935 Construction of the Korean Nosocomial Infections Surveillance System using Web-based Report and Analysis Program: A Validation Study

Sunday, March 21, 2010
Grand Hall (Hyatt Regency Atlanta)
Yong Kyun Cho, MD, PhD , Gachon University Gil Hospital, Incheon, South Korea
Sang-Oh Lee, MD, PhD , Asan Medical Center, Seoul, South Korea
Hyo Youl Kim, MD , Wonju Christian Hospital, Wonju, South Korea
Young Keun Kim, MD , Asan Medical Center, Seoul, South Korea
Hyun Kyun Ki, MD , Division of Infectious Diseases, Konkuk University Hospital, Seoul, South Korea
Eu Suk Kim, MD , Department of Infectious Diseases, Dongguk University International Hospital, Goyang, South Korea
Hong Bin Kim, MD, PhD , Seoul National University Bundang Hospital, Seongnam, Gyonggi-do, OC, REPUBLIC OF KOREA
Eui-Chong Kim, MD, PhD , Department of Laboratory Medicine, Seoul National University Hospital, Seoul, South Korea
Background: After founding the Korean Society for Nosocomial Infection Control (KOSNIC) in 1995, we had conducted several multi-center surveillance studies for nosocomial infections (NIs). Although KOSNIC gained valuable experience by running these studies, a number of limitations were encountered; the number of participating hospitals was too small to represent nationwide data, multi-center surveillance was not maintained continuously, and data management was considered rather ineffective because the data had been collected by written work-sheets. Therefore, in 2006, KOSNIC organized the Korean Nosocomial Infections Surveillance System (KONIS) to establish a nationwide database of NI rates in the intensive care units (ICUs) of Korean hospitals. The KONIS Steering Committee made a consensus of standardized practices for surveillance that was summarized in the KONIS manual. The committee conducted prospective surveillance for three major NIs: urinary tract infection (UTI), bloodstream infection (BSI), and pneumonia. The committee then educated all investigators regarding the contents of the KONIS manual. KONIS began on 1 July 2006, with data being collected using an Internet-based data entry interface (KONIS Web-based Report and Analysis Program, KONIS WRAP, http://konis.cdc.go.kr). Investigators of KONIS member hospitals should input patient-days and device-days monthly on the internet. And they report the data of nosocomial infection cases into a form, which consists of demographic data, admission date, infection date, and type of infections.Objective: We firstly performed study to evaluate the validity of the KONIS intensive care unit data.

Methods: The KONIS Steering Committee reviewed surveillance data from January through March 2008. All KONIS hospitals were divided into six groups according to the nosocomial infection rates. Each one hospital was randomly selected from six groups. The trained members of KONIS Steering Committee visited each hospital and they reviewed medical records. The samples included patients with NIs reported to KONIS and randomly selected patients without reports of these infections.

Results: We reviewed 103 (16% of total Nis) reported cases and age and sex matched 204 patients with no reported infection randomly selected from 96 ICUs of 56 hospitals. Sensitivities of UTI, BSI, and pneumonia were 93%, 87%, and 71%, respectively. And specificities of UTI, BSI, and pneumonia were 100%, 100%, and 99%, respectively.

Conclusions: Compared with previous e-mail system, using KONIS WRAP has made the surveillance activities more standardized and more efficient. The accuracy of KONIS was similar to that of other countries such as National Nosocomial Infections Surveillance Systen (NNIS) in United States.