Severe versus common COVID-19: an early warning nomogram model

The wide spread of coronavirus disease 2019 is currently the most rigorous health threat, and the clinical outcomes of severe patients are extremely poor. In this study, we establish an early warning nomogram model related to severe versus common COVID-19. A total of 1059 COVID-19 patients were analyzed in the primary cohort and divided into common and severe according to the guidelines on the Diagnosis and Treatment of COVID-19 by the National Health Commission of China (7th version). The clinical data were collected for logistic regression analysis to assess the risk factors for severe versus common type. Furthermore, 123 COVID-19 patients were reviewed as the validation cohort to assess the performance of this model. Multivariate logistic analysis revealed that age, dyspnea, lymphocyte count, C-reactive protein and interleukin-6 were independent factors for prewarning the severe type occurrence. Then, the early warning nomogram model including these risk factors for inferring the severe disease occurrence out of common type of COVID-19 was constructed. The C-index of this nomogram in the primary cohort was 0.863, 95% confidence interval (CI) (0.836–0.889). Meanwhile, in the validation cohort, the C-index of this nomogram was 0.889, 95% CI (0.828–0.950). In both the primary cohort and validation cohorts, the calibration curve showed good agreement between prediction and actual probability. The early warning model shows that data at the very beginning including age, dyspnea, lymphocyte count, CRP, and IL-6 may prewarn the severe disease occurrence to some extent, which could help clinicians early and timely treatment.