A Predicting Nomogram for Mortality in Patients With COVID-19

Area of Science:

• Medical research
• Infectious diseases
• Critical care medicine

Background:

• The COVID-19 pandemic poses a severe global health threat, with significant mortality rates.
• Accurate prediction of mortality is crucial for effective patient management and resource allocation.

Purpose of the Study:

• To develop and validate a reliable nomogram for predicting individual mortality risk in COVID-19 patients.
• To provide a clinical tool for early identification of patients requiring intensive care and close monitoring.

Main Methods:

• A retrospective single-center study involving 120 COVID-19 patients (21 deceased, 99 discharged) for nomogram construction.
• Validation of the nomogram using an independent cohort of 84 patients.
• Multivariable logistic regression analysis was employed to identify predictors and construct the model, followed by evaluation of calibration, differentiation, and clinical usefulness.

Main Results:

• The developed nomogram incorporates C-reactive protein, PaO2/FiO2 ratio, and cardiac troponin I (cTnI) as key predictors of mortality.
• The nomogram demonstrated high predictive accuracy with Areas Under the Curve (AUC) of 0.988 in the primary group and 0.956 in the validation group.
• Decision curve analysis indicated potential clinical utility for the nomogram in patient stratification.

Conclusions:

• A convenient and effective nomogram utilizing C-reactive protein, PaO2/FiO2, and cTnI has been developed to predict COVID-19 patient mortality.
• This tool can assist clinicians in assessing individual patient risk and guiding treatment decisions.
• Future research will focus on multi-center data collection to enhance the nomogram's reliability and generalizability.