Abstract
BACKGROUND: Studies have shown an association between the clinical severity of coronavirus disease (COVID-19) and sociodemographic and clinical variables in older adults. However, few studies have described the explanatory factors of the relationship between these variables and the clinical severity of COVID-19 using structural equation modeling. OBJECTIVE: To analyze the factors directly and indirectly associated with the clinical severity of coronavirus disease (COVID-19) among older adults in Minas Gerais, Brazil. DESIGN AND SETTING: Retrospective epidemiological study. METHODS: This study included 51,141 elderly adults with COVID-19 living in Minas Gerais, Brazil. Data were collected through the Individual Registration Form - Hospitalized Cases of Severe Acute Respiratory Syndrome from January 28, 2020, to January 27, 2022. RESULTS: Older age (P < 0.001), male sex (P < 0.001), dyspnea (P < 0.001), change in chest X-ray examination findings (P < 0.001), greater number of risk factors/comorbidities (P < 0.001), and longer hospitalization time (P < 0.001) were directly associated with the clinical severity of COVID-19. Female sex, mediated by the greater number of risk/comorbidity factors (β = -0.02, P < 0.001), and younger age, mediated by longer hospitalization time (β = -0.01; P < 0.001), were indirectly associated with the clinical severity of COVID-19. CONCLUSION: Demographic and clinical variables were directly associated with increased disease severity. In addition to the direct effect, a greater number of risk/comorbidity factors and longer hospitalization time mediated the association between demographic variables and outcomes.