Hospitals use many strategies to control nosocomial transmission of vancomycin-resistant enterococci (VRE). Strategies include "passive surveillance," with isolation of patients with known previous or current VRE colonization or infection, and "active surveillance," which uses admission cultures, with subsequent isolation of patients who are found to be colonized with VRE. We created a mathematical model of VRE transmission in an intensive care unit (ICU) using data from an existing active surveillance program; we used the model to generate the estimated benefits associated with active surveillance. Simulations predicted that active surveillance in a 10-bed ICU would result in a 39% reduction in the annual incidence of VRE colonization when compared with no surveillance. Initial isolation of all patients, with withdrawal of isolation if the results of surveillance cultures are negative, was predicted to result in a 65% reduction. Passive surveillance was minimally effective. Using the best available data, active surveillance is projected to be effective for reducing VRE transmission in ICU settings.