Estimation of COVID-19 spread curves integrating global data and borrowing information

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• The COVID-19 pandemic poses a significant global health threat, with limited approved treatments and scattered information.
• Effective prediction of infection spread is crucial for public health response and resource allocation.

Purpose of the Study:

• To develop a data integration approach for real-time prediction of COVID-19 infection trajectories across multiple countries.
• To create a robust predictive model that quantifies uncertainty and identifies potential risk factors for severe COVID-19.

Main Methods:

• A Bayesian hierarchical model was developed to integrate global COVID-19 data.
• The model utilizes information sharing across countries to enhance predictive accuracy.
• Joint variable selection was incorporated to identify country-level risk factors for severe disease.

Main Results:

• The proposed model demonstrated superior performance compared to individual country-based models.
• The Bayesian approach provides uncertainty quantification for predictions.
• The integrated variable selection identified potential risk factors associated with severe COVID-19 outcomes.

Conclusions:

• Bayesian hierarchical modeling offers a powerful tool for real-time COVID-19 infection prediction and risk factor identification.
• Data integration across countries improves predictive accuracy and provides a unified view of the pandemic.
• This approach aids in understanding and mitigating the impact of COVID-19 globally.