Abstract
BACKGROUND: In recent years, Japan has frequently experienced natural disasters, including major earthquakes and extreme weather events linked to climate change, forcing many people into sheltered lives. These overcrowded environments are known to increase the risk of life-threatening venous thromboembolism (VTE). The 2024 Noto Peninsula Earthquake (NPE) similarly compelled numerous residents into prolonged shelter living. AIM: This retrospective study aimed to indirectly assess VTE risk due to overcrowding in evacuation shelters using proxy indicators (occupancy rate, operational period) and identify spatial clusters of high-risk populations via geospatial and statistical analysis of open data from Nanao City, Ishikawa Prefecture, following the 2024 NPE, to inform disaster preparedness. METHODS: We analyzed open data on evacuation shelters in Nanao City, including location, capacity, and operational period, obtained from GitHub on June 3, 2024. VTE risk was assessed using three proxy indicators: occupancy rate, operational period, and the geographical concentration of overcrowding. The methods included descriptive statistics, geographic information system-based hotspot analysis, simple linear regression, and a sensitivity analysis of occupancy rates. RESULTS: Hotspot analysis identified a significant geographical cluster of the high-risk VTE population in central Nanao City (99% confidence level). Simple regression analysis showed a positive correlation between the cumulative number of evacuees and the operational period (R² = 0.48), though this trend was heavily influenced by a single significant outlier (R² = 0.35 after exclusion). The sensitivity analysis identified Ishizaki Elementary School (>90% occupancy) as the shelter with the most critical VTE risk due to overcrowding. CONCLUSION: This study demonstrated that VTE risk from overcrowding was significantly concentrated in specific shelters, suggesting a potential for recurrence in future disasters under the current system. Municipalities should reconsider shelter placement and resource allocation to mitigate this risk. The quantitative, geographical, and open-data-based approach used here offers a low-cost, rapid, and widely applicable method for disaster preparedness.