Validation of the Korean National Healthcare-associated Infections Surveillance System (KONIS): An Intensive Care Unit Module Report

Summary Background National surveillance data should be validated to identify methodological problems within the surveillance program and data quality issues. Aim This study tested the validity of healthcare-associated infection (HAI) rate data from the Korean National Healthcare-associated Infections Surveillance System (KONIS). Methods The intensive care unit (ICU) records of 12 (14.8%) of 81 participating hospitals between January and March of 2014 were examined. The validation team reviewed 406 medical records of 110 patients with 114 reported HAIs, including 34 urinary tract infections (UTIs), 57 bloodstream infections (BSIs), and 23 cases of pneumonia (PNEU), and 296 patients with no reported HAI during 1-day visits conducted in August and September 2014. The reviewers’ diagnosis of HAI was regarded as the reference standard; in ambiguous cases, the KONIS Steering Committee confirmed the diagnosis of HAI. Findings The sensitivity for UTI, BSI, and PNEU was 85.3, 74.0, and 66.7%; the specificity was 98.7, 99.1, and 98.7%; the positive predictive value was 85.3, 94.7, and 78.3%; and the negative predictive value was 98.7, 94.6, and 97.7%, respectively. The sensitivity for PNEU was lowest among the three ICU HAIs. The hospitals participating in KONIS infrequently reported conditions that were not HAIs. The sensitivity for BSIs was lower than in validation studies of KONIS conducted in 2008 and 2010. Conclusions KONIS data are generally reliable; however, the sensitivity for BSIs exhibited a decrease. This study shows the need for ongoing validation and continuous training of surveillance personnel to maintain the accuracy of surveillance data.