Development and validation of models for two‐week mortality of inpatients with COVID‐19 infection: A large prospective cohort study

Abstract

Recognizing COVID‐19 patients at a greater risk of mortality assists medical staff to identify who benefits from more serious care. We developed and validated prediction models for two‐week mortality of inpatients with COVID‐19 infection based on clinical predictors. A prospective cohort study was started in February 2020 and is still continuing. In total, 57,705 inpatients with both a positive reverse transcription‐polymerase chain reaction test and positive chest CT findings for COVID‐19 were included. The outcome was mortality within 2 weeks of admission. Three prognostic models were developed for young, adult, and senior patients. Data from the capital province (Tehran) of Iran were used for validation, and data from all other provinces were used for development of the models. The model Young, was well‐fitted to the data (p < 0.001, Nagelkerke R2 = 0.697, C‐statistics = 0.88) and the models Adult (p < 0.001, Nagelkerke R2 = 0.340, C‐statistics = 0.70) and Senior (p < 0.001, Nagelkerke R2 = 0.208, C‐statistics = 0.68) were also significant. Intubation, saturated O2 < 93%, impaired consciousness, acute respiratory distress syndrome, and cancer treatment were major risk factors. Elderly people were at greater risk of mortality. Young patients with a history of blood hypertension, vomiting, and fever; and adults with diabetes mellitus and cardiovascular disease had more mortality risk. Young people with myalgia; and adult patients with nausea, anorexia, and headache showed less risk of mortality than others.