1. Research Background and Literature Review
Currently, China is in a period of rapid demographic, social, and economic transformation. The crisis of an aging population is worsening. In 2017, the number of elderly people in China (over 60 years of age) reached 240 million [1
]. This makes China the country with the largest number of elderly people in the world [3
]. Changchun is the capital city of Jilin province, located in the plains of Northeast China, which is currently growing into an aging city. The city’s average aging population ratio has grown from 8.32% to 13.64% from 2010 to 2016. There is no doubt that nursing homes will become increasingly important urban service facilities. Differing from the usual western definition, nursing homes in this paper mainly refer to day care centers for the elderly, which can provide temporary daytime care, accommodation, health care, rehabilitation and amenities for elderly people. Thus, the accessibility of nursing homes and transportation costs associated with nursing homes is very important. In recent years, the Changchun municipal government is making a spatial plan for nursing homes, where to place the facilities for the elderly has also become a topic of common concern for scholars and policy makers. In this context, it is increasingly urgent and necessary to explore the spatial equity issue of nursing homes in Changchun.
Spatial equity usually refers to different residents (regardless of their social class, income, or race) having equal access to certain services [4
]. Early scholars believed that spatial equity only meant the uniform spatial distribution of service facilities or ensuring the equality of distance to service facilities. Particular attention was given to spatial dimensions [5
]. After that, however, the definition of spatial equity was expanded into social dimensions, such as class, income, and age [13
]. For example, Talen and Anselin (1998) believed that spatial equity refers to the coordination of facilities or services between different socioeconomic groups [8
]. Omer (2006) believed that spatial equity refers to the degree to which the services or facilities were equally distributed in different economic, ethnic, or age groups [7
]. In recent years, the integration of different trip modes into traditional accessibility models has become a burgeoning field. Some scholars believe that trip modes have significant influence on spatial equity of service facilities, because different modes of transport can produce different accessibility landscapes [14
] What is more, integrating trip modes into traditional accessibility models can better reflect different social groups’ ability to access certain services [17
Currently, the metrics measuring spatial equity mainly include the accessibility-based approaches [7
], the Lorentz curve method [20
], the Gini coefficient method [21
], the coefficient of variation (CV) [22
], and the Teil index [23
]. Among them, the accessibility-based approaches are the most commonly used methods in the assessment of spatial equity of service facilities [12
] and are widely used in the evaluation of spatial equity of urban green spaces [18
], public playgrounds [28
], educational resources [29
], public transport [14
], healthcare facilities [30
], and so forth. Among the accessibility-based approaches, the gravitational potential model, two-step floating catchment method (2SFCA) and their corresponding improved models are relatively popular methods for measuring spatial equity. For example, based on gravity model, Chang (2011) developed a spatial equity index to explore the spatial equities of urban public facilities from both accessibility and mobility perspectives [25
]. Talen (1998) evaluated the spatial equity of public playgrounds by a gravity potential model [8
]. Vadrevu (2016) assessed the spatial equity of maternal health services by an enhanced two-step floating catchment area method [22
]. Shen (2017) explored the spatial equity of public green space among different resident groups by a Gaussian-based two-step floating catchment area method [31
]. Dadashpoor (2016) developed an integrated index of spatial equity to assess the spatial inequity of urban facilities in a disaggregated and aggregated manner [32
]. Almohamad (2018) assessed the spatial equity of public green spaces by average nearest neighbor and network analysis [18
]. Taleai (2014) developed a Spatial Multi-Criteria Analysis (SMCA) method to assess the spatial equity of urban public facilities at different spatial scales [33
Distance is one of the keys in any methodology of accessibility. Network distance has been proven to be more accurate and more realistic than Euclidean distance with accessibility [34
]. As shown in Figure 1
, if A is a spatial unit and B is a facility, the red line is the network distance, the blue line is the linear distance, and the radius of the circle is the distance threshold judging whether or not a facility is accessible. If measured by linear distance, B is accessible to A. If calculated by network distance, however, B is not accessible to A. Additionally, trip modes also have an important impact on the measurement of accessibility [6
], and accessibility landscape also varies with different trip modes. The accessibility range of different trip modes in the same time are significantly different from each other. Therefore, integrating trip modes into the measurement of accessibility is very significant. Threshold is another key to accessibility research; most scholars set a time threshold between 5-min to 30-min, but when taking the walking mode into consideration, 5-min, 10-min, and 15-min are typical time divisions [16
], and the speeds of walking, public transport, and car in urban contexts are normally set to 5km/h, 25km/h, and 30km/h, respectively [14
Some scholars measured the spatial equity of service facilities by integrating trip modes into traditional accessibility models. For example, in order to explore the influence of trip modes on the spatial equity of healthcare facilities, Mao and Nekorchuk (2013) proposed a Two-Step Floating Catchment Area Method (2SFCAM) and integrated bus and car modes into a Two-Step Floating Catchment Area (2SFCA) model [38
]. Shen and Sanchez (2005), as well as Chang and Liao (2011) considered the impact of walking and driving on spatial equity and integrated those trip modes into a potential model [6
]. Coline (2015) also developed a Variable-Width Floating Catchment Area (VFCA) method and assessed accessibility to urban parks in four trip modes: bicycling, driving, public transit, and walking [15
]. Mitchel Langford (2015) claimed that combining public and private transport modes into traditional 2SFCA methodology could contribute to accurately allocating services among different social groups [17
]. Xing and Liu (2018) developed a multi-mode 2SFCA and measured spatial disparity of urban parks in three trip modes: walking, cycling and driving, and claimed a multi-mode model can better reflect accessibility [40
]. The studies conducted in the multi-mode environment show that the accessibility models integrating trip modes can better reflect social equity [15
], and that a traditional non-mode model overestimates accessibility. Therefore, the multi-modes model can provide a more realistic evaluation and offers a better guidance for practices [38
In this paper, based on network analysis, the trip modes of walking, bus, and car were integrated into the improved potential model and the spatial equity of nursing homes in Changchun are explored in 5-min, 10-min, and 15-min scenarios, respectively. The main research objectives of this paper are as follows: What is the geography of spatial equity under different trip modes and different scenarios? Will the geography of spatial equity in different scenarios show the same spatial pattern?
The remainder of this paper is organized as follows: Section 2
provides an introduction to the data sources, as well as our research methodology. Section 3
includes analysis of the spatial differences of spatial equity of nursing homes in different scenarios and its influencing factors and spatial patterns. Finally, the key conclusions and the theoretical and policy implications of this research are outlined in Section 4
4. Conclusions and Discussions
Trip modes have significant influence on spatial equity of nursing homes, and the geography of spatial equity varied with trip modes. The SEi under car mode is significantly higher than that in bus and walking scenarios, whereas the SEi under walking mode is extremely low. In walking mode, areas surrounding the urban core have higher SEi than other regions, while in public transit and car modes, the southwestern areas have superiority over other regions.
Overall, the SEi in Changchun is low, and most areas have very low access to nursing homes. In the 5-min integrated scenarios, 40 subdistricts (69%) belong to underserved areas. In the 10-min scenarios, almost 60% of the subdistricts of Changchun have low access to nursing homes. Even in the 15-min integrated situation, there are still more than half of the subdistricts (52%) belonging to the underserved areas. These underserved subdistricts are mainly located in the urban core areas and most of the urban fringes.
Service capacity of nursing home, travel costs, and the number of seniors, are the common factors that significantly affect spatial equity in all three scenarios. The competition factor, namely, the number of seniors in surrounding subdistricts, only affects the 15-min scenario. However, the number of nursing home has no significant effect on the spatial equity of nursing homes.
Although the geography of spatial equity varies with trip modes and scenarios, the geography of spatial equity in different scenarios presents a similar “low-high-low” ring structure; namely, the SEi within the urban core and at the periphery (except for southwest and southeast urban fringes) is lower than that in the intermediate areas. From the perspective of spatial statistics, southwest urban fringes are hot spot areas with high spatial equity value, whereas the urban core areas are cold spots with low spatial equity value.
From a theoretical perspective, the early studies on accessibility were carried out in a single catchment environment [13
] or in the absence of trip modes and assumed that all residents used the same way to access services [45
]. But, in reality, not all people adopt the same trip mode when they access service facilities. People of different social classes rely on different trip modes in their daily life, so a multi-trip modes accessibility model can also reflect social equity [15
]. This paper found that the landscapes of spatial equity vary with trip modes and scenarios. Thus, it is necessary to take trip modes and spatial-temporal scales into consideration when measuring spatial equity. The original spatial equity mode developed by Chang (2011) neglected the spatial competition of people among neighboring spatial units for the limited resources [6
]. This paper found that the spatial competition coming from the neighboring regions also affect spatial equity, so the competition factor, especially in a macro-spatial scale, matters. In this paper, based on the multi-mode spatial equity model developed by Chang (2011), we further modified this spatial equity model by integrating spatial competition among consumers in neighboring subdistricts for the limited resources and the attraction factor which was represented by the amounts of amenities (physical therapy equipment, fitness and recreational facilities) equipped in each nursing home into the spatial equity model. We then explored the spatial equity of nursing homes on different spatial scales. This research can provide some improvement to current research that does not consider the impact of trip modes and spatial scales on accessibility.
From a policy perspective, this research found that service capacity, travel costs, and the number of seniors, are the main factors influencing spatial equity. However, the number of nursing homes has no significant effect on the area’s spatial equity, which indicates that it is service capacity—not the number of nursing homes—that is the key to the spatial equity of nursing home resources. In addition, the “ring structure” and big spatial variance of spatial equity are highly related to the varied capacities and location preferences of different nursing homes. Because larger nursing homes always need more space to provide more services, they tend to be located in the intermediate zones between the urban core and urban fringe, while small nursing homes tend to be located in urban core areas. These findings have important policy implications to the spatial planning of nursing homes. In the spatial planning of nursing homes, it is necessary to focus on the improvement of service capacity rather than the blind increase of the number of nursing homes. What is more, it is also necessary to consider the balanced distribution of different types of nursing homes; the urban core areas are densely populated areas, but due to the scarce land provision and expensive rent, the capacities of nursing homes are limited. On the contrary, because the urban peripheries are sparsely populated areas with enough space and low land rent, they are the ideal locations for larger nursing homes. To avoid the spatial mismatch between the elderly and the nursing homes, it is necessary to consider the spatial equilibrium of nursing homes with different capacities.
This paper does have some limitations. First, factors seniors taking into account when choosing pension services are quite complex; for example, the socio-economic status of the elderly, the quality and costs of the nursing homes, etc. Due to privacy issues, some data such as the service fee and the service quality of each nursing homes cannot be collected currently, so we cannot include those factors into our analysis, we can only use the amount of the physical therapy equipment, fitness and recreational facilities to represent the attraction of each nursing home. Secondly, spatial equity was determined by both spatial factors and non-spatial factors. The spatial equity model used in this paper is mainly to address the spatial factors. How to integrate more non-spatial factors, such as individual preference, income, age, health status, into traditional spatial equity modes can be the focus of our future study.