论文推荐 | 基于深度学习方法评估考虑救护车通行能力的老年人城市紧急医疗服务可达性

题目：Assessing urban emergency medical services accessibility for older adults considering ambulance trafficability using a deep learning approach

（基于深度学习方法评估考虑救护车通行能力的老年人城市紧急医疗服务可达性）

作者：Zhuo Liu, Enjia Zhang*, Shuo Pan, Sichun Li, Ying Long, Frank Witlox

发表刊物：Sustainable Cities and Society

DOI：

Rapid urbanization and population aging have made equitable Emergency Medical Services (EMS) access for older adults a critical challenge in high-density cities. This study develops a deep learning framework to evaluate EMS accessibility, considering ambulance trafficability derived from street view images (SVIs). A Multi-Scale Vision Transformer (MSViT) model classifies SVIs into impassable, narrow, and passable categories to assess road conditions. Travel speeds are then assigned based on the classified trafficability and road hierarchy data. The framework measures the pre-hospital time through two-stage travel (emergency center-to-patient and patient-to-hospital), while evaluating accessibility through two complementary metrics: the inverse of the shortest total pre-hospital time, and a composite accessibility score with Gaussian time-decay weighting for all facilities within 15-minute service ranges. A case study in Beijing’s Old City demonstrated that considering ambulance trafficability reduces the estimated 15-minute coverage from 94.18 % to 83.14 %, revealing significant overestimation when road conditions are neglected. Spatially, peripheral areas achieved better nearest-facility response times, whereas central regions dominated in comprehensive service coverage. Additionally, EMS accessibility patterns strongly correlated with older adults’ distribution, showing limited income-based disparities. This study shifts the focus from supply-demand balancing to road system management in EMS accessibility, which can be integrated with existing methods to support more sustainable and targeted infrastructure optimization in aging cities.

论文展示（部分）