The increasing recognition of the health benefits of blue spaces highlights their crucial role in constructing Healthy Cities and advancing Sustainable Development Goals. Given that promoting recreational running represents a fundamental pathway to harnessing these benefits, integrating it into spatial planning and design is imperative. Nevertheless, this integration process necessitates substantiated evidence, especially concerning variances among population groups. To address this gap, utilising crowdsourced data and a machine learning approach, this study investigates heterogeneous spatial distributions of recreational running across various age demographics in Rotterdam, with a specific emphasis on visual perceptions and built environments. The mapping results illustrate the varied allure of blue spaces for recreational running, exhibiting a trend of increased clustering in running activities with age, extending beyond the city centre. The outcomes of GWR and spatial regression models indicate significant associations between various visual perception factors and built environment indicators with individual running preferences. Crucially, disparities and spatial heterogeneity are evident in the impacts of different environmental factors on running across age groups. Accordingly, tailored planning strategies and patterns are proposed, informed by age-specific environmental perceptions and preferences, contributing to a deeper understanding of the blue-health mechanism and offering practical insights for creating health-promoting blue spaces.

Background: Unequal access to primary healthcare (PHC) has become a critical issue in global health inequalities, requiring governments to implement policies tailored to communities’ needs and abilities. However, the place-based facility dimension of PHCs is oversimplified in current healthcare literature, and formulating the equity-oriented PHC spatial planning remains challenging without understanding the multiple impacts of community socio-spatial dynamics, particularly in remote areas. This study aims to push the boundary of PHC studies one step further by presenting a nuanced and dynamic understanding of the impact of community environments on the uneven primary healthcare supply. Methods: Focusing on Shuicheng, a remote rural area in southwestern China, multiple data are included in this village-based study, i.e., the facility-level healthcare statistics data (2016–2019), the statistical yearbooks, WorldPop, and Chinese GDP’s spatial distribution data. We evaluate villages’ PHC service capacity using the number of doctors and essential equipment per capita, which are the major components of China’s PHC delivery. The indicators describing community environments are selected based on extant literature and China’s planning paradigms, including town- and village-level factors. Gini coefficients and local spatial autocorrelation analysis are used to present the divergences of PHC capacity, and multilevel regression model and (heterogeneous) difference in difference model are used to examine the driving role of community environments and the dynamics under the policy intervention. Results: Despite the general improvement, PHC inequalities remain significant in remote rural areas. The village’s location, aging, topography, ethnic autonomy, and economic conditions significantly influence village-level PHC capacity, while demographic characteristics and healthcare delivery at the town level are also important. Although it may improve the hardware setting in village clinics (coef. = 0.350), the recent equity-oriented policy attempts may accelerate the loss of rural doctors (coef. = − 0.517). Notably, the associations between PHC and community environments are affected inconsistently by this round of policy intervention. The town healthcare centers with higher inpatient service capacity (coef. = − 0.514) and more licensed doctors (coef. = − 0.587) and nurses (coef. = − 0.344) may indicate more detrimental policy effects that reduced the number of rural doctors, while the centers with more professional equipment (coef. = 0.504) and nurses (coef. = 0.184) are beneficial for the improvement of hardware setting in clinics. Conclusions: The findings suggest that the PHC inequalities are increasingly a result of joint social, economic, and institutional forces in recent years, underlining the increased complexity of the PHC resource allocation mechanism. Therefore, we claim the necessity to incorporate a broader understanding of community orientation in PHC delivery, particularly the interdisciplinary knowledge of the spatial lens of community, to support its sustainable development. Our findings also provide timely policy insights for ongoing primary healthcare reform in China.