Spatial accessibility to healthcare and wide-area evacuation following the 2024 Noto Peninsula Earthquake: a GIS-based cross-sectional study

Wide-area evacuation beyond disaster-affected regions can disrupt continuity of care and increase health risks, particularly for older adults and individuals requiring long-term care. However, the community-level factors associated with such evacuation remain poorly understood. This retrospective observational study examined characteristics associated with wide-area evacuation following the 2024 Noto Peninsula Earthquake, Japan. Aggregated and anonymized human mobility data at the 250-m mesh level were used to identify population displacement outside the affected region during January 2024. Explanatory variables included demographic characteristics, spatial accessibility indicators (distance to evacuation shelters and healthcare-related facilities), and measures of earthquake-related damage. Multivariable logistic regression analysis was performed to evaluate independent associations with wide-area evacuation. Among 375 analyzed meshes, wide-area evacuation occurred in 27 (7.2%). Greater distance to evacuation shelters, shelters with dispatched nurses, hospitals, and home-visit nursing care offices was independently associated with wide-area evacuation, whereas demographic characteristics, including population aging and estimated prevalence of long-term care needs, were not. Residential building damage was also not independently associated. These findings indicate that spatial accessibility to essential services, rather than demographic vulnerability alone, plays a key role in shaping evacuation patterns, highlighting the importance of incorporating structural accessibility into disaster preparedness and response planning.