isolation rooms with negative ventilation; 12% reported isolation rooms with neutral pressure. Conclusion: The practice in Minnesota hospitals surveyed was reasonably consistent with the critical elements defined in the 1990 CDC guidelines for an effective TB control program. The rate of positive TST results was low, with programs in place before 1992. Unified TST programs permitting evaluation of programs are endorsed. Respiratory protection should be protective, not excessive. Engineering controls of isolation rooms should be ensured. Policy development should be driven by the prevalence and incidence of disease in the HCW's place of employment.