Correlates of In-Hospital COVID-19 Deaths: A Competing Risks Survival Time Analysis of Retrospective Mortality Data

INTRODUCTION: Several aspects of the coronavirus disease 2019 (COVID-19) pandemic remain ambiguous, including its transmission, severity, geographic, and racial differences in mortality. These variations merit elaboration of local patterns to inform wider national policies. METHODS: In a retrospective analysis, data of patients treated at a dedicated COVID hospital with moderate and severe illness during 8 wk of the pandemic were reviewed with attention to mortality in a competing risks framework. RESULTS: A total of 1147 patients were hospitalized, and 312 (27.2%) died in hospital. Those who died were older (56.5 vs 47.6 y; P < 0.0001). Of these, 885 (77.2%) had tested positive on reverse transcriptase polymerase chain reaction (RT-PCR), with 219 (24.2%) deaths (incidence rate, 1.9 per 100 person-days). Median time from onset of symptoms to death was 11 days. A competing risks analysis for in-hospital death revealed an adjusted cause-specific hazard ratio of 1.4 for each decade increase in age. CONCLUSIONS: This retrospective analysis provides broad patterns of disease presentation and mortality. Even COVID test-negative patients will receive treatment at dedicated facilities, and 33% presenting cases may die within the first 72 h, most with comorbid illness. This should be considered while planning distribution of services for effective health-care delivery