A one-day pilot point-prevalence survey of healthcare-associated infections in three Korean hospitals: risk factors, clinical burden, and implications for nationwide surveillance

초록

Background: Healthcare-associated infections (HAIs) remain a major cause of morbidity and mortality in hospitalized patients. We assessed the prevalence, clinical burden, and predictors of HAIs across multiple hospitals using point-prevalence and matched case -control designs. Methods: A point-prevalence survey was conducted among all inpatients at three hospitals. For clinical burden analysis, HAI cases were matched to non-infected controls in a 1:3 ratio based on sex, age (+/- 10 years), ward type, length of stay up to the survey date, clinical department, and surgical history (for surgical site infections). Propensity score matching was performed to generate 1:1 matched pairs. Multi-variable logistic regression analyses were conducted to identify factors associated with HAI occurrence and to evaluate predictors of in-hospital and 30-day mortality. Results: Among 2226 patients, 119 had HAIs (prevalence, 5.35%). Recent microbiological culture testing, purulent discharge, documentation suggestive of infection in medical records, and fever were independently associated with HAIs in the propensity score -matched analysis. HAIs were also independently associated with increased in-hospital mortality, whereas the association with 30-day mortality was not statistically significant after adjustment for comorbidities and McCabe classification. Conclusions: In this multi-centre study, HAIs were associated with substantial clinical burden and increased mortality. Specific clinical indicators were strongly associated with HAI detection and may support surveillance strategies. These findings highlight the need for strengthened infection prevention and the integration of early identification systems into routine hospital surveillance. Trial registration: Not applicable. & ordf; 2026 The Healthcare Infection Society. Published by Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

키워드