Epidemiological and Genomic analysis of a Sydney Hospital COVID-19 Outbreak

Abstract

Abstract Australia’s early COVID-19 experience involved clusters in northern Sydney, including hospital and aged-care facility (ACF) outbreaks. We explore transmission dynamics, drivers and outcomes of a metropolitan hospital COVID-19 outbreak that occurred in the context of established local community transmission. A retrospective cohort analysis is presented, with integration of viral genome sequencing, clinical and epidemiological data. We demonstrate using genomic epidemiology that the hospital outbreak (n=23) was linked to a concurrent outbreak at a local aged care facility, but was phylogenetically distinct from other community clusters. Thirty day survival was 50% for hospitalised patients (an elderly cohort with significant comorbidities) and 100% for staff. Staff who acquired infection were unable to attend work for a median of 26.5 days (range 14-191); an additional 140 staff were furloughed for quarantine. Transmission from index cases showed a wide dispersion (mean 3.5 persons infected for every patient case and 0.6 persons infected for every staff case). One patient, who received regular nebulised medication prior to their diagnosis being known, acted as an apparent superspreader. No secondary transmissions occurred from isolated cases or contacts who were quarantined prior to becoming infectious. This analysis elaborates the wide-ranging impacts on patients and staff of nosocomial COVID-19 transmission and highlights the utility of genomic analysis as an adjunct to traditional epidemiological investigations. Delayed case recognition resulted in nosocomial transmission but once recognised, prompt action by the outbreak management team and isolation with contact and droplet (without airborne) precautions were sufficient to prevent transmission within this cohort. Our findings support current PPE recommendations in Australia but demonstrate the risk of administering nebulised medications when COVID-19 is circulating locally.