During the early years of our annual tuberculosis skin testing program for hospital employees, we experienced conversion rates that were inordinately high. In an effort to eliminate the booster phenomenon in serial tuberculin testing as a contributing factor, a program of repeat skin testing performed within 7 to 28 days in new employees with less than 10-mm induration on initial skin testing was begun in 1978. Employees demonstrating a 6-mm increase in reaction size to a total induration of 10 mm or greater on repeat testing were classified as boosters. The booster rate for 1978 was 8.3%, and elimination of these employees from subsequent annual testing resulted in a decrease in the conversion rate from 8.7% in 1978 to 2.9% in 1979 (p less than 0.01). The rate of conversion was age-dependent in 1978, but this effect was absent in 1979 when the boosters were eliminated. On the basis of the results for the first 2 yr of our program, we concluded that repeat tuberculin skin testing is a practical approach to reduce falsely elevated conversion rates and can potentially reduce the number of persons exposed to the risk of isoniazid toxicity.