Modelo pronóstico de muerte por la COVID-19

Introduction: Tools have been used to predict the risk of COVID-19, but none have evaluated their external validity.Objective: To evaluate the capacity of a model, based on prognostic factors, to predict the risk of dying from COVID-19.Material and Methods: A cohort analytical study was carried out on patients with COVID-19 at the “Carlos Manuel de Céspedes” provincial general hospital in the Bayamo municipality, province of Granma, from January 1st, 2020 to March 31st, 2023. Confirmed COVID-19 patients were selected. Death was defined as the dependent variable and sociodemographic variables were considered as independent variables for evaluation, including: toxic habits, comorbidity, biomarkers, phenotypes, and prognostic scales.Results: The Cox proportional regression model demonstrated that the factors associated with the prognosis of death were: classifying phenotype 5 (HR= 6.41; 95% CI= 1.49 to 13.44; p= 0.015); history of arterial hypertension (HR= 2.01; 95% CI= 1.34 to 2.98; p= 0.001); and the RALE scale at 4 or more points (HR= 1.94; 95% CI=1, 47 to 7.90; p= 0.047). The discriminative capacity of the model (C statistic = 0.891) and its calibration were adequate (X2 = 5.384; p = 0.637).Conclusions: The predictive capacity and calibration of the model were adequate. The model can be used as a clinical and epidemiological surveillance instrument, by staging the risk of dying in the most vulnerable subjects. Introducción: Se han empleado herramientas para pronosticar el riesgo de la COVID-19, pero en ninguna se ha evaluado su validez externa.Objetivo: Evaluar la capacidad de un modelo, basado en factores pronósticos, de predecir el riesgo de morir por la COVID-19.Material y método: Se realizó un estudio de cohorte en pacientes con la COVID-19 en el hospital provincial general “Carlos Manuel de Céspedes” del municipio Bayamo, provincia de Granma, desde el 1ero de enero 2020 hasta el 31 de marzo de 2023. Se seleccionaron a los pacientes confirmados de la COVID-19. Se delimitó como variable dependiente a la muerte y variables independientes a evaluar a las sociodemográficas, hábitos tóxicos, comorbilidad, biomarcadores, fenotipos y escalas pronósticas.Resultados: El modelo de regresión proporcional de Cox, demostró que los factores asociados al pronóstico de muerte fueron: clasificar en fenotipo 5 (HR= 6,41; IC 95 %= 1,49 a 13,44; p= 0,015) el antecedente de hipertensión arterial (HR= 2,01; IC 95 %= 1,34 a 2,98; p= 0,001) y la escala de RALE en 4 ó más p