Mortality Due to Covid-19 in Hospitalized Patients: A Prediction Model Based on Different Risk Factors.

Abstract

Purpose: Since 2020, COVID-19 severely affected the world population, generating numerous deaths and a great socioeconomic impact that affected the healthcare system. This investigation aimed to analyze a prediction model for COVID-19 mortality on the basis of different risk factors.

Patients and methods: Retrospective, cross-sectional study in a sample of 2000 hospitalized patients. Biological and clinical factors (signs and symptoms), laboratory/diagnostic results and comorbidities were taken into account. The SPSS version 29 statistical package was used to process the information, performing a bivariate and multivariate analysis with binary logistic regression using the intro methods.

Results: Most of the deceased were male, older than 60 years, blood type O positive, hypertensive, type 2 diabetic, obese. The most common symptoms were fever, malaise, shortness of breath and fatigue, the most common tomography findings were bilateral ground glass with BiRad 5 scale in more seriously impaired patients.

Conclusion: An adequate model was obtained with a 76% prognostic rate. The variables included in the predictive model for COVID-19 mortality were age, fever, productive cough, sore throat, fatigue, shortness of breath, unilateral consolidation on CT scan, hemoglobin level, leucocyte count, lymphocytes, platelets, urea, and ferritin.