Olivier Bedard-Lapointe, Xuejiao Shelly Wei, Nick Daneman, Kevin Brown, Fizza Manzoor, Mona Golmohammadzadeh, Dominik Mertz, Zain Chagla, Jennifer A Flemming, Sudeep S Gill, Anthony D Bai
In this cohort study of 114 hospitals, there was significant variability in hospital HOB rates, suggesting that HOB may be a useful quality metric in discriminating between hospitals.
IMPORTANCE: Hospital-onset bacteremia and fungemia (HOB) is being proposed as a new hospital quality metric for health care-associated infections. There is a lack of data on the ability of HOB to discriminate between hospitals from a large regional perspective.
OBJECTIVE: To describe and compare the HOB rates across hospitals for the province of Ontario in Canada.
DESIGN, SETTING, AND PARTICIPANTS: This population-based retrospective cohort study used ICES Ontario databases. Participants were adult patients admitted to an acute care hospital from 2017 to 2024 in Ontario, Canada.
EXPOSURE: Local care processes and practices in different hospitals.
MAIN OUTCOMES AND MEASURES: HOB was defined as a positive blood culture collected on or after hospital day 4. HOB rates per admission-days were described across hospital sites. A generalized linear mixed-effects model was used to compare variation in HOB rates across hospitals while adjusting for patient and hospital characteristics.
RESULTS: At 114 acute-care hospitals in Ontario, Canada, there were 3 325 900 hospital admissions (median [IQR] age, 71 [57-81] years; 1 727 228 female patients [51.9%]). In total, there were 61 452 HOB episodes observed over 29 721 745 admission-days, equal to an overall rate of 2.07 episodes per 1000 admission-days. Across the 114 acute-care hospitals, the HOB rates were 0.39 episodes per 1000 admission-days for the 10th percentile, 0.71 episodes per 1000 admission-days for the 25th percentile, 1.08 episodes per 1000 admission-days for the 50th percentile, 1.68 episodes per 1000 admission-days for the 75th percentile, and 2.27 episodes per 1000 admission-days for the 90th percentile. In a generalized linear mixed-effects model, the random intercept for hospital sites had an SD of 0.36 (95% CI, 0.29-0.42), which corresponded to a rate ratio of 1.43 (95% CI, 1.34-1.52) per 1 SD.
CONCLUSIONS AND RELEVANCE: In this cohort study of 114 hospitals, there was significant variability in hospital HOB rates, suggesting that HOB may be a useful quality metric in discriminating between hospitals.