Surveillance for the detection of nesocomial infections and the potential for nesocomial outbreaks I. Microbiology culture surveillance is an effective method of detecting nosocomial infection

Abstract

A study was undertaken to determine the sensitivity and specificity of a laboratory-based surveillance system using numbers of positive cultures in our institution. A randomized retrospective review of 575 medical-surgical patient charts revealed 70 nosocomial infections in 41 patients. Review of microbiology records indicated positive cultures for 71%, cultures obtained but inadequate for 6%, no culture necessary for 7%, and no culture obtained for 16%. Patients over 60 years old were significantly more likely to have a nosocomial infection cultured than were younger patients (p < 0.025). There was no significant difference in rate of nosocomial infections classified by clinical service, site of infection, or length of stay in the hospital. By randomly selecting 100 positive cultures and reviewing the charts of the affected patients, we found that 48% represented nosocomial infections. Thus laboratory-based surveillance was 84% sensitive and 48% specific for detecting nosocomial infections.