Abstract

Avian influenza (AI) is a highly contagious disease that causes significant mortality and economic losses in the poultry industry. This study integrates meta-analysis with spatial analysis, specifically Ordinary Least Squares (OLS), Geographically Weighted Regression (GWR), and Multiscale Geographically Weighted Regression (MGWR), to identify high-risk areas for AI in Thailand. The independent variables were selected through a systematic meta-analysis, with data sourced from secondary empirical datasets. The dependent variable was based on previous HPAI outbreak data in Thailand. The findings indicate that the central and lower northern regions of Thailand are high-risk areas, strongly linked to rice paddy fields and elevated poultry density. Model comparisons demonstrate that MGWR, utilizing the neighbor-based approach, outperforms both OLS and GWR, achieving higher predictive accuracy (adjusted R² = 0.96, AUC = 0.89, 95 % CI = 0.87–0.91). MGWR's ability to assign variable-specific spatial scales enhances its capacity to capture spatial heterogeneity. The integration of these methodologies enhances predictive modeling accuracy, allowing authorities to design more efficient surveillance systems and implement focused prevention strategies to reduce the public health risks of avian influenza.