Healthcare Spatial Accessibility Study Based On Analysis Of Road Network Centrality

At present, Chinese Government has concerned it as a crucial issue to promote healthcare service equity. There have already appropriate potential accessibility measures and related framework perform well to evaluate equity of healthcare services, i.e., two-step floating catchment area method (or rather framework), for such measures account for interaction between healthcare service providers and the need. However, the dependence of healthcare service providers is likely to be ignored when considering the accessibility towards healthcare equity. It seems that more attention has been paid to issues of distance decay at the idea of spatial interaction, rather than the city form and structural properties of the road network. With the development in theories and methodologies of complex network, more and more geographers began to notice the impact of spatial network properties on socio-economics. Many studies have found that, urban road network, as the carrier of socio-economic activities, its topological properties and statistical distributions of related measures reflect the distributions and intensity of urban land use, and they have influence on attraction of socio-economic activities. It is meaningful to incorporate topological properties of urban network structure into healthcare service interactions in the methodology of accessibility analysis, for this would benefit equity evaluation of healthcare system. The main content of this paper is to combine network structure analysis originated from complex network methodology with conventional spatial interaction models that stem from geography science, and facilitate the combination in analysis of healthcare service accessibility. We would be honored if our method contribution to the related inter-disciplinary fields. The research contents are as follows:Methods and applications of healthcare service accessibility and complex networks methods applied in geography are systematically reviewed. Among all accessibility methods, two-step floating catchment area (2SFCA) framework is commonly approved and widely used. However, we found it flawed by taking relatively simple and fixed scheme to define the radius of the catchment such as fixed or segmented fixed (better but still not dynamic) travel distance or time. This ignores the healthcare facility impact of its location, such as centrality, since different locations bring different degree of attraction. After examining advantages of disadvantages of current network structural analysis methods used in geographical networks, we found some adapted methods that would be of help. We find oud it would be benefit to adjust the catchment radii with the some physical-distance-considered centrality measures. Then, two methods, Road-Network-Centrality-Concerned Accessibility 2SFCA (RNC-2SFCA) methods and Topological-Distance-based Accessibility 2SFCA (TD-2SFCA) together with an Actual-Accessibility-based Index System (AAIS) are introduced. In RNC-2SFCA, we first construct the urban road network using the primal approach, based on 2SFCA, and then calculate the centralities of the road intersections and applied them to determine the radius of catchment of healthcare facility in a 2SFCA framework. RNC-2SFCA is an improved 2SFCA method. In TD-2SFCA, we alternatively contruct the urban road network using the dual approach, and determine the catchment radius by certain topological distance. This approach is used to explore the role of physical distance in healthcare service accessibility. In AAIS, population-based indexes and hospital-based indexes are introduced to analysis the healthcare service utilization data, by facilitate the methods to the geographical units same as planning ones to evaluate the current healthcare resource allocation and planning schedule; moreover, AAIS would be useful to evaluate the outcome of potential accessibility methods. The main work includes:(1) to define an exponential function as distance decay function in the healthcare services accessibility method by facilitating the discharge records of Shenzhen hospitals; (2) to generate community level population-weighted centroids (relative population) using multiple GIS spatial analyzing tools, which serve as locations (quantity) of the demands in healthcare service accessibility method; (3) to construct urban road network with the primal approach, and to simulate a log-form function of catchment radius from evidence, and finally to calculate the accessibilities using RNC-2SFCA; (4) to construct urban natural road-subway-hybrid network with the dual approach, and to calculate the accessibilities using TD-2SFCA; (5) to applied several basic methods, such as the nearest provider together with Euclidian-straight-line-based 2FSCA and traveling-time-based 2SFCA to analyzing accessibilities which would be compared with RNC-2SFCA and TD-2SFCA results; (6) to calculate AAIS indexes of hospital discharge records provided by Shenzhen Center for Health Information to analyze hospital utilization rates, thus to evaluate all kinds of potential accessibility methods and to evaluate 2010-2015 healthcare planning policies of Shenzhen.