Mortality amongst hospitalized COVID-19 cases by acquisition and pandemic wave in Wales, UK, February 2020–March 2022

Abstract

Hospital populations are vulnerable to COVID-19, but the relative severity of hospital acquisition compared to community is unknown. We investigated differences in mortality between hospital and community acquired cases in Wales. We linked SARS-CoV-2 PCR tests from Feb 2020-Mar 2022 to hospital admissions, to identify likely nosocomial acquisition. We measured death from any cause 28-days post positive specimen (mortality), by acquisition. We used multivariable logistic regression to compare mortality by acquisition, adjusting for confounders, computing adjusted odds ratio (aOR) with 95% confidence intervals (95%CI). There were 25,263 COVID-19 hospital cases and 5490 (22%) deaths in the study period. Although significant in univariate analysis, adjustment for confounding showed no association with increased mortality in nosocomial cases compared to cases admitted with COVID-19 (aOR= 0.8, 95%CI= 0.7-0.8). Vaccination (aOR= 0.6, 95%CI= 0.5-0.7) and infection in later pandemic waves (aOR= 0.5, 95%CI= 0.4-0.6) were associated with lower mortality; older age (85+ vs <25 years: aOR= 76.4, 95%CI= 41.8-160.5) and male sex (aOR= 1.5, 95%CI= 1.4-1.6) were associated with higher mortality. One in five hospitalised COVID-19 cases died within a month of infection. Mortality in nosocomial cases was not worse than those admitted with COVID-19, maybe reflecting early identification of nosocomial cases through screening.