Survival Follow-up and Mortality Prediction Model for Elderly Severe COVID-19 Patients following the complete relaxation of pandemic restrictions in China

The global outbreak of the novel coronavirus has been a significant public health crisis in recent years. Elderly patients with severe COVID − 19 are a special group with extremely high morbidity and mortality rates. Early identification of important factors influencing prognosis and implementation of intervention measures can significantly improve the prognosis of elderly patients with severe COVID − 19.A retrospective cohort study was conducted on elderly patients (≥ 60 years) with severe COVID-19 admitted to the emergency room and Emergency Intensive Care Unit (EICU) of North Jiangsu People's Hospital affiliated with Yangzhou University between December 2022 and January 2023. Patients were randomly divided into training and validation sets (8:2 ratio). Variable selection was performed using LASSO and multifactorial stepwise backward Cox regression. Cox regression was then utilized to construct survival probability plots at 1-month and 1-year. Model performance was assessed using calibration curves, time-dependent ROC curves, and decision curve analysis (DCA). The study included 199 elderly patients with severe COVID-19. The mean survival time in the mortality group was 80.092 days, with a median survival time of 22.000 days. The analysis identified gender (male), lung consolidation, age, lymphocyte ratio, lactate levels, and endotracheal intubation as factors associated with poorer survival prognosis (HR > 1), while gender (female), hospitalization, and oxygenation index were associated with better prognosis (HR < 1). A prediction model based on these factors demonstrated good predictive performance, with a C-index of 0.842 (training set) and 0.809 (validation set). The AUC values for the model were 0.920 (training set) and 0.930 (validation set) for 1-month mortality prediction, and 0.941 (training set) and 0.947 (validation set) for 1-year mortality prediction, indicating strong diagnostic accuracy. Calibration curves showed good model fit. DCA indicated the model's clinical utility. Clinical interventions targeting elderly severe COVID-19 patients, optimizing pulmonary infection control, enhancing oxygenation and circulation, and improving immune function (lymphocyte counts) may significantly improve outcomes in critically ill patients with severe COVID-19.