Background: Growing evidence has documented the disproportionate impact of COVID-19 across racial, economic, and clinical risk factors, but the extent to which these patterns hold during and after hospitalization is not as well understood. Analysis of these in-hospital determinants and temporal trends is needed to inform further health system responses. 

Methods: We examined outcomes in 2500 individuals hospitalized with COVID-19 in 5 hospitals in the University of Pennsylvania Health System by September 17, 2020 using electronic health records to assess outcomes through 8 weeks post-admission. Hospital discharge and mortality were analyzed as competing risks using a multivariable cause-specific hazards model with patient socio-demographics, comorbidities, disease severity at admission, treating hospital, and month of admission as independent risk factors.

Results: Morality decreased markedly over time, with cumulative incidence (95% CI) 30 days post-admission of 19.1% (17.2, 21.3) in March-April versus 6.3% (4.3, 8.9) in July-September; 19% of deaths occurred after discharge. Compared to Caucasian, Black race was associated with more severe disease at admission, prevalent co-morbidities and residence in low-income zip code. In multivariable models, there were no detectable differences in mortality risk by race, while treating hospital, increasing age, admission early in the pandemic and severe disease at admission were associated with increased mortality hazard.

Conclusions: Mortality for hospitalized COVID-19 patients decreased over time, but post-discharge mortality was non-trivial. Poor outcomes were associated with advanced age, disease severity at admission, and treating hospital. After adjusting for baseline risk factors, we found no clear association between race and mortality.