OBJECTIVES: Nosocomial pneumonia (NP) is costly in terms of resource utilization and mortality. Patients with ventilator-associated pneumonia (VAP) have a higher risk of death than those with NP from other sources. The differences in risk and costs are attributable in part, to differences in the underlying pathogens. The purpose of this study is to characterize the outcomes and costs associated with NP, specifically examining the impact of mechanical ventilation. METHODS: We used Premier's 2003 hospital dataset for this study. These data originate from 1500 hospitals in the United States. Records with a non-missing admission code and a diagnosis of pneumonia sometime during the stay were retained. To restrict the sample to those with NP, records with a diagnosis of pneumonia at admission as well as those with an antibiotic on day one were deleted. CPT-codes were used to identify mechanical ventilation and text strings were searched to identify antibiotic therapy. Length of stay and costs were calculated by ward type. RESULTS: The final analytical sample contained 26,349 admissions: 762 stays involved mechanical ventilation, 9,495 had ICU time, and 16,092 had neither. Hospital mortality rates were 27.7%, 23.8% and 7.6% for these three groups, respectively. Median cost per stay was substantially higher for those with ventilation ($39,493) versus those with ICU time but no ventilation ($25,798) and those with neither ($7,261). Average length of stay in the ICU was 14 days and 9 days for those with and without ventilation, respectively. Average anti-infective drug costs were 1.79 times higher in the ventilator group compared with the ICU group. CONCLUSION: VAP is an area of high unmet need. Among these 1500 hospitals, 2003 costs for those with mechanical ventilation were 1.5 times higher than a group of NP cases that were fairly complex, as indicated by some receipt of intensive care services.