Using GAM functions and Markov-Switching models in an evaluation framework to assess countries' performance in...

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• COVID-19 pandemic necessitates accurate monitoring of its global evolution.
• Cross-country comparisons of COVID-19 data can be misleading due to diverse geopolitical and socio-economic factors.
• Existing monitoring methods may not adequately capture pandemic dynamics like spikes and structural breaks.

Purpose of the Study:

• To develop a new evaluation framework for monitoring and assessing COVID-19 evolution across countries.
• To provide stakeholders with tools to adequately assess pandemic control measures.
• To identify factors influencing the spread and mortality rates of COVID-19.

Main Methods:

• Generalized Additive Models (GAMs) for modeling cumulative and daily COVID-19 case and death curves.
• Local min-max function for identifying local maxima in time series data.
• Markov-Switching and Markov-Switching GARCH models for detecting structural breaks in COVID-19 time series.
• A quadrant system for evaluating country-level efficiency in controlling virus spread and deaths.

Main Results:

• The proposed methodology effectively explains COVID-19 evolution, outperforming traditional growth functions.
• Identified inflection points and regime-switching moments linked to public commitment levels.
• Europe exhibits a high number of countries with upward trends in cases and deaths, indicating inefficiency.

Conclusions:

• The developed framework enables stakeholders to evaluate the effectiveness of public policies and interventions against COVID-19.
• Facilitates the identification of best practices in pandemic control for wider adoption.
• Supports evidence-based decision-making to improve global pandemic response strategies.