Do Primary Healthcare Facilities in More Remote Areas Provide More Medical Services? Spatial Evidence from Rural Western China

Shen, Chi; Lai, Sha; Deng, Qiwei; Dong, Wanyue; Cao, Dan; Zhao, Dantong; Zhao, Yaxin; Zhou, Zhongliang; Chen, Xi

doi:https://doi.org/10.1155/2023/6131486

Health & Social Care in the Community

On this page

Abstract Introduction Materials and Methods Results Discussion Conclusions Data Availability Conflicts of Interest Acknowledgments Supplementary Materials References Copyright Related Articles

Research Article | Open Access

Volume 2023 | Article ID 6131486 | https://doi.org/10.1155/2023/6131486

Do Primary Healthcare Facilities in More Remote Areas Provide More Medical Services? Spatial Evidence from Rural Western China

Chi Shen,¹Sha Lai,¹Qiwei Deng,²Wanyue Dong,³Dan Cao,¹Dantong Zhao,¹Yaxin Zhao,²Zhongliang Zhou,¹and Xi Chen⁴

Academic Editor: Kathiravan Srinivasan

Received03 Jan 2023

Revised23 May 2023

Accepted30 Jun 2023

Published18 Jul 2023

Abstract

Primary healthcare institutions (PHIs) in China have experienced a sizable decline in medical services in recent years. Despite the large regional disparities in China, there is a lack of evidence on the differential patterns of medical services offered by PHIs, especially from a spatial perspective. This study examines whether residents in more remote areas use more medical services offered by township healthcare centers (THCs), a main type of PHIs. Linking medical visits to 923 THCs in a western Chinese province in 2020 with the driving time and geographic coordinates from the Gaode map, a leading map navigation provider in China, we applied a multilevel linear model and a geographically weighted regression to examine spatial heterogeneity in medical service utilization. We showed that a one-hour increase in the shortest driving time between THCs and the local county hospitals was associated with an average 6% increase in THCs outpatient visits and a 0.6% increase in THCs inpatient visits. Our findings suggest that THCs located in more remote areas provided more medical services, especially outpatient services. This study may improve our understanding of the spatial disparity in medical services provision for THCs in western China. We suggest that more attention should be paid to enhance the quality of THCs in remote areas, which deserve a higher priority in the process of improving the quality of primary healthcare. Health policymaking in developing regions such as western China should take geographic factors into account.

1. Introduction

Primary healthcare institutions (PHIs) in China have experienced a significant decline in medical services over the past few years, with the proportion of outpatient services provided by PHIs decreasing from 58.5% to 54.1% from 2015 to 2019, and the proportion of inpatient services decreasing from 20.1% to 16.9% [1]. One of the main concerns is whether all PHIs in China suffered from the same level of reduction in medical services.

An important backdrop to the shrinking of medical services in PHIs in China is the new healthcare reform that began in 2009. In this reform, three major policies were applied to PHIs: increased government investment, elimination of drug markup as a source of funding, and provision of basic public health services [2]. Studies have explored the reasons why PHIs have suffered significant declines in medical services. Li et al. [3] argued that a policy that increases government investment creates incentives similar to the so-called iron rice bowl policy (occupations with job security and benefits) and leaves PHIs in a low productivity state. Shen et al. [4] also found that township healthcare centers, a form of PHIs in China with large government subsidies, reduced the volume of medical services after 2011. Ma et al. [5] pointed out that the declining use of primary healthcare as a proportion of total healthcare was an unintended consequence of the reforms because the performance-based salary system did not link with quality to ensure adequate compensation for primary healthcare professionals in China. Yip et al. [6] noted that aligning the incentives and governance of primary healthcare systems through the establishment of an integrated delivery system based on primary healthcare would improve the quantity and quality of primary healthcare providers in China. Despite the large regional differences in China, there have been gaps in the literature that aims to understand the heterogeneity of medical services and their decline in PHIs, especially from a spatial perspective.

Spatial factors have been found to play an important role in health systems around the world. We summarized the impact of spatial factors on health sectors into two areas: health-seeking behavior and healthcare utilization; and health outcomes. In terms of health-seeking behavior and healthcare utilization, suburban Chinese with poor access to high-grade hospitals are more likely to perform self-treatment compared with urban residents [7]. A survey study conducted in northern China found that the rate of receiving hypertension management services for people living closer to PHIs was about 10 times (OR = 10.360, 95% CI 2.090 to 51.343) as high as those living closer to higher-level hospitals [8]. Jiang et al. [9] explain that the preference for seeking healthcare services nearby is as important as being driven by price or seeking a higher quality of care. Even in online healthcare utilization, geographic distance between physician and patient is negatively correlated with the use of online medical services, although this effect is 40% to 50% of the size for offline medical services [10]. The same pattern has been observed in the United States; in the Texas Coastal Bend, older adults with poorer access to general practices but easier access to hospitals had a higher rate of ambulatory care-sensitive conditions in the emergency department [11]. Similarly, in low-middle income countries such as Ghana, Haiti, and Kenya, spatial accessibility to healthcare facilities played a significant role in birthing and maternal services. An increase in distance by one kilometer was associated with a 6.7% significant reduction in the prevalence of women giving birth in health facilities in eastern Ghana [12]. In rural Haiti and Kenya, lower-level use of maternity services was observed in some areas with longer distances to facilities and low quality of care [13].

In terms of health outcomes, large-scale road construction programs in rural India have improved access to medical facilities and led to better care for pregnant women and subsequently better health outcomes such as lower rates of delivery complications [14]. For elderly people, the low accessibility to healthcare services due to limited public transport service widens health inequalities between older adults living in urban and rural areas [15], and promoting active travel modes among the older population and community-level health facility planning to reduce travel time can improve the use and satisfaction of older adults with primary care visits [16].

As mentioned above, spatial accessibility to PHIs and hospitals and socioeconomic factors may shape residents’ health-seeking behavior and thus generate different healthcare utilization in different institutions. Therefore, this study examines whether residents in more remote areas use more medical services offered by PHIs. Remote areas in this study are defined as townships with long travel times between their PHIs and local county hospitals. We hypothesize that health-seeking behavior in western rural China tends to be spatially dependent, i.e., residents in remote areas rely more on PHIs for basic medical services due to a lack of accessibility to higher-level health facilities and therefore fewer health-seeking choices.

In general, five types of medical institutions are classified as PHIs in China, which are township healthcare centers (THCs), community healthcare centers (CHCs), outpatient departments, private clinics, and village clinics [17]. CHCs and THCs are the core primary care providers [18]. THCs and CHCs are similar in function but different in distribution; THCs are usually built in rural areas of China, while CHCs are usually built in urban areas of China. THCs and CHCs typically provide a wide range of basic medical services to local residents, including preventive care, diagnosis and treatment of minor illnesses, health education, and medication management. Compared to healthcare systems in other countries such as the UK and US, THCs and CHCs in China typically offer a more extensive range of healthcare services and are considered more accessible to the general population due to their location and coverage areas. The primary goal of these healthcare centers is to provide quality healthcare services that address the needs of the local population in a manner that is efficient, effective, and accessible to all. According to the China Health Statistical Yearbook in 2021, there were 1.03 million physicians in PHIs in 2020, with CHCs and THCs accounting for 14.4% and 30.5%, respectively [17]. As this study focused on PHIs in rural China, we selected THCs as a research sample.

We linked a dataset of medical utilizations in 923 PHIs in 2020 with drive time and geographic coordinates downloaded from the Gaode map (one of the leading map navigation providers in China) on July 4, 2020, and applied a multilevel linear model and geographically weighted regression to examine the spatial dependency of health-seeking behavior in western China. This study may enrich our understanding of the spatial disparity in medical service capacity and utilization of township healthcare centers in western China and attempt to fill the gap in understanding the factors that influence utilization of primary healthcare services in low-and middle-income countries. Our findings highlight the urgency of improving primary medical service capabilities in underdeveloped areas and further promoting a hierarchical medical system in developed areas.

2. Materials and Methods

2.1. Sample Selection and Data Sources

Shaanxi province, a developing province located in western China, was chosen as our study site. There are two parts of data sources used in this study. First, we collected healthcare service workload, workforce, and resources of PHIs and county-level hospitals from the Annual Report on Health Statistics (ARHS) of Shaanxi province in 2020. ARHS generates statistics from the health resources and medical services statistical survey, which is an annual administrative affair for all levels of Chinese health administration departments. This census survey covers all types of public and private healthcare institutions and is designed to collect basic information on healthcare institutions, including but not limited to operational status and resource allocation [19]. Of the 1535 THCs from 78 counties and county-level cities in Shaanxi province in 2020, a total of 923 THCs were included in this study based on the availability of information on key variables. The flow chart of THCs selection can be found in Figure A1 in Supplementary Materials. The spatial distribution of 923 THCs is shown in Figure A2 in Supplementary Materials.

Second, another essential part of the data in this study is a measure of the remoteness of THCs. We used the shortest driving time between THCs and local county hospitals to measure this remoteness. The data about the shortest driving time was downloaded from the route direction application programming interface (API) of Gaode Map (https://lbs.amap.com/api/webservice/guide/api/direction/), one of the leading map navigation providers in China, and the data collection process was performed by a self-developed crawler written in Python. The detailed process of data collection is described in our previous study [20]. The data collection strategy was as follows: (1) we firstly used the geocoding function provided by the geocoding API of Gaode Map to obtain the geographic coordinates of THCs based on their institution name, (2) the latitude and longitude of THCs and the local county hospitals were set as the starting and ending points, respectively, (3) Gaode Map offered the shortest driving time with a strategy that gave navigation speed priority but did not take the highway, and (4) we conducted two crawls on July 3, 2020, in the morning (start at 10:17 AM and end at 11:54 AM) and afternoon (start at 15:38 PM and end at 15:57 PM), respectively. Finally, a total of 1,846 (923 multiplied by 2 times) shortest driving times in minutes were collected for 923 THCs, and we used a histogram to show the distribution of the 1,846 shortest driving times (see Figure A3 in Supplementary).

The reason why we did not choose the road map in ArcGIS to calculate the travel time, such as in [21, 22], is because real-time travel cost metrics based on network map navigation are more accurate when considering traffic lights and traffic congestions [15]. Moreover, it is appropriate to use the real-time travel time between PHIs in a town and county-level hospital to evaluate the remoteness because we define the remote area at township level in our study and PHIs are located almost in the economic and population center of towns in rural China, which can reflect the average degree of remoteness of each village or household in the township.

2.2. Variable Measures

As this study focuses on the relationship between the remoteness of the location of THCs and the workload of medical services, we chose outpatient visit and inpatient visit to measure the workload, which are also the dependent variables in our study. It is a common and widely used practice to apply outpatient and inpatient visit to reflect the workload of medical services in a healthcare facility.

In order to mitigate the biases that may affect outpatient and inpatient visits of THCs other than remoteness of location, we controlled for a range of features of THCs, such as number of physicians, number of served populations, percentage of served populations aged 65 and above, government subsidies as a percentage of revenue [4], and type of THCs. Moreover, considering the possible crowding-out effect of local county hospitals on THCs [23–25], we included the number of available bed days in local county hospitals and county categories in regression models. For better presentation, we summarized the details about the values and attributes of the abovementioned variables in Table 1. Descriptive statistics for variables are presented in Table 2.

2.2.1. Empirical Strategy

Three parts of empirical strategies were used in this study. First, our baseline model uses an ordinary least squares (OLS) regression to estimate the impact of remoteness on the medical service workload of THCs. The model is shown as follows:where Y_ij is a measure of medical workload, the number of outpatient visits and inpatient visits, in THCs i in the county j. Remoteness_ij which denotes the shortest driving time from THCs i to the county j. X_ij is a set of THCs-level variables listed in Table 1. Z_j is a set of county-level variables that account for county characteristics, such as county category and number of hospital beds, and ɛ_ij is the error term. The coefficient β identifies our interest effect.

However, in terms of health management, county governments in China have a certain degree of decision-making autonomy in health policies such as resource allocation and market regulation, which leads to a situation where THCs follow a similar pattern within the same county. This means that THCs from the same county may be correlated and nested within the county. Then, considering the need to eliminate the bias caused by the correlation, we also used a two-level multilevel linear model (MLM) to estimate the coefficient β in the equation above. The first level in our dataset is THCs, and the secondary level is counties. We used intraclass correlation (ICC), calculated from an “empty” MLM model (only intercept, no predictors), to measure the correlation between pairs of THCs located in the same county. The ICCs for outpatient and inpatient services were 0.534 and 0.328, respectively, indicating that MLM was an appropriate choice in this study.

Further, both OLS and MLM so far assume homogeneous effects for each unit, regardless of geographic location, while the relationship between remoteness and THCs medical workload can be spatially heterogeneous and vary by location of the THCs [26–28]. To account for this spatial heterogeneity in THCs, we used geographically weighted regression (GWR), a derivative of OLS, to fit a series of locally OLS models for subsets of 923 THCs within a given bandwidth of each THC’s location instead of fitting a unique model to the entire 923 THCs. The subsets of 923 THCs were constructed based on the moving window method [29], which simply means that a search window (bandwidth) is used to cover the available THCs, moving from one THC location to another, and all other THCs around it and within the search window are identified as a subset. The equation of GWR in this study is shown as follows:where Y_i is a measure of the number of outpatient visits and inpatient visits of THCs, and i devotes the THCs, k devotes the number of independent variables, and X_ik means independent variables. is the spatial coordinates of THCs i, is the spatial weight kernel and is the weight matrix of THCs i based on and . We used adaptive Bi-square as spatial weights kernel and selected the optimal bandwidth by cross-validation. Fotheringham et al. [30] provided a comprehensive description of GWR.

Data manipulation and visualization were performed by using Python 3.8.5, and MLMs were performed using the lme4 [31] package in R 4.0.5 [32]. The standard errors of coefficients in OLS models are clustered at the county level. GWR was performed in the mgwr [33] module in Python.

3. Results

3.1. Exploratory Analysis

First, we mapped the outpatient visit and inpatient per service population of 923 THCs in scatter charts in geographic coordinates (Figure 1), which shows that THCs located near city boundaries had more outpatient visit and inpatient visit per service population than THCs in other locations. To clearly observe this characteristic, we aggregated outpatient visits and inpatient visits per service population of 923 THCs by 78 counties or county-level cities and plotted them in a geographic map. As illustrated in Figure 2, the average number of outpatient visit and inpatient visit per service population is greater in counties or county-level cities near the border of Shaanxi province than in counties or county-level cities located in the center of Shaanxi province. In general, the central areas of Shaanxi province, where the capital city is located, are the transportation, economic, and population center and have a low degree of geographic remoteness. In addition, we drew a scatter plot at the level of THCs with the shortest driving time on the horizontal axis and medical workload on the vertical axis to explore the relationship between these two variables, and then we obtained a slope that slopes upward to the right (see Figure A4 in Supplementary Materials).

(a)

(b)

(a)

(b)

3.2. Regression Analysis

The exploratory analysis indicated that the THCs located in remote areas of Shaanxi Province appeared to provide more medical services, and regression models quantified this effect while controlling for confounding factors. As shown in Table 3, the OLS coefficients of shortest driving time are consistent with or without controlling for covariates, which indicates that each 1-minute (1-hour) increase in shortest driving time between THCs and the local county hospitals was associated with an average 0.2% (12%) and 0.01% (0.6%) increase in outpatient and inpatient visits of THCs, separately.

As illustrated in Table 4, the MLM coefficients of the shortest driving time for outpatient visits of THCs are half their OLS coefficients after eliminating the bias caused by intracounty correlations, but there was no change in the coefficient for inpatient visits, which supports that each 1-minute (1-hour) increase in the shortest driving time between THCs and the local county hospitals was associated with an average of 0.1% (6%) and 0.01% (0.6%) increase in outpatient and inpatient visits for THCs, separately.

The basic information of the GWR models can be found in Table A1 in Supplementary Materials; only the significant variables in OLS and MLM were included as control variables in GWR, such as the number of physicians in THCs, service population, percentage of population over 65, share of subsidy on revenue, and number of available bed days in local county hospital. As presented in Table A1, the AICs of GWR are smaller than those of OLS and MLM, which indicates GWR removes the effect of spatial heterogeneity on regression coefficient estimates. Typically, the local coefficients for each unit in the GWR are presented as a map, as shown in Figure 3, where the coefficient for the shortest driving time for each THCs is marked in red (coefficient positive) and blue (coefficient negative) colors. A positive coefficient indicates that the longer the driving time or the more remote the THCs, the more outpatient and inpatient visits the THCs will have, and vice versa. We can clearly find that most THCs are located in the southern and northern Shaanxi province, where there are mountainous and developing areas, had positive coefficients, and this finding was more obvious in outpatient visits than in inpatient visits.

(a)

(b)

3.3. Robustness Analysis

Considering that the 2020 COVID-19 pandemic shocked urban and suburban areas of China more than rural areas, this leads to a possible differential impact of the 2020 COVID-19 pandemic on medical services in THCs, which would make the less frequent outpatient and inpatient visits in THCs with less remoteness confounded by the 2020 COVID-19 pandemic [34]. Therefore, we performed the same analysis using the 2018 data mentioned above to check the robustness of our findings and presented only the results of MLM in 2018 here; the rest of the results can be found in Table A2 in the Supplementary Materials. As shown in Table 5, the same findings can be observed in our 2018 sample, and the results are in line with the main analysis based on data in 2020.

4. Discussion

This study aims to examine the hypothesis that PHIs located in remote areas service more patients in China, and our findings support this hypothesis. Results of OLS, MLM, and GWR all point to a spatial disparity in medical service capacity of THCs in western China. Specifically, a one-hour increase in the shortest driving time between the THC and the local county hospitals was associated with an average increase of 6% and 0.6% in outpatient and inpatient admissions to THCs, respectively. This suggests that THCs located in remote areas provided more medical services, especially outpatient services.

Our findings are partly in line with some previous studies that found that the inpatient healthcare needs of patients in rural China were significantly influenced by the distance to the hospital; as the distance to the hospital increased, patients’ utilization of visits to that hospital decreased [35]. PHIs in areas with higher spatial accessibility of hospitals had a lower capacity of medical services, and residents were more likely to utilize inpatient services in hospitals [36]. Chinese patients prefer not to go to a primary care facility for outpatient services when they have the same spatial accessibility to hospitals and primary care facilities [37]. However, our study provides more insights into the impact of spatial factors on health-seeking behavior in Chinese residents. Specifically, patients in western China prefer to go to a primary care facility for outpatient services when they have less access to higher-level hospitals. In previous studies [38–40], concerns were raised about the lack of popularity of primary care in China. However, our findings highlight the possibility that these studies may have overlooked the geographical heterogeneity underlying this phenomenon. Although both urban and rural residents in China tend to seek medical treatment at large hospitals, residents living in remote rural areas in China are more likely to seek outpatient medical services at primary healthcare facilities due to restrictions imposed by their geographical environment. Although our study cannot distinguish whether this tendency is voluntary or forced, it is a fact.

We interpret the “spatial dependence effect of health-seeking behavior” in rural China by three effects. The first we call the substitution effect. It is well known that China has a regional disparity of health resource allocation between rural and urban areas; in terms of all kinds of health resources, urban areas are rich and rural areas are poor [41–44]. For residents in resource-rich areas, there are sufficient substitutions of healthcare facilities when residents are seeking medical care, and they are more willing to seek medical services in hospitals rather than PHIs [45]. However, remote areas in rural China are far away from public resource centers and have limited health resources. Consequently, there are insufficient substitute healthcare facilities for residents seeking medical care, and they have to rely on THCs that are of relatively high quality compared to village clinics.

The second effect we call the income effect, as income is another key factor affecting residents’ health-seeking behavior. In remote areas of China, residents have lower incomes, and a higher incidence of catastrophic health expenditures, and are more sensitive to healthcare spending [46]. In addition, the price of medical care in PHIs is the lowest in China’s healthcare delivery system; therefore, residents in remote areas used more medical services in THCs.

The third is the residential self-selection effect. Several studies have revealed that the neighborhood and built environment can impact residents’ travel [47, 48] and physical activity behaviors [49], and even unsafe walking circumstances due to high daytime and nighttime crime rates lead to higher diabetes incidence [50]. The same is true for health-seeking behaviors, where the transportation conditions and allocation of health facilities within a community may influence residents’ decisions to seek healthcare.

However, the above interpretations can only account for the fact that THCs in remote areas provide more medical services, but they cannot explain the difference between outpatient and inpatient services and why the effect on inpatient services is only one-tenth that of outpatient services. We think this difference can be explained by traditional Chinese culture. The Chinese traditionally believe that “if you have green hills, you will not be afraid to burn wood,” which means that the Chinese value life or health so much that when a life-threatening illness requires hospitalization, the sensitivity to expenses no longer has the highest priority. Although seeking inpatient care in high-level hospitals costs more in medical care, travel, and lost wages, it provides a higher quality of care than hospitalization in THCs, which is cheap but lower quality care. Compared to reduce spending, life is much more important, and regaining health and the ability to work as soon as possible are primary concerns.

Several policy implications can be drawn from our findings. First, more attention should be paid to improving the quality of THC in remote areas of China when constructing the capacity system for primary healthcare. Widespread gaps in the quality of primary healthcare are a long-term challenge for Chinese health policymakers. Building a primary healthcare-based integrated delivery system and strengthening the coordination between PHIs and hospitals to improve the quality of primary healthcare in China is a consensus [6, 51]. Our study suggests that remote areas deserve more attention and priority in the process of improving the quality of primary healthcare since the poor and vulnerable residents in remote areas rely more on primary care services. Specifically, the government can utilize telemedicine technology to connect remote communities with health care providers and provide professional development opportunities and mentoring programs to primary health care workers in remote regions to enhance their skills and knowledge. Second, health policy makers should take into account the influence of geographic factors on residents’ healthcare utilization and the supply of medical institutions and pay attention to the promotion function of infrastructure construction for public transportation, such as high-speed railways, on individuals’ health status, healthcare utilization, and household income [52–54].

Moreover, with the rapid development of telehealth in China―the 2021 China E-hospital Development Report showed that 1,004 Internet hospitals had been built in China by the end of 2020 [55]―the impact of telehealth on residents’ health-seeking behaviors cannot be ignored. Telehealth may weaken the effect of spatial factors on outpatient visits in THCs in China, as telehealth could break through the limits of spatial distance. Further research should investigate the long-term effects of telehealth on the relationship between spatial factors and health-seeking behavior.

This study has the following limitations: first, there is a lack of demand-side data to manifest comprehensive health-seeking behaviors, this prevents us from exploring in depth and explaining the mechanisms by which PHIs located in remote areas serviced more patients in China. Our findings need to be tested and supported by further future research by surveying residents in remote areas to obtain comprehensive information on their healthcare preferences, choices, and experiences. Second, we focus not only on the quantity but also on the quality of care at THCs in remote areas; however, due to the lack of data related to quality of care, our study could not go deeper to explain whether the increased outpatient and inpatient services of remote THCs led to better quality of care or whether residents in remote areas received better quality primary care services. Third, our sample was limited to THCs, one of five types of PHIs in China. Although THCs are the main primary care provider in rural China, our findings do not cover all medical services. Fourth, 612 THCs were dropped from our sample due to missing key variables, which resulted in our findings not reflecting the overall situation in Shaanxi Province and threatened the external validity of our findings, so relative caution should be exercised in extending our findings. Fifth, although the variation in travel time collected by one mode of transportation is sufficient to capture the variation in the remoteness of an area, different modes of transportation may have an impact on the variation in travel time and thus on the measurement of the variation in remoteness. Furthermore studies are needed to explore this impact.

5. Conclusions

Our findings suggest that THCs located in remote areas provided more medical services, especially outpatient services, which indicate that patients in western China prefer to go to a primary care facility for outpatient services when they have less access to higher-level hospitals. We suggest that more attention should be paid to enhance the quality of THCs in remote areas, which deserve a higher priority in the process of improving the quality of primary healthcare since the poor and vulnerable residents in remote areas rely more on primary care services. The provision of medical services in more remote areas could be improved through targeted telehealth interventions that address geographical and socioeconomic inequalities. Prioritizing greater investment in telehealth construction in remote areas may be an effective solution to promote the quality of primary healthcare and residents’ access to high-quality medical services. Health policymaking in developing regions such as western China should take geographic factors into account.

Data Availability

Restrictions are on the availability of the data and therefore are not publicly available but are available from the corresponding authors upon request.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

The authors thank the National Natural Science Foundation of China and Health Commission of Shaanxi Province for providing fund and data support. The authors also thank the China Scholarship Council for providing financial support to Chi Shen to visit Yale University. This research was funded by National Natural Science Foundation of China (72104195 and 71874137), China Postdoctoral Science Foundation (2021M702577), China Scholarship Council (520190628017), Career Development Award (K01AG053408), and a major research grant (R01AG077529) from the National Institute on Aging.

Supplementary Materials

Supplementary Table A1: geographically weighted regression model results. Supplementary Table A2: OLS model results of outpatient and inpatient visits per service population of THCs in 2018. Supplementary Figure A1: flow chart of data clean. Supplementary Figure A2: geographic distribution of 923 THCs in Shaanxi province. Supplementary Figure A3: histogram of the shortest driving time from 923 THCs to the local county hospital. Supplementary Figure A4: relationship between the shortest driving time and outpatient and inpatient visits per service population. (Supplementary Materials)

References

National Health Commission of the People’ s Republic of China, China Health Statistical Yearbook 2020, 5-1-2: Outpatient Services at Health Care Facilities in 2019 and 5-3-2: Inpatient Services at Health Care Facilities in 2019, Peking Union Medical College Publishing House, Beijing, China, 2020.
Q. Meng, A. Mills, L. Wang, and Q. Han, “What can we learn from China’s health system reform?” BMJ, vol. 365, Article ID l2349, 2019.
View at: Publisher Site | Google Scholar
X. Li, J. Lu, S. Hu et al., “The primary health-care system in China,” The Lancet, vol. 390, no. 10112, pp. 2584–2594, 2017.
View at: Publisher Site | Google Scholar
C. Shen, Z. Zhou, S. Lai et al., “Whether high government subsidies reduce the healthcare provision of township healthcare centers in rural China,” BMC Health Services Research, vol. 21, no. 1, p. 1184, 2021.
View at: Publisher Site | Google Scholar
X. Ma, H. Wang, L. Yang, L. Shi, and X. Liu, “Realigning the incentive system for china’s primary healthcare providers,” BMJ, vol. 365, Article ID 12406, 2019.
View at: Google Scholar
W. Yip, H. Fu, A. T. Chen et al., “10 years of health-care reform in China: progress and gaps in Universal Health Coverage,” The Lancet, vol. 394, no. 10204, pp. 1192–1204, 2019.
View at: Publisher Site | Google Scholar
Y. Shen and Y. Tao, “Associations between spatial access to medical facilities and health-seeking behaviors: a mixed geographically weighted regression analysis in Shanghai, China,” Applied Geography, vol. 139, Article ID 102644, 2022.
View at: Publisher Site | Google Scholar
J. Liu, H. Yin, T. Zheng et al., “Primary health institutions preference by hypertensive patients: effect of distance, trust and quality of management in the rural Heilongjiang province of China,” BMC Health Services Research, vol. 19, no. 1, p. 852, 2019.
View at: Publisher Site | Google Scholar
M. Jiang, Q. Fu, J. Xiong et al., “Preferences heterogeneity of health care utilization of community residents in China: a stated preference discrete choice experiment,” BMC Health Services Research, vol. 20, no. 1, p. 430, 2020.
View at: Publisher Site | Google Scholar
Q. Chen, D. Xu, H. Fu, and W. Yip, “Distance effects and home bias in patient choice on the Internet: evidence from an online healthcare platform in China,” China Economic Review, vol. 72, Article ID 101757, 2022.
View at: Publisher Site | Google Scholar
Y. Huang, P. Meyer, and L. Jin, “Neighborhood socioeconomic characteristics, healthcare spatial access, and emergency department visits for ambulatory care sensitive conditions for elderly,” Preventive Medicine Reports, vol. 12, pp. 101–105, 2018.
View at: Publisher Site | Google Scholar
W. Dotse-Gborgbortsi, D. Dwomoh, V. Alegana, A. Hill, A. J. Tatem, and J. Wright, “The influence of distance and quality on utilisation of birthing services at health facilities in Eastern Region, Ghana,” BMJ Global Health, vol. 4, no. 5, Article ID e002020, 2020.
View at: Publisher Site | Google Scholar
X. Gao and D. W. Kelley, “Understanding how distance to facility and quality of care affect maternal health service utilization in Kenya and Haiti: a comparative geographic information system study,” Geospatial Health, vol. 4, no. 1, 2019.
View at: Publisher Site | Google Scholar
S. Aggarwal, “The long road to health: healthcare utilization impacts of a road pavement policy in rural India,” Journal of Development Economics, vol. 151, Article ID 102667, 2021.
View at: Publisher Site | Google Scholar
G. Chen, C. C. Wang, P. Jin et al., “Evaluation of healthcare inequity for older adults: a spatio-temporal perspective,” Journal of Transport & Health, vol. 19, Article ID 100911, 2020.
View at: Publisher Site | Google Scholar
X. Li, H. M. Krumholz, W. Yip et al., “Quality of primary health care in China: challenges and recommendations,” The Lancet, vol. 395, no. 10239, pp. 1802–1812, 2020.
View at: Publisher Site | Google Scholar
National Health Commission of the People’ s Republic of China, China Health Statistical Yearbook 2021, Peking Union Medical College Publishing House, Beijing, China, 2021.
W. C. M. Yip, W. Hsiao, Q. Meng, W. Chen, and X. Sun, “Realignment of incentives for health-care providers in China,” The Lancet, vol. 375, no. 9720, pp. 1120–1130, 2010.
View at: Publisher Site | Google Scholar
National Bureau of Statistics, “National health resources and medical services statistical survey system,” 2021, http://www.stats.gov.cn/tjfw/bmdcxmsp/bmzd/202201/t20220121_1826888.html.
View at: Google Scholar
C. Shen, Z. Zhou, S. Lai et al., “Measuring spatial accessibility and within-province disparities in accessibility to county hospitals in Shaanxi Province of Western China based on web mapping navigation data,” International Journal for Equity in Health, vol. 19, no. 1, p. 99, 2020.
View at: Publisher Site | Google Scholar
J. Pan, H. Zhao, X. Wang, and X. Shi, “Assessing spatial access to public and private hospitals in Sichuan, China: the influence of the private sector on the healthcare geography in China,” Social Science & Medicine, vol. 170, pp. 35–45, 2016.
View at: Publisher Site | Google Scholar
X. Wang, H. Yang, Z. Duan, and J. Pan, “Spatial accessibility of primary health care in China: a case study in Sichuan Province,” Social Science & Medicine, vol. 209, pp. 14–24, 2018.
View at: Publisher Site | Google Scholar
M. Jia, F. Wang, J. Ma, M. Tian, M. Zhao, and L. Shen, “Implementation and early impacts of an integrated care pilot program in China: case study of county-level integrated health organizations in zhejiang province,” International Journal of Integrated Care, vol. 21, no. 3, p. 7, 2021.
View at: Publisher Site | Google Scholar
J. Wu, J. He, and G. Christakos, Quantitative Analysis and Modeling of Earth and Environmental Data: Space-Time and Spacetime Data Considerations, Elsevier, Amsterdam, Netherlands, 1st edition, 2021.
S. Yuan, F. Fan, J. van de Klundert, and J. van Wijngaarden, “Primary healthcare professionals’ perspective on vertical integration of healthcare system in China: a qualitative study,” BMJ Open, vol. 12, no. 2, Article ID e057063, 2022.
View at: Publisher Site | Google Scholar
L. J. Thomas, P. Huang, F. Yin et al., “Spatial heterogeneity can lead to substantial local variations in COVID-19 timing and severity,” Proceedings of the National Academy of Sciences, vol. 117, no. 39, pp. 24180–24187, 2020.
View at: Publisher Site | Google Scholar
S. Wang and J. Wu, “Spatial heterogeneity of the associations of economic and health care factors with infant mortality in China using geographically weighted regression and spatial clustering,” Social Science & Medicine, vol. 263, Article ID 113287, 2020.
View at: Publisher Site | Google Scholar
J. F. Wang, T. L. Zhang, and B. J. Fu, “A measure of spatial stratified heterogeneity,” Ecological Indicators, vol. 67, pp. 250–256, 2016.
View at: Publisher Site | Google Scholar
C. Wu, Y. Tu, Z. Li, and J. Yu, “An early assessment of the County Medical Community reform in China: a case study of Zhejiang province,” Journal of Chinese Governance, vol. 6, no. 4, pp. 463–485, 2021.
View at: Publisher Site | Google Scholar
A. S. Fotheringham, C. F. Brunsdon, and M. E. Charlton, Geographically Weighted Regression: The Analysis of Spatially Varying Relationships, John Wiley & Sons, Hoboken, NJ, USA, 2002.
D. Bates, M. Mächler, B. Bolker, and S. Walker, “Fitting linear mixed-effects models using lme4,” Journal of Statistical Software, vol. 67, no. 1, 2015.
View at: Publisher Site | Google Scholar
R Core Team, R: A Language and Environment for Statistical Computing, R Core Team, Vienna, Austria, 2021.
T. Oshan, Z. Li, W. Kang, L. Wolf, and A. Fotheringham, “Mgwr: a Python implementation of multiscale geographically weighted regression for investigating process spatial heterogeneity and scale,” ISPRS International Journal of Geo-Information, vol. 8, no. 6, p. 269, 2019.
View at: Publisher Site | Google Scholar
L. Xu, L. Zhuo, J. Zhang et al., “Impact of the COVID-19 pandemic on outpatient service in primary healthcare institutions: an inspiration from yinchuan of China,” International Journal of Health Policy and Management, vol. 11, no. 9, pp. 1926–1933, 2021.
View at: Publisher Site | Google Scholar
L. Li, J. Wang, and J. Yuan, “The impact of hospital distance on the Chinese inpatient service demand in rural areas: the application of discrete choice model (in Chinese),” Chinese Health Economics, vol. 33, no. 1, pp. 11–13, 2014.
View at: Google Scholar
Y. Zhang, Y. Niu, and L. Zhang, “Analysis on rural residents’ selection of the type of medical service in Macheng,Hubei Province (In Chinese),” Chinese Health Resources, vol. 21, no. 1, pp. 51–55, 2018.
View at: Google Scholar
Y. Li and Y. Xing, “Emperical study on the choice behavior of medical treatment in outpatients in China under the back⁃ ground of hierarchical diagnosis and treatment (in Chinese),” Chinese Hospital Management, vol. 40, no. 6, pp. 50–54, 2020.
View at: Google Scholar
S. Tang, H. Brixi, and H. Bekedam, “Advancing universal coverage of healthcare in China: translating political will into policy and practice: universal coverage of healthcare in China,” The International Journal of Health Planning and Management, vol. 29, no. 2, pp. 160–174, 2014.
View at: Publisher Site | Google Scholar
J. Xu, T. Powell-Jackson, and A. Mills, “Effectiveness of primary care gatekeeping: difference-in-differences evaluation of a pilot scheme in China,” BMJ Global Health, vol. 5, no. 8, Article ID e002792, 2020.
View at: Publisher Site | Google Scholar
W. C. M. Yip, W. C. Hsiao, W. Chen, S. Hu, J. Ma, and A. Maynard, “Early appraisal of China’s huge and complex health-care reforms,” The Lancet, vol. 379, no. 9818, pp. 833–842, 2012.
View at: Publisher Site | Google Scholar
F. Chen, X. Hao, and Z. Chen, “Can high-speed rail improve health and alleviate health inequality? Evidence from China,” Transport Policy, vol. 114, pp. 266–279, 2021.
View at: Publisher Site | Google Scholar
E. Dong, J. Xu, X. Sun, T. Xu, L. Zhang, and T. Wang, “Differences in regional distribution and inequality in health-resource allocation on institutions, beds, and workforce: a longitudinal study in China,” Archives of Public Health, vol. 79, no. 1, p. 78, 2021.
View at: Publisher Site | Google Scholar
L. Fu, K. Xu, F. Liu, L. Liang, and Z. Wang, “Regional disparity and patients mobility: benefits and spillover effects of the spatial network structure of the health services in China,” International Journal of Environmental Research and Public Health, vol. 18, no. 3, p. 1096, 2021.
View at: Publisher Site | Google Scholar
D. Li, Z. Zhou, Y. Si et al., “Unequal distribution of health human resource in mainland China: what are the determinants from a comprehensive perspective?” International Journal for Equity in Health, vol. 17, no. 1, p. 29, 2018.
View at: Publisher Site | Google Scholar
Y. Ta, Y. Zhu, and H. Fu, “Trends in access to health services, financial protection and satisfaction between 2010 and 2016: has China achieved the goals of its health system reform?” Social Science & Medicine, vol. 245, Article ID 112715, 2020.
View at: Publisher Site | Google Scholar
Y. Wang and C. Zhang, “A study on the risks and influencing factors of catastrophic health expenditure of rural poor families: based on the 2018 CHARLS data (In Chinese),” Chinese Journal of Health Policy, vol. 14, no. 1, pp. 44–49, 2021.
View at: Google Scholar
T. Lin, D. Wang, and X. Guan, “The built environment, travel attitude, and travel behavior: residential self-selection or residential determination?” Journal of Transport Geography, vol. 65, pp. 111–122, 2017.
View at: Publisher Site | Google Scholar
P. Zang, Y. Lu, J. Ma, B. Xie, R. Wang, and Y. Liu, “Disentangling residential self-selection from impacts of built environment characteristics on travel behaviors for older adults,” Social Science & Medicine, vol. 238, Article ID 112515, 2019.
View at: Publisher Site | Google Scholar
R. Wang, G. Grekousis, and Y. Lu, “Rethinking the link between the availability of neighborhood PA facilities and PA behavior: a comparison between private and public housing,” Building and Environment, vol. 207, Article ID 108401, 2022.
View at: Publisher Site | Google Scholar
T. Dendup, T. Astell-Burt, and X. Feng, “Residential self-selection, perceived built environment and type 2 diabetes incidence: a longitudinal analysis of 36,224 middle to older age adults,” Health & Place, vol. 58, Article ID 102154, 2019.
View at: Publisher Site | Google Scholar
S. Li, Y. Zhang, H. Ruan, E. Guerra, D. Burnette, and D. Burnette, “The role of transportation in older adults’ use of and satisfaction with primary care in China,” Journal of Transport & Health, vol. 18, Article ID 100898, 2020.
View at: Publisher Site | Google Scholar
J. Chen, Z. Lin, L. Li et al., “Ten years of China’s new healthcare reform: a longitudinal study on changes in health resources,” BMC Public Health, vol. 21, no. 1, p. 2272, 2021.
View at: Publisher Site | Google Scholar
G. G. Liu, C. Tang, Y. Liu, T. Bu, and D. Tang, “Will high-speed railway influence the healthcare seeking behaviour of patients? Quasi-experimental evidence from China,” Health Policy and Planning, vol. 36, no. 10, pp. 1633–1643, 2021.
View at: Publisher Site | Google Scholar
Y. Wang, N. Luo, and G. Zhou, “Is the opening of high-speed railway conducive to improving residents’ health? (In Chinese),” Journal of Finance and Economics, vol. 46, no. 9, pp. 92–107, 2020.
View at: Google Scholar
National Telemedicine and Connected HealthCare Center, “China e-hospital development report,” 2021, https://zk.cn-healthcare.com/doc-show-53644.html.
View at: Google Scholar

Copyright

Copyright © 2023 Chi Shen et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies