Impact of traffic on the spatiotemporal variations of spatial accessibility of emergency medical services in inner-city Shanghai

Abstract

Variability in spatial accessibility of emergency medical services has become a major concern in evaluating the quality of emergency medical services in China. Unlike some other public services, response time is critical in the provision of emergency medical services. Traffic congestion may significantly affect response time, especially in large cities. This study uses a transportation simulation model to estimate the travel time under free-flow and congested road conditions and measure the corresponding spatial accessibility of emergency medical services for various hours of a day in inner-city Shanghai. When traffic congestion is considered, the overall spatial accessibility is significantly reduced, and the effect is further magnified in certain congested areas. The results help policy makers in planning the emergency medical services resource that is sensitive to the spatiotemporal variation of its accessibility.

Keywords

Spatial accessibility emergency medical services macro-traffic simulation traffic congestion spatiotemporal variability Shanghai

Background

The mission of emergency medical services (EMSs) is to save human lives, and time is critical in fulfilling this mission. The overall EMS time is comprised of four intervals, namely activation, response, on-scene, and transport (Carr et al., 2006). This study focuses on an ambulance’s response interval, i.e. the time from its dispatch to arrival at the scene. Any delay in the arrival of EMS, e.g. due to traffic congestion (Griffin and McGwin, 2013; Sisiopiku et al., 2011), lowers a patient’s survival rate. The flashing light and a siren on an EMS vehicle is for minimizing the delay caused by traffic (Shah, 2006). However, unlike most developed countries, when the siren from an ambulance is wailing, drivers of other vehicles in China, India, Thailand, and other countries do not yield; this often results in lost lives (Thai PBS, 2017; The Time of India City, 2016; The Urban Country, 2012). The problem is augmented in inner-city areas, where there is more traffic. It is therefore important to assess how traffic congestion affects access to EMS in these areas and where people are most vulnerable.

Analysis of spatial accessibility is a valid approach to evaluating the spatial distribution of a service and the variability of convenience of accessing the service. Spatial accessibility is determined by the distributions of supply and demand and how they are connected in space. A large body of research on spatial accessibility is concerned with access to health care (Apparicio et al., 2008; Langford and Higgs, 2006; Luo and Qi, 2009; Luo and Wang, 2003; McGrail and Humphreys, 2009; Wan et al., 2012), jobs (Parks, 2004; Wang and Minor, 2002), education (Williams and Wang, 2014), and recreation facilities (Smoyer-Tomic et al., 2004; Talen and Anselin, 1998). Measures of spatial accessibility intend to capture a function of two factors: availability of service choices and travel cost (Chen et al., 2015; Guagliardo, 2004; Yiannakoulias et al., 2013). Population-to-provider ratios are easily calculated and simple to understand, and thus a popular indicator for policy analysts (McGrail and Humphreys, 2009); however, the measure typically assumes that the interaction between supply and demand is confined within an administrative unit and also ignores the variability of accessibility within the unit (Wang, 2015: 95). Both shortcomings have been overcome by the popular “two-step floating catchment area (2SFCA) method” (Wang, 2012; see “Measuring the accessibility of EMS” section for details).

Most recently, Baloyi et al. (2017) evaluated the accessibility of EMS by consulting with city officials to define road speeds, and Szymon (2017) used the 2SFCA method to assess the accessibility of fire services by assuming vehicles to travel at maximum permitted speeds. Neither considered the impact of traffic congestion on the road network. The development of geographic information systems (GIS) technology has made it possible to estimate the travel time on street networks more accurately (Wu et al., 2004). However, most researchers have not taken advantage of the full capacity of GIS and still assume constant travel speeds throughout the day (Weber and Kwan, 2002). Accessibility of EMS in inner-city neighborhoods, often populated by disadvantaged population groups, may have adversely affected worsened traffic congestions more than any other areas. Studies ignoring the traffic factor may not be able to reveal the full extent of disparity in EMS access, especially the temporal variability in such a disparity. The problem is most pronounced in large cities in developing countries, usually plagued by traffic congestion to some extent throughout the day.

This study evaluates the impact of spatiotemporal variation of traffic congestion on spatial accessibility of EMS in inner-city Shanghai in a GIS environment. The aims of the study are to model the traffic flow based on a macro-traffic simulation, to estimate the travel time at various periods during the day, and to examine the EMS accessibility accordingly. By revealing disparities in spatial accessibility of EMS across the study area under both free-flow and congested conditions, we hope to shed light on the possible policy remedies to improve the efficiency of EMS delivery and reduce inequality in its access.

Study area and data

The study area is the inner city of Shanghai, China (Figure 1). The area is bordered by the Inner Ring Road and covers about 120 km² with a population over 3.4 million according to the Census 2010. The area has a complex road network with a road density as high as 8.124 km/km² (the road density outside the inner city but within the outer ring of Shanghai is 5.54 km/km²). It includes over 3800 links, among which there are three elevated highways. With the central business district as its core, this area is mixed with the highest concentration of commercial land use and the highest density of residential land use, typical of many inner-city areas of China. It is certainly one of the most important areas of Shanghai in terms of social and economic development. Given the high intensity of both population and business activities, the impact of quality EMS is amplified. However, with an increasing car ownership in China and especially so in Shanghai, traffic congestion in the area has become a major problem. Moreover, the geographic distribution of traffic congestion changes significantly during the day. Therefore, it offers an excellent case study to highlight the spatiotemporal variation of EMS accessibility.

Figure 1.

Study area and EMS stations in inner-city Shanghai. EMS: emergency medical service.

Data for the study begin with two types of point locations: facility locations, where EMS stations are located; and demand locations, where prospective users of EMS reside. The EMS data for Shanghai are extracted from the Amap web page (http://ditu.amap.com/) and the Shanghai municipality government web page (http://big5.shanghai.gov.cn/gate/big5/www.shanghai.gov.cn/nw2/nw2314/nw2319/nw12344/u26aw17406.html). There are 17 EMS stations located in the study area, as shown in Figure 1. The EMS stations are unevenly distributed, as most of the stations (14) are located west of Huangpu River, and only three stations in the east across the river. Residential area (“juweihui”) is chosen as the analysis unit for the population distribution. Based on the sixth National Population Census of China in 2010, there are 963 residential areas in the study area, with a total population of 3,420,890. Given their relatively small areas with an average of 0.12 km², it is adequate to represent residential areas by their centroids.

The street network data are extracted from the OpenStreetMap (OSM) including road levels. The OSM data in China are not well developed, and the preliminary road networks derived from the OSM are enhanced by using images from the ArcGIS online and the Amap website. According to the Standardization Administration of the People’s Republic of China (2012), the road segments are categorized as five levels (I, II, III, IV, and V) and are assigned speed limits of 80, 60, 50, 40, and 30 km/h, respectively. Other information such as one-way streets and maximum traffic volumes for lanes is also accounted for. The final street network is saved as an ArcGIS transport network format.

Methods

Modeling traffic congestion

In general, serious traffic congestion occurs as the traffic volume exceeds the road capacity. In this study, the degree of congestion was defined using the ratio of vehicle volume (V) to capacity (C), namely V/C. Since road capacity is not variable during a day if there are no contingencies like traffic accidents or adverse weather, the major procedures of modeling congestion are to estimate and count the vehicle volume on each road link.

Basically, the traffic volume on a given road link depends on two factors: the overall traffic demand (the total number of vehicles in the road network) and the utilization priority of the link (the number of vehicles using this link). Our method of estimating the traffic volume is extracted from the Four-Step Transport Demand Model, which comprises two principle steps: (1) Estimate the traffic demand. As a simplifying evaluation, different scenarios of trip generation (the number of OD pairs) were tested first. In each scenario, all trips were distributed over the study area according to the distribution of automobile journeys presented in Shanghai’s transportation annual report for 2015 (see black lines in Figure 2). (2) Assign the trip path. Travelers usually tend to take the shortest or fastest path which can be calculated by various algorithms. This study used the Dijkstra (1959) algorithm to calculate the trip path with minimum travel time between each pair of origin (O) and destination (D).

Figure 2.

Traffic monitoring stations and relative traffic volumes (based on Amap).

Moreover, travel time is calculated by road length/speed, and the initial speed was set as the speed limit which varies with the road levels. But speed usually decreases notably when the traffic volume gradually increases on the road, which enlarges the travel time as a consequence (Li et al., 2017). Thus, the Greenshields et al. (1935) model was used to model the dynamic variation of speed. The model is defined as

{\begin{array}{l} T_{0} = L / V_{0} \\ T_{V} = T_{0} / (1 - V / C) (V / C \leq 0.9) \\ T_{v} = T_{0} / 0.2 (V / C > 0.9) \end{array}

(1)

where T_v is the travel time at traffic flow V (updated every 2 min), T₀ is the free-flow travel time, L is the route length, V₀ is the speed limit, and C is the road capacity. In this case, the traffic volume may increase very close to capacity and cause serious traffic jam. As a result, the travel speed approaches zero. When V/C is very close to 1, T_v reaches infinity, but it is not in line with reality. Thus, we modified the parameters of the model based on the comparison of the travel time calculated by Amap Direction API with the travel time calculated by the model (see “Measuring ambulance travel time with the effect of traffic congestion” section for details). Finally, the travel time on extremely congested roads (V/C > 0.9) can be five times as much as that under free-flow occasions.

To test the different scenarios of trip generation to see which scenario best matches the reliability, a comparison was made between the traffic simulation results and the actual monitored traffic volume at 15 road intersections (see red dots in Figure 2). It is noted that the actual data were collected during the rush hours. The linear regression analysis (Table 1) shows the relationship between the simulated and the monitored traffic volume. Obviously, if the travel demand was set as 145,500 OD pairs during the rush hours (slope value is 1.07), R² would be the highest (0.74), which indicated that the simulated traffic volume would be best correlated with the actual traffic volume at the 15 traffic intersections.

Table 1.

Estimated traffic flows versus real traffic flows at intersections.

Interval	Ratio of traffic flow	Cars representing each pair of OD (pcu)	Estimated traffic volume (pcu)	Slope	Regression R²
6.00–7.00 am	0.2	1	48,500	0.35	0.68
8.00–9.00 am (rush hour)	0.6	3	145,500	1.07	0.74
10.00–11.00 am	0.45	2.25	109,125	0.80	0.73
12.00 am–1.00 pm	0.4	2	97,000	0.70	0.68
2.00–3.00 pm	0.45	2.25	109,125	0.80	0.73
4.00–5.00 pm	0.5	2.5	121,250	0.89	0.73
6.00–7.00 pm (rush hour)	0.6	3	145,500	1.07	0.74
8.00–9.00 pm	0.35	1.75	84,875	0.62	0.72
10.00–11.00 pm	0.3	1.5	72,750	0.53	0.70

Further, in order to identify the spatiotemporal variation of the traffic congestion, traffic volumes during other hours of a day were also estimated. The actual ratio of V/C was obtained from Shanghai’s transportation annual report for 2015. Two peaks arise during 8–9 am and 6–7 pm, respectively. Finally, we generated nine scenarios of OD pairs during nine different time periods (Table 1). Thus, a total of 48,500 trips (pairs of origins (O) and destinations (D)) were generated to model the traffic flows. All trips departure randomly during 1 h. To model traffic congestion during different hours of a day, a pair of OD represented different vehicle numbers (Table 1).

Based on the above process, we calculated traffic flows on all links, and then calculated V/C values for all links for various time intervals. Meanwhile, the Road Service Level was classified into four categories (Figure 3), I, II, III, and IV, in order of deteriorating traffic conditions (Shen et al., 2007), which reflected the distribution of traffic congestion intuitively. It should be possible for ambulances to avoid congested roads once supplied with information about congested links.

Figure 3.

Spatiotemporal distribution of vehicle volume/capacity (V/C).

Measuring ambulance travel time with the effect of traffic congestion

Travel cost is one of the most important factors in determining the spatial accessibility of EMS. An ambulance always takes the fastest but not necessarily the shortest route, and the route may vary at different times of the day (Kumar and Benedict, 2011). Thus, we need to calculate the travel time for calculating spatial accessibility of EMS.

We used all the streets to establish network datasets to compute travel time between census tracts and EMS stations based on network analysis in ArcMap. Since network analysis is often computationally demanding, in this study, each intersection of a census tract or EMS station that was nearest to it was considered as the location of the census tract or the EMS station. In addition, the source feature dataset must have fields representing network impedance values such as distance and travel time. Thus, calculating the time to pass a link, link lengths and vehicle speeds is required. However, on the same road, the vehicle speed varies at different times due to changes in traffic flow. We calculated the traffic flows on each link based on the above traffic simulation, thereby calculating the exact time passing link combined with the relationship between traffic flow and travel time given by the Greenshields model. Finally, the EMS stations were loaded as the set of origins and the census tracts were loaded as the set of destinations. Previous research has demonstrated that a key marker of quality EMS care is the ability to meet an 8 min response time (Pons et al., 2005); therefore, the default cutoff value was 8 min, which was used to solve the OD Cost Matrix.

In order to validate the travel cost calculated by our tool, we randomly selected 1500 OD pairs (origins are from the EMS stations and destinations are from the centroids of census tracts) and calculated the travel time between each pair using the Amap Direction API during different time periods. All calculations were conducted under good weather conditions. A comparison between the two sets of results indicates that the average relative error of travel cost calculated by our tool is below 26% (Table 2). In addition, there is a strong correlation between the travel time calculated by our tool and the Amap Direction API with R² greater than 0.6. That is to say, the estimated travel time by our tool is reasonably reliable. The discrepancy may be caused by our tool not being able to account for specific road conditions as such data (such as accidents, traffic-reduction measures and etc.) are not available to our research team.

Table 2.

Estimated travel time versus real travel time based on the Amap Direction API.

	6–7 am	8–9 am	10–11 am	12–1 pm	2–3 pm	4–5 pm	6–7 pm	8–9 pm	10–11 pm
Regression R²	0.67	0.72	0.71	0.71	0.69	0.67	0.68	0.7	0.71
Average relative error	22.79%	22.29%	21.70%	23.94%	23.04%	25.39%	25.27%	21.79%	21.30%

Measuring the accessibility of EMS

The complex relationship between supply and demand is considered in the calculation of spatial accessibility by the 2SFCA method. In the first step of the 2SFCA, the service area of each EMS station (j) is defined by a threshold travel time (d₀). After summing up all demands within the service area, the availability indicator R_j in the area is calculated as follows

R_{j} = \frac{10, 00, 000}{_{\sum}^{k \in {d_{k j} \leq d_{0}}} P_{k}}

(2)

where d_kj is the travel time (after accounting for congestion effect) between residential area k and EMS station j, and P_k is the demand at residential location k that falls within the area (i.e. d_kj≤ d₀). Due to the lack of data for the number of ambulance or medical emergency personnel in each EMS station, we apply a uniform multiplier 1,000,000 to represent the capacity of each EMS station.

In the second step of the 2SFCA, the search area around each population location (i) is also defined by d₀. Search for all EMS stations (j) that are within the area, and sum up the availability ratios R_j associated with these EMS stations to yield the accessibility at residential location k

A_{i}^{F} = \sum_{j \in {d_{i j} \leq d_{0}}} R_{j}

(3)

In summary, the first step is to assign a ratio of EMS station to population location as a measure of supply availability; and the second step is to sum the ratios in the overlapping service areas to measure the accessibility for a population location, where residents have access to multiple supply locations. The multiplier in equation (2) is set as 1,000,000, and the accessibility score in equation (3) can be interpreted as the number of EMS stations reachable per one million residents.

Results and discussion

Differences in the number of routes from EMS station to census tract

Calculations of traffic flow on a street network at various times are made through the process of traffic simulation. The number of routes between census tracts and EMS stations at various time periods during the day within 2 min driving time intervals in the study area is shown in Figure 4. The number of travel routes at a required time varies considerably using the travel cost from the traffic simulation under different time periods. Ignoring traffic conditions, there are over 5000 routes within an 8 min drive time of the EMS stations, which is almost twice the number of routes for those with traffic conditions from 6.00 to 7.00 am. In addition, with increasing traffic volumes, the number of journey routes within 8 min decreases, so that at rush hour the minimum number of residents may be served by the closest EMS station during the 8 min “golden period.” Meanwhile, Figure 4 also shows the average driving speed under different traffic conditions. Ideally, ambulances are much faster speed without considering traffic condition, which is almost twice as much as the speed at rush hour. When there is high-density traffic, the driving speed becomes slower, thereby increasing the travel time from EMS stations to spots. Qualitatively, the map of V/C values (Figure 3) makes intuitive sense, with the highest commuter traffic flow at rush hours, and thus the minimum number of journey routes travel paths within an 8 min drive time and slowest speed. Different traffic conditions during the day have a large impact on the calculation of trip paths.

Figure 4.

Number of trip paths and average speed under free-flow and congested road conditions.

The map of V/C values (Figure 3) demonstrates that highest commuter traffic flow is on elevated roads and major arterials, especially, on the South–North Viaduct and the Yan’an Viaduct, and the lowest flow is on minor residential roadways. Consequently, EMS personnel need to carefully consider the traffic conditions and plan their routes to avoid congested roads, to minimize the travel time to the emergency site.

Differences in accessibility of EMS

As stated previously, the 2SFCA-derived accessibility score is interpreted as the number of reachable EMS stations per million people, and a higher score corresponds to better access. Most of the accessibility scores under the free-flow traffic condition range between 4.5 and 8.0 (SD = 2.07), while the scores during rush hours have values of 0, 3, and 5 (SD = 2.46). This is to say, traffic congestion reduces the accessibility scores overall and enlarges the gap between areas of better access and poorer access.

To further illustrate the spatial accessibility at different hours of a day, we group the accessibility scores into five levels, ranging from Q1 to Q5, in a descending order (Figure 5). Q5 highlights areas not reachable by EMS within 8 min; for the remaining areas, Q1 for the top 25th percentile accessibility, Q4 for the bottom 25th percentile, and Q2 and Q3 for the two middle groups (Guagliardo, 2004). Accessibility generally decreases as a function of distance from the EMS stations along the road network, and worse traffic further reduces its value. Overall, the east side of the Huangpu River with fewer EMS stations has lower accessibility than the west side. The areas falling in Q5 are also more prominent in the east side and expand significantly during the peak hours. Also note the absence of any Q5 areas in the west side under the free-flow condition, and the presence and enlargement of a Q5 pocket at the northwest corner in the west side during the peak hours. Accessibility in most areas declines due to traffic congestion. Among over 500 census tracts with changed accessibility during the hours of peak traffic, over 360 tracts (about 72%) have lower accessibility levels. However, a small portion of areas have actually improved accessibility.

Figure 5.

EMS accessibility at different hours of the day.

Finally, the focus is turned on the areas, classified in Q5, with deficient access to EMS. It calls for policy makers to improve the provision of EMS and expand the coverage areas within the critical 8 min range. Table 3 shows the size of Q5 areas and its percentage under different traffic conditions. When assuming the free-flow condition, nearly all areas are covered by the required response time of 8 min, except for a merely 1% area on the east edge and 0.21% population. However, as traffic volume drives up travel time, the areas left out of the coverage expand to about 12% area and 5.96% population, which is very significant. The Q5 areas are initially limited to the periphery of the study area in the early morning time (6–7 am) and pop out even in the center of the study during the morning peak hour (e.g. 8–9 am). It calls for special policy measures in EMS delivery to ensure the safety of all citizens.

Table 3.

Area and population of Q5 based on various traffic conditions.

Interval	Area in km² (in %)		Population (in %)
Free-flow speed	1.14	(1.00)	935	(0.21)
6.00–7.00 am	1.14	(1.00)	935	(0.21)
8.00–9.00 am (rush hour)	13.51	(11.86)	26,688	(5.96)
10.00–11.00 am	8.76	(7.69)	13,501	(3.02)
12.00 am–1.00 pm	7.54	(6.62)	9226	(2.06)
2.00–3.00 pm	8.76	(7.69)	13,501	(3.02)
4.00–5.00 pm	9.21	(8.09)	15,733	(3.52)
6.00–7.00 pm (rush hour)	13.51	(11.86)	26,688	(5.96)
8.00–9.00 pm	6.68	(5.87)	6395	(1.43)
10.00–11.00 pm	5.66	(4.97)	3655	(0.82)

Discussion

Two travel times are estimated between each census tract and each EMS facility: one when the traffic flows freely, and the other when travel is affected by traffic congestion at various times based on traffic simulations. Our results show that the number of travel routes and driving speed with traffic congestion are about half of those under the free-flow condition, and subsequently different accessibility scores are obtained in the two scenarios. Most of the census tracts have different relative levels of spatial accessibility when considering traffic congestion. Various methods of travel time calculation result in varying spatial accessibility levels in outlying areas or in areas with inefficient road networks. Even though road networks are fairly efficient with relatively low traffic congestion in the eastern part of the study area, the area is less affected by traffic, but still suffers from poorer accessibility of EMS because of shortage of EMS stations. Traffic congestion may also lead to a smaller range of services for EMS, thus less census tract for within a short range of EMS station and improved accessibility for these areas. However, for areas farther away from EMS stations, ambulances may not reach within 8 min, and therefore congestion reduces accessibility there. In short, the congestion further polarizes the EMS accessibility between the haves (with developed road network and EMS) and the have-nots (with poor road network and EMS), drives the disparity. This calls for policy remedies for placing more EMS stations in areas with inadequate access more attributable to their relatively remote locations, and increasing the number of ambulances in areas of poor access more attributable to their high concentration of population.

Some recent publications cover the general area of spatial accessibility under varying traffic conditions (e.g. Chen et al., 2016; Guagliardo, 2004). However, none examines the effect of temporal variation on accessibility on a daily basis. Our study shows that the quality of accessibility deteriorates most at times of peak traffic. For example, the area with inadequate access to EMS (i.e. Q5) is 13.51% of the total area at times of peak traffic, which is nearly two times during nonworking hours. When we limit the analysis of the study area without considering the supply–demand interaction , the edge effect surfaces. In other words, most census tracts near the central area have the highest accessibility and areas on the periphery have poor access. This is especially evident in the absence of traffic congestion. However, when congestion becomes severe at times of peak traffic, new areas of poor access emerge in the central area. In other words, the central part of the study area may at certain times be as poorly served as those on the periphery. Temporal variation in accessibility adds another layer of complexity to the planning of adequate EMS provision. One may consider using medical emergency motorcycles driven by EMS technicians as a way to provide timely care in crowded urban areas during peak congestion periods (Kiefe and Soares-Oliveira, 2008). Another option is to use mobile EMS stations with flexible work hours that may be deployed to those highly congested areas.

Conclusions

This study examines the spatial accessibility of EMS in inner-city Shanghai, under varying traffic conditions. By calculating travel times from EMS stations to scenes under free-flow versus congested road conditions, our results demonstrate that the spatial accessibility of EMS under the free-flow condition is overestimated, and congestion during times of peak traffic sharply reduces the accessibility in many of the high-density areas. Congestion may also enlarge the disparity in access of EMS between the well-served and the under-served areas, an often missed repercussion by researchers. Policy makers need to consider deploying EMS personnel on motorcycles or adding mobile EMS stations with flexible hours in these heavily congested areas or in order to meet the mandatory minimum response time for EMS.

Several limitations need to be pointed out. The study simply uses population to define demand in measuring spatial accessibility and does not account for the impact of demographic composition on demand (e.g. seniors, residents with chronic diseases). Moreover, the study area is limited to the inner city of Shanghai and does not include the areas beyond the inner ring. The edge effect is a major concern. Future work will be expanded to include the whole municipality with a wide range of population densities and thus areas of various levels of congestion. Besides, due to the lack of real-life response times driving by ambulances, we just calculated the travel times by the Amap Direction API to improve the validity and to verify the model. In addition, extreme weather events have become increasingly frequent and intense as a result of climate change. It will be valuable to examine the impacts of heavy rain and flooding on traffic conditions and subsequently on the spatiotemporal accessibility of EMS.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/ or publication of this article: This work was financially supported by the Major Program of National Social Science Foundation of China(18ZDA105), the National Natural Science Foundation of China (51761135024 and 41671095), and the National Key Research and Development Program of China (2017YFE0107400).

Wenyan Hu is a master’s student in Physical Geography at East China Normal University. Her research interests are in urban risk disaster and spatial geography. And her work has been published in Progress in Geography.

Jinkai Tan is a PhD candidate in Physical Geography at East China Normal University. His research interests are in mathematical modeling and machine learning, flood numerical simulation, climate change and its effect on tropical cyclone. And his work has been published in some journals, including Theoretical & Applied Climatology.

Mengya Li is a PhD candidate in Physical Geography at East China Normal University. Her research interests are in transport geography, travel behavior, and urban risk disaster. Her work has been published in Applied Geography, Transactions in GIS, and Science of the Total Environment.

Jun Wang is a Professor in the School of Geographic Science at East China Normal University. His research interests are in urban risk disaster, urban geography, and flood numerical simulation. His work has been published in Climatic Change, Science of The Total Environment, and Journal of Transport Geography, etc.

Fahui Wang is James J. Parsons Professor and chair of the Department of Geography and Anthropology, Louisiana State University. His research interests include GIS applications in human geography (urban, economic, transportation), city and regional planning, and public policy (crime and health).

References

Apparicio

Abdelmajid

Riva

, et al. (2008) Comparing alternative approaches to measuring the geographical accessibility of urban health services: Distance types and aggregation-error issues. International Journal of Health Geographics 7: 7.

Baloyi

Mokgalaka

Green

, et al. (2017) Evaluating public ambulance service levels by applying a GIS based accessibility analysis approach. South African Journal of Geomatics 6(2): 172.

Carr

Caplan

Pryor

, et al. (2006) A meta-analysis of prehospital care times for trauma. Prehospital Emergency Care 10: 198–206.

Chen

Zhou

Liu

, et al. (2016) Estimating the effect of traffic congestion on accessibility to emergency medical care services: Take Guangzhou as an example. Progress in Geography 35: 431–439.

Chen

Wheeler

Huang

, et al. (2015) Evaluation of spatial accessibility to Ohio trauma centers using a GIS-based gravity model. British Journal of Medicine & Medical Research 10: 1–12.

Dijkstra

EW (

, 1959) A note on two problems in connection with graphs. Numerische Mathematik 1: 269–271.

Greenshields

Bibbins

Channing

, et al. (1935) A study of traffic capacity. In: Highway research board proceedings. Washington, D.C., 6–7 December 1934, Highway Research Board; 14: 448–477.

Griffin

McGwin

G Jr

(2013) Emergency medical service providers’ experiences with traffic congestion. Journal of Emergency Medicine 44: 398–405.

Guagliardo

(2004) Spatial accessibility of primary care: Concepts, methods and challenges. International Journal of Health Geographics 3: 3.

10.

Kiefe

Soares-Oliveira

(2008) Medical emergency motorcycles: Are they safe? European Journal of Emergency Medicine 15: 40–42.

11.

Kumar

Benedict

(2011) Development of route information system for ambulance services using GPS and GIS. International Journal of Geomatics & Geosciences 2: 147–156.

12.

Langford

Higgs

(2006) Measuring potential access to primary healthcare services: The influence of alternative spatial representations of population. Professional Geographer 58: 294–306.

13.

Huang

Yin

, et al. (2017) Modeling the traffic disruption caused by pluvial flash flood on intra-urban road network. Transactions in GIS 22: 311–322

14.

Luo

(2009) An enhanced two-step floating catchment area (E2SFCA) method for measuring spatial accessibility to primary care physicians. Health & Place 15: 1100.

15.

Luo

Wang

(2003) Measures of spatial accessibility to health care in a GIS environment: Synthesis and a case study in the Chicago Region. Environment & Planning B Planning & Design 30: 865–884.

16.

McGrail

Humphreys

(2009) Measuring spatial accessibility to primary care in rural areas: Improving the effectiveness of the two-step floating catchment area method. Applied Geography 29: 533–541.

17.

Parks

(2004) Access to work: The effects of spatial and social accessibility on unemployment for native-born black and immigrant women in Los Angeles. Economic Geography 80: 141–172.

18.

Pons

Haukoos

Bludworth

, et al. (2005) Paramedic response time: Does it affect patient survival? Academic Emergency Medicine 12: 594–600.

19.

Shah

(2006) The formation of the emergency medical services system. American Journal of Public Health 96: 414–423.

20.

Shen

Zhu

Ying-Jun

(2007) Research on calculation method for road saturation. Communications Standardization 1: 125–129.

21.

Sisiopiku

Sullivan

Foster

, et al. (2011) Emergency response and traffic congestion: The dispatcher’s perspective. Disaster Management 119: 125–135.

22.

Smoyer-Tomic

Hewko

Hodgson

(2004) Spatial accessibility and equity of playgrounds in Edmonton, Canada. Canadian Geographer 48: 287–302.

23.

Szymon

(2017) The use of network analyst tool and 2SFCA method to assess fire service effectiveness in a city as exemplified by Łódź. Geometics, Landmanagement and Landscape 1: 147–158.

24.

Talen

Anselin

(1998) Assessing spatial equity: An evaluation of measures of accessibility to public playgrounds. Environment & Planning A 30: 595–613.

25.

Thai PBS (2017) 20 percent of critically-ill patients die in ambulances. Available at: http://www.pattayamail.com/thailandnews/20-percent-critically-ill-patients-die-ambulances-161550 (accessed 8 March 2018).

26.

The Standardization Administration of the People’s Republic of China (2012) The industry standard of the people’s republic of China: Code for design of urban road engineering (CJJ 37-2012). Available at: http://www.gbstandards.org/China_standards/CJ/CJJ%2037-2012.htm (accessed 8 March 2018).

27.

The Times of India City (2016) Chocking life: 12 on way to hospitals die in traffic jam. Available at: https://timesofindia.indiatimes.com/city/raipur/Choking-life-12-on-way-to-hospitals-die-in-traffic-jams/articleshow/54910538.cms (accessed 8 March 2018).

28.

The Urban Country (2012) Beijing woman dies because ambulance stuck in traffic. Available at: http://www.theurbancountry.com/2012/12/beijing-woman-dies-because-ambulance.html (accessed 8 March 2018).

29.

Wan

Zhan

Zou

, et al. (2012) A relative spatial access assessment approach for analyzing potential spatial access to colorectal cancer services in Texas. Applied Geography 32: 291–299.

30.

Wang

F (

, 2012) Measurement, optimization and impact of health care accessibility: A methodological review. Annals of the Association of American Geographers 102: 1104–1112.

31.

Wang

(2015) Quantitative Methods and Socio-economic Applications in GIS. Boca Raton, FL: CRC Press.

32.

Wang

Minor

(2002) Where the jobs are: Employment access and crime patterns in Cleveland. Annals of the Association of American Geographers 92: 435–450.

33.

Weber

Kwan

(2002) Bringing time back in: A study on the influence of travel time variations and facility opening hours on individual accessibility. Professional Geographer 54: 226–240.

34.

Williams

Wang

(2014) Disparities in accessibility of public high schools, in metropolitan Baton Rouge, Louisiana 1990–2010. Urban Geography 35: 1066–1083.

35.

Lee

(2004) Travel-time prediction with support vector regression. IEEE Transactions on Intelligent Transportation Systems 5: 276–281.

36.

Yiannakoulias

Bland

Svenson

(2013) Estimating the effect of turn penalties and traffic congestion on measuring spatial accessibility to primary health care. Applied Geography 39: 172–182.