Related Experiment Video
Updated: Oct 20, 2025
05:34
Author Spotlight: Advancing Pathogen Diagnostics with Standardized LAMP
Published on: September 8, 2023
958
Early outcome detection for COVID-19 patients
Alina Sîrbu1, Greta Barbieri2, Francesco Faita3
1Department of Computer Science, University of Pisa, Pisa, Italy. alina.sirbu@unipi.it.
Scientific Reports
|September 17, 2021
Summary
This study developed a predictive model to identify COVID-19 patient mortality risk using six key clinical variables. The model achieved over 85% accuracy, aiding pandemic management.
Area of Science:
- Medical Informatics
- Public Health
- Epidemiology
Background:
- The COVID-19 pandemic strained healthcare systems globally.
- Clinical decision support systems (CDSS) are crucial for pandemic management.
- Predictive models can enhance patient care and resource allocation.
Purpose of the Study:
- To develop and validate a predictive model for COVID-19 patient mortality.
- To identify key clinical variables with high predictive power for mortality.
- To create a robust model applicable to diverse patient cohorts and data quality.
Main Methods:
- Utilized a feature selection method based on genetic algorithms to identify significant clinical variables.
- Developed a predictive model using data from the first wave of COVID-19 patients.
- Validated the model on a new cohort from the second wave, including patients with imputed missing values.
Main Results:
- Identified six clinical variables with the largest predictive power for COVID-19 mortality.
- The final predictive model demonstrated accuracy exceeding 85% on test data.
- The model performed well on new patient cohorts and with imputed data, showing resilience to missing values.
Conclusions:
- The developed predictive model effectively forecasts COVID-19 patient mortality.
- The identified clinical variables are crucial indicators of mortality risk.
- This CDSS tool can support clinical decision-making and improve pandemic management strategies.
