RESUMO
Purpose: To assess survival patterns and predictors of mortality among patients admitted with COVID-19 to treatment centers in the Oromia region of Ethiopia from April 1 to August 31, 2021. Methods: A prospective cohort study design was employed, taking a sample of 854 patients selected from eight treatment centers in the region. The follow-up duration was the time interval from admission to the treatment center until the final disposition of patients at discharge (death, recovery, or failed to recover). Data were collected by computer tablet with an interviewer-administered questionnaire and checklist designed using CSPro 7.5 and exported to Stata 13 for analysis. Descriptive analysis was used to explore the characteristics of patients. The mortality rate was estimated by number of deaths per 1,000 person-days of observation. The survival duration was estimated by medians with IQR. The Kaplan-Meier method was used to compare the survival experiences of patients. To identify the predictors of time to death after hospitalization, a Cox proportional-hazard model was used. The magnitude of association was estimated using HRs with 95% CIs, and statistical significance was set at P<0.05. Results: The mortality rate among hospitalized patients was 9.9 per 1,000 person-days of observation and the median survival time after admission was 9 (IQR 9-10) days. Higher hazard of death was observed among patients who drank alcohol (AHR 2.0, 95% CI 1.2-3.3), required anticoagulants (AHR 10, 95% CI 1.2-91.5), glucocorticoids (AHR 1.7, 95% CI 1.1-2.8), and oxygen (AHR 4.7, 95% CI 1.1-22.0), those with acute respiratory distress syndrome (AHR 2.9, 95% CI 1.7-5.1), and critical patients admitted to intensive care units (AHR 3.4, 95% CI 2.0-5.9). Conclusion: The hazard of death is significantly predicted by alcohol use, requiring anticoagulants, glucocorticoids, or oxygen medication, acute respiratory distress syndrome complication, and being critical when admitted to intensive care units.