OBJECTIVES: In a Phase-3 clinical trial we compared hospital and outpatient antibiotic treatment costs for MRSS patients treated with linezolid to those treated with vancomycin. Linezolid, first of a new class of antibiotics, the oxazolidinones, is active against gram positive bacteria sensitive and resistant to other antibiotics and is available in bioequivalent oral and IV formulations. Hospital length of stay data from this trial (Li et al, 2001), suggests linezolid use may result in earlier discharge than vancomycin use. METHODS: Hospitalized patients with pneumonia, skin/soft tissue infection, right-sided endocarditis, urinary tract infection, or bacteremia caused by MRSS were treated with linezolid (IV/oral, n=240) or vancomycin (IV only, n=220) in a randomized multi-center clinical trial. Of the 460 intent-to-treat (ITT) patients, 254 were clinically evaluable. Patients received up to four weeks of treatment followed by up to four weeks of post-treatment observation, including outpatient use of study medication. Total hospital costs were calculated using observed resource use (inpatient days by hospital unit type, non-investigational therapies and medications, and study medications) and 1999 unit costs from 8 countries, adjusted to 2001 price levels. Using a societal perspective, total costs are compared using mean, median and regression-adjusted (linear and log) differences. RESULTS: Mean and median total costs in the ITT sample were higher for linezolid patients than vancomycin patients, but no differences were statistically different from zero. Non-normality caused unadjusted mean differences to be much higher than median differences. Regression analysis explained about 40% of the cost variance and indicated no significant total cost differences between therapies. Clinically evaluable sample results were qualitatively similar to the ITT sample. CONCLUSION: Substantial variation in hospital costs in this moderately large multinational Phase-3 trial does not allow definitive conclusions regarding whether the length of stay differences seen earlier result in total treatment cost differences. In future research, combining data with other similar trials may allow for more precise point estimates of cost differences.