We agree with Dr. Lai's restatement of our findings. In addition, Dr. Lai asks specifically about the cost-effectiveness of rectal surveillance cultures in an intensive care unit in which vancomycin-resistant enterococci are already endemic.

Most patients who are colonized by vancomycin-resistant enterococci never develop infections, and those who do also have multiple comorbid conditions that complicate the assessment of treatment costs. Moreover, the incremental benefit of surveillance cultures, beyond that of other control measures, is difficult to measure. Therefore, the cost-effectiveness of rectal surveillance cultures might be difficult to ascertain, even in the context of a prospective, controlled clinical trial. We have found that surveillance cultures are most helpful in tracking responses to an intervention and that they allow epidemiologic typing of isolates. If isolates are found to represent a single strain, evaluation for a possible common source of contamination should be done. In most instances, however, dissemination of a single strain of vancomycin-resistant enterococci results from lapses of hygiene that result in indirect patient-to-patient spread of organisms via the unwashed hands of health care workers and environmental contamination. The presence of a single strain suggests that maximal control efforts [1] may be effective. Outbreaks of multiple strains of vancomycin-resistant enterococci seem to be more difficult to control; this may reflect the repeated introduction of resistant strains from other units, other hospitals, or nursing homes [2] or the presence of a promiscuous resistance plasmid. If typing of surveillance isolates indicates that multiple strains are circulating, additional control strategies may include developing a source-patient profile; evaluating newly admitted patients as possible sources; introducing more stringent antibiotic controls; developing a vancomycin-resistant enterococci cohort system; and using more aggressive standard precautions, as described below.

With regard to Dr. Lai's second question, we note that the durability of colonization with vancomycin-resistant enterococci may depend on patients' comorbid conditions and antibiotic exposures. In our intensive care unit [3], as in Dr. Lai's [4], patients seem to remain colonized during their entire stay. For this reason and because patients may introduce enterococci from other units or other institutions, we recommend that gloves and handwashing be used for contact with all acutely ill patients and their environments in hospitals in which vancomycin-resistant enterococci are endemic. This approach goes beyond standard precautions [5] to include the intact skin and the patient's environment.