External validation method of the Korean Healthcare-associated Infections Surveillance System (KONIS) intensive care unit (ICU) module
| Survey year | Target HAI of ICU | Number of surveyed patients | Target period | Selection of hospitals or ICUs | Selection of patients | Investigators | Survey method |
|---|---|---|---|---|---|---|---|
| 2008 | UTI, BSI, PNEU | 307 | January 2008-March 2008 | Hospitals were divided into 6 groups based on HAI rate percentiles: <10, 10-25, 25-50, 50-75, 75-90, and >90. One hospital was then randomly selected from each group, total 6 hospitals | All reported HAIs and randomly selected non-HAI ICU patients In hospitals ≤10 HAIs: non-HAI patients were selected at a ratio of 1:4 In hospitals >10 HAIs: non-HAI patients were selected at a ratio of 1:2 |
A 2-member validation team visited each hospital for 1 day and conducted a retrospective chart review | Discrepant cases were discussed by the KONIS Steering Committee. |
| 2010 | UTI, BSI, PNEU | 321 | July 2009-September 2009 | The same as above | The same as above | A 2-member validation team visited each hospital for 1 day and conducted a retrospective chart review | Discrepant cases were discussed by the KONIS Steering Committee. |
| 2012 | PNEU | 55 | January 2012-March 2012 | Selection of 5 hospitals randomly based on the PNEU rate | All reported 11 PNEU patients and non-PNEU ICU patients randomly selected at a 1:4 ratio. | A 2-member validation team visited each hospital for 1 day and conducted a retrospective chart review | Discrepant cases were discussed by the KONIS Steering Committee. |
| 2014 | UTI, BSI, PNEU | 406 | January 2014-March 2014 | Hospitals were divided into 6 groups based on HAI rate percentiles: <10, 10-25, 25-50, 50-75, 75-90, and >90. Then, two hospitals were randomly selected from each group, resulting in a total of 12 hospitals | All reported HAI patients were included in the review. For each hospital, a number of non-HAI patients, two-fold (for the hospitals that reported ≥10 HAI patients) or four-fold (for the hospitals that reported <10 HAI patients) that of HAI patients, were selected at random. | A trained one- to three-member validation team visited each hospital for one day and conducted a retrospective chart review | Discrepant cases were discussed by the KONIS Steering Committee. The participating hospitals were notified of the results of the validation survey |
| 2017 | UTI, BSI, PNEU | 282 | July 2016 - September 2016 |
Hospitals were divided into five groups based on the HAI rate percentiles: 0, 0-25, 25-50, 50-75, 75-100. Hospitals within each percentile group were then stratified into new participating hospitals and existing participating hospitals. After this stratification, a total of 10 hospitals were randomly selected, with one from each of the 10 groups. | All HAI patients were included in the review. The control patients without HAI was selected randomly from all patients aged ≥15 years in the participating of each hospital who had positive blood or urine cultures during the validation period. Hospitalized for < 3 days were excluded. | A 2-member validation team visited each hospital for 1 day and conducted a retrospective chart review | The validation team provided the survey results to the staff in charge at each hospital and confirmed the final diagnosis after a post-review discussion. Cases in which it was difficult for the investigators to determine whether HAI had occurred were discussed by the KONIS Steering Committee. |
Abbreviations: HAI, healthcare-associated infection; UTI, urinary tract infection; BSI, bloodstream infection; PNEU, pneumonia.
Korean J healthc assoc Infect Control Prev 2023;28:192~202 https://doi.org/10.14192/kjicp.2023.28.2.192
© Korean J healthc assoc Infect Control Prev