Tuberculosis (TB) skin testing practices and the prevalence and inherent risk of TB infection among hospital employees in 167 North Carolina (NC) hospitals were determined from a 79% (132/167) response to a tuberculosis screening questionnaire. Preemployment TB skin testing was performed by 98% of responding hospitals, primarily (87%) by the Mantoux method. TB skin test reactions of ≥ 10 mm were interpreted as significant by 72% and at the appropriate time interval of 48 to 72 hours after administration by 80%. The booster test was routinely performed in 12% of the hospitals. TB infection prevalence among new employees during 1983 was 6.3% (260/4137) in 30 hospitals supplying these data. A positive correlation was noted between employee infection prevalence and county TB case rates (P = .014). Skin test conversion data from 56 hospitals across the state revealed a five year mean conversion rate of 1.14% among 71,253 personnel. There was an association between the incidence of TB in the general population and the frequency of conversions among hospital employees in corresponding geographical regions. Similarly, the incidence of TB among approximately 100,000 NC hospital employees in 1983 and 1984 was less than the incidence in the general population. These associations suggest that the incidence of TB infection among hospital personnel may reflect the prevalence of tuberculosis in the community rather than an occupational hazard. Annual TB skin testing of hospital employees may be justified in eastern North Carolina where the incidence of tuberculosis (22-30 cases/ 100,000) is greater than the national average and where the risk of new TB infection among hospital employees is relatively common (≥1.5%).