Contact isolation has been recommended by the Centers for Disease Control and Prevention for the prevention of nosocomial transmission of methicillin-resistant Staphylococcus aureus (MRSA), but there are few data which prospectively quantitate the effectiveness of contact isolation for this purpose. During an outbreak of MRSA in a neonatal intensive care unit between July 18, 1991 and January 30, 1992, weekly surveillance cultures were performed on all patients. Sixteen of 331 admissions became colonized with MRSA, and 3 (19%) developed infections: bacteremia, conjunctivitis, and dialysis catheter site infection. The isolates from all 16 patients were submitted to plasmid profile analysis and restriction enzyme analysis of whole cell DNA. All of the patients had identical chromosomal patterns and plasmid profiles, which differed from control isolates from other wards, indicating that the outbreak resulted from spread of a unique strain. None of 144 personnel who were cultured after recent contact with newly colonized patients during the outbreak were found to carry MRSA, which suggests that patients were the reservoir for transmission rather than caregivers. The most probable source for each individual transmission was determined based on proximity in time and space and shared exposure to caregivers. The rate of transmission of MRSA from patients on contact isolation was significantly lower (0.009 transmissions per day on isolation) than the rate for patients not on isolation (0.140 transmissions per day unisolated, relative risk = 15.6, 95% confidence interval 5.3-45.6, p < 0.0001). The authors conclude that the risk of nosocomial transmission of MRSA was reduced 16-fold by contact isolation during the outbreak in this neonatal intensive care unit. These data confirm the results of previous studies which have suggested that contact isolation was effective in controlling the epidemic spread of methicillin-resistant Staphylococcus aureus.