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

Area of Science:

• Infectious Diseases
• Clinical Medicine
• Biostatistics

Background:

• Coronavirus disease 2019 (COVID-19) poses a significant global health threat.
• Severe cases of COVID-19 have poor clinical outcomes, necessitating early identification.
• Distinguishing severe from common COVID-19 is crucial for effective patient management.

Purpose of the Study:

• To establish and validate an early warning nomogram model for predicting severe COVID-19.
• To identify key clinical factors associated with severe COVID-19 progression.
• To aid clinicians in early and timely treatment decisions for COVID-19 patients.

Main Methods:

• Analysis of 1059 COVID-19 patients in a primary cohort and 123 in a validation cohort.
• Logistic regression analysis to identify independent risk factors for severe COVID-19.
• Construction and performance evaluation of a nomogram model using identified risk factors.

Main Results:

• Multivariate analysis identified age, dyspnea, lymphocyte count, C-reactive protein (CRP), and interleukin-6 (IL-6) as independent predictors of severe COVID-19.
• The nomogram model demonstrated strong predictive performance with a C-index of 0.863 in the primary cohort and 0.889 in the validation cohort.
• Calibration curves confirmed good agreement between predicted and actual probabilities in both cohorts.

Conclusions:

• An early warning nomogram model incorporating age, dyspnea, lymphocyte count, CRP, and IL-6 can effectively predict severe COVID-19.
• This model facilitates early identification of patients at risk for severe disease.
• Timely intervention based on nomogram predictions can potentially improve clinical outcomes for COVID-19 patients.