Development and external validation of a COVID-19 mortality risk prediction algorithm: a multicentre retrospective...

Area of Science:

• Medical Informatics
• Epidemiology
• Public Health

Background:

• The COVID-19 pandemic necessitated accurate tools for predicting patient mortality.
• Risk stratification is crucial for effective management of severe cases.

Purpose of the Study:

• To develop and externally validate a predictive algorithm for COVID-19 mortality.
• To provide a tool for acute risk classification of COVID-19 patients.

Main Methods:

• A retrospective cohort study involving 1364 adult COVID-19 patients from Hubei province, China.
• Development of two models (full and simple) using data from 1088 patients and external validation on 276 patients.
• Assessment of model discrimination using AUC and calibration using calibration plots and decision curve analysis.

Main Results:

• The full model achieved an AUC of 0.96 in the derivation cohort and 0.97 in the validation cohort.
• The simple model achieved an AUC of 0.92 in the derivation cohort and 0.88 in the validation cohort.
• Both models demonstrated good calibration accuracy in both cohorts.

Conclusions:

• The developed prediction models effectively identify COVID-19 patients at high risk of 60-day mortality.
• These models show potential utility for acute risk classification and clinical decision-making.
• A freely accessible web calculator is available for practical application.