Sensitivity of Laboratory-based Surveillance for Detecting Nosocomial Infections

Jeong, Jae Sim; Ryu, Sun Joo; Kim, Yang Soo; Ryu, Ji So; Lee, Seong Hee; Pai, Chik Hyun

Abstract

BACKGROUND
Laboratory-based surveillance is an effective method' to detect nosocomial infections with limited personnel and time, and also can be used to identify clusters of organisms by unit and site. However it will fail to detect a significant number of infections because of cultures being negative or not submitted at all. This study was performed to determine the sensitivity of laboratory-based surveillance method in detecting nosocomial infections.
METHODS
Four nursing units, two each of medical and surgical units, of Asan Medical Center were chosen and surveillance for nosocomial infections was performed with the total surveillance method for the 6-month period from July to December of 1995 by an infection control nurse. Proportion of the nosocomial infections that would have been detected by culture results alone was identified to determine the sensitivity of the laboratory-based surveillance method.
RESULTS
A total of 164 cases of nosocomial infections were identified by the total surveillance method. Of these, 119 (72.6%) cases would have been detected by the laboratory-based surveillance method alone. Sensitivities were 100% in urinary tract infection and bacteremia, and 44.7%, 52.6% and 54.8% in lower respiratory tract, surgical site and other site infections, respectively. The 45 cases of infections that would have been undetected by microbiology data alone included 30 cases with culture negatives and 15 cases in which cultures were not submitted.
CONCLUSIONS
The laboratory-based surveillance revealed different sensitivities depending on the sites of infections and patient services. The method is efficient with an acceptable level of sensitivities, but infection control teams should be aware of its shortcoming by a periodic monitoring of its sensitivity in the detection of nosocomial infections at their own hospitals.