Competing risk survival analysis of time to in-hospital death and recovery among COVID-19 patients in Hawassa University Comprehensive Speciality Hospital, Hawassa, Ethiopia: a retrospective cohort study

The impact of various factors and complications on COVID-19 outcomes remains a significant public health challenge. In Sidama, Ethiopia, there is limited data on COVID-19 mortality risk factors, and no previous studies have employed competing risk approaches to analyze this data. This study aims to identify determinants of in-hospital mortality among COVID-19 patients, explicitly considering recovery as a competing event. A retrospective cohort study was conducted on 804 patients treated at Hawassa University Comprehensive Specialized Hospital between September 2020 and November 2021. Data were extracted from electronic medical records. To account for the competing nature of the outcomes (death vs. recovery), we applied both cause-specific hazard (CSH) and sub-distribution hazard (SDH) regression models. Among the 804 patients, 21.5% died while 74.0% fully recovered. The median age at death was 60 years, compared to 40 years for those who recovered. The median hospital stay was 5 days for deceased patients and 12 days for those who recovered. In both the CSH and SDH models, factors such as high disease severity on admission, older age, and diabetes were associated with an increased risk of mortality and a decreased likelihood of recovery. Early detection of symptoms like shortness of breath and body weakness was associated with a reduced risk of death. This study underscores the importance of employing appropriate competing risk models in COVID-19 prognosis. Traditional methods, such as the Kaplan-Meier estimator, tend to overestimate the risk of an event. Using the cumulative incidence function (CIF) alongside sub-distribution hazard models provides more accurate estimates of competing events, such as recovery and death. These methods improve prognostic accuracy and can inform better clinical and public health decisions for managing COVID-19. The study underscores the significance of competitive risk analysis in COVID-19 patient care, emphasizing early symptom diagnosis to reduce morbidity and improve quality of life, and controlling comorbidities for faster recovery. Since many survival studies do not employ competing risk analysis, it offers methods for analyzing and interpreting survival data in the face of conflicting events in low-resource settings. It focuses on in-hospital patient studies, which are rarely taken into account in settings with limited resources.