Previous studies of risk factors for hepatitis B virus infection among hospital employees have been based on surveys in single institutions or have been analyzed with univariate techniques. From November 1980 through August 1981, the authors performed a multi-institutional seroepidemiologic survey of hospital employees screened for entrance into a hepatitis B vaccine trial who represented groups at high risk for hepatitis B infection. Using a logistic regression model for the analysis of risk factors, the investigators determined the relative odds and 95% confidence intervals for risk of hepatitis B infection to be as follows: race (nonwhite/white: 3.4; 2.4-4.8) (p less than 0.001); history of acute viral hepatitis of an unspecified type (3.6; 2.2-5.9) (p less than 0.001); and employment at hospitals 1 through 5 as compared with hospital 6 (1.8; 1.1-2.9) (p = 0.015). In addition, certain job categories and the duration of employment within some of these categories were associated with increasing risk for hepatitis B infection over time. Laboratory workers (1.4; 1.2-1.7), surgical staff (1.2; 1.1-1.4), and medical staff (1.3; 1.1-1.5) had significant (p less than 0.05) increased risk of prior infection with longer duration of employment. Such time-job interaction was not demonstrable for nursing staff, anesthesiology staff, dental personnel, pathology staff, or ancillary personnel. The logistic regression model also shows that the highest gradient of risk for laboratory workers, surgeons, and medical staff occurs during the first five years of employment. An effective preventive strategy, such as the use of hepatitis B vaccine, should be targeted for these groups at the time of initial employment.