BACKGROUND Methicillin-resistant Staphylococcus aureus (MRSA) generates concern in nursing homes. Restrictive isolation precautions may be applied for indefinite periods. Adverse events driving these concerns include transmission and infection. METHODS The 721-bed Wisconsin Veterans Home in King performs approximately 645 cultures annually. The site, severity, and number of MRSA infections were determined for 69 months. Pulsed-field gel electrophoresis was performed on all initial isolates, followed by a statistical cluster analysis looking for evidence of transmission. RESULTS Sixty-seven MRSA infections were identified (1.6 per 100 residents per year); many were polymicrobial, and it was difficult to determine the proportionate role of MRSA in morbidity or mortality. There was an episode of rapidly fatal MRSA septicemia in which empiric antibiotic therapy was ineffective. Twenty-one genetic strains were encountered. Statistical analysis identified 13 clusters of genetically identical strains clustered in time and space (P<.05). CONCLUSIONS Infections with MRSA were identified at relatively low rates; however, the etiology of many serious nursing home infections is not determined, especially pneumonia. Statistical analysis revealed clustering and evidence of transmission. Nursing home practitioners should consider MRSA when applying empiric treatment to serious infections. We recommend a program including (1) judicious use of antibiotics, including topical agents, to reduce selection of resistant organisms; (2) obtaining and tracking cultures of infectious secretions to diagnose MRSA infections and focus antibiotic therapy; (3) universal standard secretion precautions because any resident could be a carrier; and (4) a detailed assessment and care plan for the carrier that maximizes containment of secretions and independence in activities. However, basic hygiene cannot be maintained in communal areas by some residents without restriction of activities of daily living.