Reduced use of surface cultures for suspected neonatal sepsis and surveillance.

Data on infection in a neonatal unit were collected prospectively for seven years. After the first four years, the number of surface cultures obtained from neonates with suspected sepsis and for surveillance was reduced. Rates of systemic infection (sepsis and meningitis) were not significantly different in the four years before and the three years after this change. Reduction in surface culture information made no observable difference to detection of colonisation in neonates with early onset sepsis (within first 48 hours of life) nor to antibiotic choice in late onset sepsis. Decisions concerning the length of antibiotic course in suspected infection were not adversely affected. Reduction in the number of surface cultures led to considerable saving of time, effort, and cost while appearing safe in terms of clinical practice and outcome.