Risk Estimation of Severe COVID-19 Based on Initial Biomarker Assessment Across Racial and Ethnic Groups

Area of Science:

• Medical research
• Infectious disease epidemiology
• Biomarker discovery

Background:

• COVID-19 presents a wide spectrum of disease severity.
• Early prediction of disease severity is essential for appropriate patient management.
• Limited data exists on predictive factors at the time of initial diagnosis.

Purpose of the Study:

• To develop predictive models for COVID-19 severity.
• Utilize demographic, clinical, and laboratory data from initial patient encounters.
• Identify key predictors for distinguishing severe from mild COVID-19 outcomes.

Main Methods:

• Retrospective analysis of 14,147 COVID-19 patients' deidentified data.
• Backward stepwise logistic regression modeling.
• Inclusion of demographic, clinical, and laboratory variables at diagnosis.

Main Results:

• Four proficient predictive models for COVID-19 severity were developed.
• Key predictors consistently identified across models include age, albumin, diastolic blood pressure, ferritin, lactic dehydrogenase, socioeconomic status, procalcitonin, B-type natriuretic peptide, and platelet count.
• 18% of patients experienced severe outcomes, and 24% experienced mild outcomes.

Conclusions:

• Biomarkers identified in specific and sensitive models are valuable for initial COVID-19 severity assessment.
• These findings can assist healthcare providers in early evaluation and management decisions.
• The study highlights the utility of readily available data for predicting disease trajectory.