AN ADVANCED COMPUTATIONAL MODEL FOR EARLY SEVERITY STRATIFICATION OF COVID-19 PATIENTS IN EMERGENCY CARE SETTINGS

Abstract

This thesis examines the development and clinical implementation of a mathematical model designed to predict the severity of COVID‑19 using demographic, clinical, and laboratory parameters. The study analyzes 1,145 hospitalized patients and compares traditional physician‑based stratification with an automated program that assigns a probabilistic severity index. The system significantly reduced stratification errors across all severity categories and improved patient routing, enabling earlier initiation of intensive therapy for high‑risk individuals. The expanded text further explores the integration of machine‑learning methods, temporal data analysis, and cross‑pathology application. The findings highlight the clinical value of computational tools in standardizing triage and enhancing decision‑making efficiency during pandemics.