12c - Automated detection of hospital outbreaks of antimicrobial-resistant bacteria in one Italian region, 2022 to 2024
Fireside Abstracts
Information
Background : Antimicrobial resistance (AMR) is a critical public health issue. Automated tools for detecting hospital outbreaks of drug-resistant pathogens are essential for informing prevention and control interventions. For the first time in Italy, WHONET-SaTScan was used at a regional level in Lombardy to identify hospital outbreaks of pathogens associated with healthcare-associated infections (HAIs) beyond the EARS-Net priority pathogens.
Methods : Microbiology data from the regional surveillance system, MICRO-BIO, covering 30 laboratories and 76 hospitals were analysed using WHONET-SaTScan to identify spatiotemporal clusters of drug-resistant pathogens. Cluster detection analyses were performed employing a simulated prospective analysis covering the first six months of 2024 and using the previous 24 months as historical data. A space–time uniform algorithm was used, and clusters were identified at three spatial levels: ward, groups of wards (‘meta-wards’), hospital using a recurrence interval of 365 days for statistical significance.
Results : The simulated prospective analysis included 41,435 isolates from 25,839 patients. Overall, 285 clusters were identified, with Enterococcus faecalis and Pseudomonas aeruginosa being the most frequently detected organisms. General medicine wards had the highest number of clusters detected (n=25, 6.9%). Clusters were primarily detected in large hospitals (92.9%) and at the meta-ward level (36.1%). There was no evidence for a difference in cluster length, cluster size, recurrence intervals, or days to first alert detected between the three spatial levels. Compared to small and medium-sized hospitals, clusters in large hospitals were larger (p<0.001) and accumulated cases for longer periods (p=0.04).
Conclusions : WHONET-SaTScan effectively identified clusters of drug-resistant pathogens consistent with hospital outbreaks. Its implementation as a real-time outbreak detection tool to support prevention and control interventions.would be particularly valuable in large hospitals.
