A simple point score model for prediction of Covid-19 in some Egyptian patients

Abstract

On a global scale, a substantial number of newly diagnosed cases of the covid-19 virus and a considerable number of associated fatalities are recorded weekly. Its laboratory detection depends on the costly and time-consuming real-time PCR analysis. A simple way to facilitate the diagnosis of Covid-19 is still required. Here, it was aimed to generate a simple point score as a prediction model for fast diagnosis of Covid-19 using simple laboratory analyses. 121 adult individuals with qRT-PCR results served as a training group, whereas 35 individuals were used as a validation group. Different laboratory analyses, including complete blood count (CBC), differential count, D-dimer, C-Reactive Protein (CRP), and Ferritin, have been recruited as predictors using the Receiver Operating Characteristic (ROC) analysis. The results revealed three models, depending on the predictor parameters' ROC area (AUC). The simplest model consisted of the data of the three predictors: lymphocytopenia, CRP, and D-dimer, and resulted in a ROC AUC value of 0.9773. The use of the three models on the validation group provided support for the conclusion that the calculation of lymphocyte count, CRP, and D-dimer is enough for predicting the occurrence of Covid-19.