Open AccessRisk Identification and Prediction for COVID-19 Mortality
Author(s) -
H. Bryant Nguyen,
Qin Shao
Publication year - 2021
Publication title -
translation
Language(s) - English
Resource type - Journals
ISSN - 2469-6706
DOI - 10.46570/utjms.vol9-2021-462
Subject(s) - case fatality rate , logistic regression , receiver operating characteristic , mortality rate , covid-19 , medicine , outbreak , statistics , outcome (game theory) , identification (biology) , risk of mortality , demography , emergency medicine , medical emergency , epidemiology , mathematics , virology , disease , biology , botany , mathematical economics , sociology , infectious disease (medical specialty)
This paper studies several key metrics for COVID-19 using a public surveillance system data set. It compares the difference between two case fatality rates: the naive case fatality rate, which has been frequently mentioned in media outlets, and one which is the sample estimate for the mortality rate. A logistic regression model is applied to modeling the daily mortality rate. The conclusion is that time, gender, age and some of their interactions, appear to have a significant impact on the mortality rate; the daily mortality rate has been decreasing since the outbreak; males older than 60 has been the most vulnerable group. The receiver operating characteristics curve and the curve under the area show that the proposed logistic model is capable of predicting the outcome of a reported case with accuracy as high as 89%. These findings are helpful in assessing the magnitude of the risk posed by the COVID-19 virus to certain groups, predicting outcome severity, and optimally allocating medical resources such as intensive care units and ventilators.