Healthcare-Associated COVID-19 across Five Pandemic Waves: Prediction Models and Genomic Analyses

Abstract

Background: Healthcare-associated SARS-CoV-2 infections need to be explored further. Our study is an analysis of hospital-acquired infections (HAIs) and ambulatory healthcare workers (aHCWs) with SARS-CoV-2 across the pandemic in a Belgian university hospital. Methods: We compared HAIs with community-associated infections (CAIs) to identify the factors associated with having an HAI. We then performed a genomic cluster analysis of HAIs and aHCWs. We used this alongside the European Centre for Disease Control (ECDC) case source classifications of an HAI. Results: Between March 2020 and March 2022, 269 patients had an HAI. A lower BMI, a worse frailty index, lower C-reactive protein (CRP), and a higher thrombocyte count as well as death and length of stay were significantly associated with having an HAI. Using those variables to predict HAIs versus CAIs, we obtained a positive predictive value (PPV) of 83.6% and a negative predictive value (NPV) of 82.2%; the area under the ROC was 0.89. Genomic cluster analyses and representations on epicurves and minimal spanning trees delivered further insights into HAI dynamics across different pandemic waves. The genomic data were also compared with the clinical ECDC definitions for HAIs; we found that 90.0% of the ‘definite’, 87.8% of the ‘probable’, and 70.3% of the ‘indeterminate’ HAIs belonged to one of the twenty-two COVID-19 genomic clusters we identified. Conclusions: We propose a novel prediction model for HAIs. In addition, we show that the management of nosocomial outbreaks will benefit from genome sequencing analyses.

Surfing the Waves: Differences in Hospitalised COVID-19 Patients across 4 Variant Waves in a Belgian University Hospital

“…Although sequencing capacity increased with time in our centre, we chose not to use individual variant identification, as this would have led to excessive missing data. We are confident that the Belgian national surveillance data reflects our catchment area’s epidemiology, as shown in another project published by our group recently [ 29 ]. We were also reassured to read a recent paper by Meurisse et al [ 11 ], who found that exposure ascertainment based on time periods leads to the most precise results when studying severity per variant wave.…”