Individualized prediction nomograms for disease progression in mild COVID‐19

Abstract

Background The coronavirus disease 2019 (COVID‐19) has evolved into a pandemic rapidly. The majority of COVID‐19 patients are with mild syndromes. This study aimed to develop models for predicting disease progression in mild cases. Methods The risk factors for the requirement of oxygen support in mild COVID‐19 were explored using multivariate logistic regression. Nomogram as visualization of the models was developed using R software. Results A total of 344 patients with mild COVID‐19 were included in the final analysis, 45 of whom progressed and needed high‐flow oxygen therapy or mechanical ventilation after admission. There were 188 (54.7%) males, and the average age of the cohort was 52.9 ± 16.8 years. When the laboratory data were not included in multivariate analysis, diabetes, coronary heart disease, T >=38.5℃ and sputum were independent risk factors of progressive COVID‐19 (Model 1). When the blood routine test was included the CHD, T >=38.5℃ and NLR were found to be independent predictors (Model 2). The AUROC of model 2 was larger than model 1 (0.872 vs. 0.849, P =0.023). The negative predictive value of both models was greater than 96%, indicating they could serve as simple tools for ruling out the possibility of disease progression. Conclusions In conclusion, two models comprised common symptoms (fever and sputum), underlying diseases (diabetes and coronary heart disease) and blood routine test are developed for predicting the future requirement of oxygen support in mild COVID‐19 cases.