Data-driven modeling and forecasting of COVID-19 outbreak for public policy making

Area of Science:

• Epidemiology
• Mathematical Modeling
• Public Health

Background:

• COVID-19 pandemic necessitates effective control strategies.
• Non-Pharmaceutical Interventions (NPIs) are crucial for outbreak management.
• Accurate forecasting models are vital for policy decisions.

Purpose of the Study:

• To develop a data-driven model for COVID-19 forecasting.
• To estimate the time-varying effective reproduction number (R_t) using real-world data.
• To assess the impact of NPIs on disease transmission and forecast future scenarios.

Main Methods:

• Application of an extended Kalman filter (EKF) to a discrete-time stochastic compartmental model.
• Utilizing daily confirmed cases, active cases, recovered cases, deceased cases, Case-Fatality-Rate (CFR), and infectious time as model inputs.
• Definition of a Transmission Index (TI) based on R_t to quantify intervention effectiveness.

Main Results:

• The model successfully estimates R_t and provides a TI for assessing transmission.
• Forecasts demonstrate the potential impact of adjusting NPIs (e.g., physical distancing, lockdowns).
• Case studies in three countries validate the model's practicability.

Conclusions:

• The proposed data-driven approach offers a valuable tool for public policy and decision-makers.
• The model facilitates scenario forecasting to inform the relaxation or tightening of public health measures.
• This methodology aids in controlling COVID-19 outbreaks through evidence-based NPI strategies.