COVID-19 ICU Prediction

Studies that explicitly show how to use repeatedly measured predictors to get and update predictions for future binary or continuous outcomes are lacking. Examples of patients’ biomarkers that are repeatedly measured over time during follow-up visits include CD4 count, number of lymph nodes positive for tumor cells, creatinine level, IgG titers, white blood cell (WBC), neutrophil, lymphocyte counts and many other. Examples of future events include death, disease progression, disease recurrence, admission to ICU, or development of a new medical condition. There are different methods to harness the value of repeated measurements of predictor variables for clinical risk prediction, and each of them has its own advantages and limitations. The objective of this study is to review the current epidemiology literature for the application of various statistical approaches for prediction of binary and continuous outcomes based on repeatedly measured predictors. Secondly, this study aims to apply and compare the two-stage method and joint modeling approach to predict admission to ICU of patients hospitalized for Covid 19 based on repeated biomarkers.

Edward De Vol, Jameela Edathodu.