Spatial accessibility and equity of primary healthcare in Zhejiang, China

Abstract

Background

Enhancing the accessibility and equity of primary healthcare (PHC) is a crucial objective of China’s healthcare reform. However, spatial barriers remain a significant factor contributing to the inequitable access to PHC services among residents.

Objective

This study aims to quantify the spatial accessibility (SA) and evaluate the equity of PHC resources in a pilot province for healthcare reform, and its municipalities, thereby providing insights that can be generalized to the broader context of China.

Methods

The study used the navigation function provided by Gaode Map to estimate the time it takes for residents to visit PHC institutions. The two-step floating catchment area (2SFCA) method, weighted by a Gaussian function, was employed to measure the SA of PHC institutions, general practitioners (GPs), and beds across various residential areas. The Gini coefficient was utilized to assess disparities in SA among different regions. Additionally, Getis-Ord Gi* analysis was conducted to visualize these spatial disparities.

Results

The study analyzed 25,601 residential and 1,451 healthcare points in Zhejiang Province, revealing significant disparities in SA of PHC. Urban residents reached PHC institutions faster than rural ones (7.05 ± 4.7 min vs. 9.17 ± 7.78 min, P < 0.001). Within 30 min, 98.4% of residential points and 99.1% of the population accessed PHC institutions. Disparities in PHC resources were notable, with Lishui having the highest SA for institutions and GPs, and Shaoxing for beds. Equity assessment showed high inequity for institutions (Gini 0.553), moderate for beds (Gini 0.497), and reasonable for GPs (Gini 0.332)​. Getis-Ord Gi* analysis demonstrated that areas further from urban centers were more likely to exhibit clusters of hotspots or cold spots.

Conclusion

The study highlights substantial disparities in SA of PHC and equity across Zhejiang, underscoring the need for strategic resource distribution. Future research should include diverse transportation modes and more precise demand point data to enhance understanding of accessibility dynamics.

Introduction

Primary health care (PHC) serves as the cornerstone for the sustainable development of national health systems [1, 2]. Strengthening PHC is also a key component of the “Health for All” agenda outlined in the Alma-Ata Declaration [3]. The World Health Organization (WHO) defines PHC as a whole-of-society approach to health, emphasizing that it must be equitably distributed and as close as possible to resident’s everyday environments [4]. Empirical evidence further demonstrates that improving the accessibility of PHC not only reduces overall health expenditures but also mitigates health-related inequalities [5,6,7]. Therefore, the establishment of a person-centered, equitable, and accessible PHC system is a key objective in health system reforms worldwide. Since 2009, China has launched a nationwide healthcare reform aimed at providing safe, effective, accessible, and affordable medical services to more than one billion residents [8]. A key focus of this reform has been enhancing the accessibility and equity of PHC [9]. The role of PHC institutions in China includes the diagnosis and treatment of common diseases, management of chronic illnesses, and prevention of infectious diseases. Under the current hierarchical medical system and the national basic medical insurance scheme [10], patients are encouraged to seek initial consultations at PHC institutions. General practitioners (GPs) in these institutions assess patients’ conditions and determine whether to provide treatment at the PHC level or refer patients to higher-level hospitals for more complex care. This approach aims to improve the efficiency of local health services and reduce overall healthcare costs.

Disparities in access to PHC are a well-documented global health issue, particularly pronounced in low- and middle-income countries [11]. Spatial barriers significantly contribute to these disparities, a view increasingly supported by research. Authoritative evidence indicates that as travel time increases, the likelihood of utilizing healthcare services decreases [12]. Consequently, residents with varying travel times to healthcare facilities inevitably experience inequities in healthcare accessibility. To quantitatively analyze these issues and develop strategies for ensuring the spatial justice of public health resource distribution, the concept of spatial accessibility (SA) has been introduced into the health domain and integrated into medical geography. SA refers to the potential or opportunity for individuals to obtain desired resources within a given spatial context [13]. Initially based on Reilly’s 1931 gravity model for commercial distribution [14], SA has evolved through various analytical approaches, including transportation network-based models introduced by Hansen in 1959 [15]. Advances in Geographic Information Systems (GIS) have refined SA measurement methods, such as kernel density estimation [16], the two-step floating catchment area (2SFCA) method [17], and composite gravity models [18]. These methodologies have further stimulated the application of accessibility studies across different fields, enhancing our understanding of spatial disparities in healthcare access and informing policy interventions aimed at improving equity.

Recent studies provide clear evidence of social inequalities in the spatial distribution of healthcare providers, including PHC providers. Studies from Malaysia and Indonesia have demonstrated significant spatial disparities in PHC access opportunities in surveyed rural areas [19, 20]. Similarly, studies from economically developed regions such as Canada, France, the United States, Poland, and Hong Kong indicates pronounced differences in SA to PHC among residents of various areas [12, 13, 21,22,23].

For China, a rapidly developing country with a population of 1.4 billion, the equitable distribution of PHC resources remains a challenge despite significant progress in healthcare reform. Issues such as inadequate PHC resource allocation and substantial urban-rural disparities have been widely reported [24]. Previous studies have attempted to explore the spatial accessibility (SA) of PHC in various regions of China (Hainan [25], Guangzhou [26], Shenzhen [27], Sichuan [28], Shanghai [29], and Beijing [30]) and have provided valuable insights into regional inequalities in PHC distribution and the spatial characteristics of resource allocation. However, these studies have limitations, including narrow research areas, low precision of indicator data, and reliance on single-method approaches, such as vague demand point locations, rough estimations of travel impedance, municipalities-level focus, and the use of singular methods for interpreting PHC accessibility.

Building on the experiences of previous studies, this research focuses on a healthcare reform pilot province where similar studies have not yet been conducted. More precise indicator data were collected (locations of residential points and travel times), and a comprehensive approach (including the 2SFCA, the Gini coefficient, and Getis-Ord Gi*) was employed to thoroughly assess the SA of PHC. This approach aims to offer valuable insights for informing local healthcare reform strategies.

Methods

Study area

Zhejiang Province, situated in the southeastern coastal region of China and an integral part of the Yangtze River Delta Economic Zone, was selected as the study area for this research. During the study period, Zhejiang comprised 11 municipalities with a population nearing 65 million and a per capita disposable income of 7596 USD (ranking third nationwide). In 2016, Zhejiang was designated by the Chinese government as a pilot province for healthcare reform, serving as a model to inform reforms in other similar regions. Given its unique geographic and economic advantages, achieving an equitable distribution of public health resources has been a critical focus of the province’s ongoing development efforts. Therefore, selecting Zhejiang Province as the study area is highly representative and practically valuable.

Given that this study aims to explore the accessibility of PHC from a spatial perspective, it is essential to introduce some spatial information about the sample area, including natural topography, transportation networks, and population distribution (as illustrated in Fig. 1). Zhejiang Province’s natural topography is characterized by mountainous regions in the southwest, alluvial plains in the northeast, coastal hills and plains in the east, and the Jin-Qu Basin in the central area. Consequently, the population and transportation networks are predominantly concentrated in the relatively flat plain areas. These regions are also the most economically developed in the province and serve as key destinations for internal migration.

Fig. 1

Characteristics of the study area. A natural topography; B transportation networks; C population distribution

Study data

Point-of-supply data

This study included a total of 1,451 PHC institutions across the province, comprising 973 township health centers and 478 community health centers. We focused exclusively on PHC services provided by township and community health centers, as these institutions are considered the main components of PHC in China and are responsible for delivering the majority of such services [28]. We selected three key indicators: institution, GP, and bed. The institution reflects the distribution and coverage of primary care services. The GP indicates the capacity for providing continuous and comprehensive care. The bed represents the institutional capacity to manage inpatient care needs. Data on these indicators were obtained from the Zhejiang Provincial Health Commission, the governing body overseeing health affairs in the province. The data represent the status as of the end of 2020. Additionally, we utilized the free Gaode Coordinate Picker to retrieve the latitude and longitude coordinates of each institution based on their names and addresses, preparing for subsequent spatial analysis. The specific spatial distribution of PHC institutions is illustrated in Fig. 2A.

Fig. 2

The spatial distribution of primary healthcare institutions (A) and residential points (B)

Point-of-demand data

For the purpose of analysis, this study selected the locations of community committees and village committees as the demand points (residential points) for PHC. These are the smallest administrative organizations in urban and rural China, generally characterized by self-governance. Additionally, to facilitate the provision of public services to residents, these committees are typically located as close as possible to the centers of communities or villages. We obtained the names of 25,601 villages and communities in Zhejiang Province from the National Bureau of Statistics website. According to the classification by the National Bureau of Statistics, there are a total of 5,231 residential points located in urban areas and 20,370 residential points in rural areas. Supplementary Table 1 provides detailed information on the number of urban and rural residential points across different cities. We wrote code in the Python programming environment to interface with the Gaode Maps (AutoNavi Map) open platform Application Programming Interface (API), enabling us to batch retrieve their coordinates (Fig. 2B). Additionally, the population data was sourced from the publicly available WorldPop website and calibrated using the Zhejiang Provincial Statistical Yearbook as of 2020. Typically, population data are aggregated at the township or sub-district level (the administrative level above villages or communities). To estimate the population for each village, we calculated the average population of the corresponding township (i.e., the total township population divided by the number of villages).

Travel time data

Travel time to PHC institutions is a crucial parameter for calculating SA. Some studies simulate travel time based on transportation networks. However, due to the complexity of real-world traffic, this method can lack precision. In this study, we utilized Gaode Maps, a navigation website, to obtain estimated driving travel times. The underlying logic of this calculation involves entering the starting and ending locations into the website, which then provides an estimated arrival time. We developed code in the Python programming environment to interface with the platform’s free route planning API, allowing us to batch process estimated travel times from residential points to nearby PHC institutions. Additionally, we measured travel times during peak traffic periods and regular periods, taking the average of these two measurements to represent the actual residents’ travel time.

Research methods

Two-step floating catchment area method(2SFCA)

This study employed the widely used 2SFCA method to measure the SA of PHC for residents. This method, developed by the renowned American scholar John Radke, is a spatial decomposition approach based on the floating catchment area technique, aimed at assessing the accessibility of public services [31]. The principle of this method involves two steps. In the first step, each supply point (e.g., PHC institution) is used as the center to search for demand points within its catchment area, calculating the supply-to-demand ratio to evaluate the service capacity of each supply point. In the second step, each demand point (e.g., residential location) is used as the center to search for supply points within its threshold distance. The sum of the supply-to-demand ratios of all identified supply points is then calculated to determine the SA for each demand point. A detailed description of this method can be found in a previous study [17].

The subsequent section presents the 2SFCA calculation formulas. Formula (1) calculates the supply-to-demand ratio for supply points, while Formula (2) computes the SA for demand points.

$${R}_{j}=\frac{{S}_{j}}{\sum\:_{k\in\:{\{d}_{kj}\le\:{d}_{0}\}}{D}_{i}}$$

(1)

$${A}_{i}=\sum\:_{j\in\:{\{d}_{ij}\le\:{d}_{0}\}}{R}_{j}$$

(2)

Where, A_i is the SA scores of demand point i; R_j is the supply-to-demand ratio of supply point j within the threshold time d₀ (set to 30 min here [17]); S_j is the service capacity or size of supply point j, such as the number of GPs or beds; D_i is the demand intensity of demand point i, which in this context refers to the population size; d_i is the travel time from demand point k to supply point j.

One limitation of the 2SFCA method is that all demand points within the threshold distance are considered to have the same level of SA (e.g., residents 1 min away and those 29 min away have the same SA), this calculation does not accurately reflect reality. Therefore, an attenuation function needs to be introduced to differentiate the SA for different travel times within the catchment area. According to prior research, the Gaussian function is commonly used as the attenuation function, featuring an “S” shaped decay that accelerates initially and then slows down as distance increases (Formula 3) [32].

$$G\left({d}_{ij}\right)=\left\{\begin{array}{c}\frac{{e}^{-\frac{1}{2}\times\:{\left(\frac{{d}_{ij}}{{d}_{0}}\right)}^{2}}-{e}^{-\frac{1}{2}}}{1-{e}^{-\frac{1}{2}}}\\\:0{,d}_{ij}\ge\:{d}_{0}\end{array}\right.,\:{d}_{ij}\le\:{d}_{0}$$

(3)

Equity assessment

This study utilizes the Gini coefficient to evaluate the spatial equity of accessibility to rural cooperative medical centers. Tao ZL [33] used various measures to analyses the spatial equity of urban medical facilities. The findings indicated that the Gini coefficient is more appropriate for assessing spatial equity. The calculation formula is as follows:

$$Gc=\frac{\sum\:_{ij}\left|{A}_{j}-{A}_{i}\right|}{2{n}^{2}\stackrel{-}{A}}$$

(4)

where A_i and A_j are the SA scores of any two demand point i and j, respectively; $\:\stackrel{-}{A}$ is the average SA in the study area; and n is the total number of units. Gini coefficient reflects the degree of deviation between the distribution of SA and uniform distribution, with a range of 0 to 1. A higher Gini coefficient indicates lower SA equity. Based on previous research [34, 35], A Gini coefficient below 0.2 indicates “high equity” or “absolute equity” in SA distribution; 0.2 ≤ Gc<0.3 indicates “relative equity”; 0.3 ≤ Gc < 0.4 indicates " proper equity”; 0.4 ≤ Gc < 0.5 indicates “significant disparity”; and Gc ≥ 0.5 indicates “high inequity”.

Getis-Ord Gi*

This study utilizes the Getis-Ord Gi* analysis function in ArcGIS 10.3 to identify the spatial clustering characteristics of accessibility hot spots and cold spots across various residential areas [36]. To facilitate the observation of internal conditions within each municipality, we conducted separate Getis-Ord Gi* analyses for each municipality and combined the results for display. This analytical method compares the degree of accessibility aggregation in local areas relative to their surrounding regions at specific significance levels. The results categorize hot spots and cold spots into seven intervals based on significance levels of 0.01, 0.05, 0.1, and > 0.1 (Not Significant). For example, a result of Hot spot-99% Confidence indicates that the accessibility of a particular residential area is highly clustered with 99% confidence.

Statistical analysis

We conducted a significance test to examine the differences in travel times between rural and urban residential areas using SPSS 25. Given that the two samples did not follow a normal distribution, the Mann-Whitney U test was employed to assess whether the observed differences were statistically significant. A p-value of less than 0.05 was considered statistically significant.

Results

Coverage area

The study encompassed 25,601 demand points (residential points) and 1,451 supply points (PHC institutions) across Zhejiang Province, resulting in 466,477 valid healthcare pathways. The analysis revealed that urban residents reach the nearest PHC institution significantly faster than rural residents (7.05 ± 4.7 min vs. 9.17 ± 7.78 min, P < 0.001). Regarding coverage within different time frames, 89.5% of residential points and 92.2% of the population can reach a PHC institution within 15 min. These values increase to 98.4% and 99.1%, respectively, within 30 min, and a 60-minute travel time covers all residential points. Table 1 details the coverage rates of demand points and population across various municipalities for different travel times. The results indicated that Hangzhou, Huzhou, Jiaxing, Quzhou, and Shaoxing had over 99.0% of their residential points and population covered within 30 min. Additionally, more than 90% of residential points and population in Hangzhou, Jinhua, Ningbo, Quzhou, Shaoxing, and Taizhou were within a 15-minute coverage area. It was also observed that, except for Jiaxing, the coverage rate of residential points was lower than that of the population in other municipalities, suggesting that areas with poor accessibility were concentrated in less populated communities or villages. The coverage areas for different travel times are visually represented in Fig. 3. From a geographical distribution perspective, the majority of the sample areas are visually covered within a 15-minute travel time. Regions requiring travel times exceeding 15 min are primarily located in the western mountainous areas of the sample, including parts of Hangzhou and Quzhou, as well as the southwestern mountainous areas of Lishui and some northeastern islands in Zhoushan.

Table 1 Coverage of demand points and population by travel time to primary healthcare institutions in Zhejiang province municipalities

Fig. 3

Coverage areas at different travel times for primary health care

Results of SA

The SA of PHC resources across Zhejiang Province is summarized in Fig. 4. At the provincial level, the average SA scores for institutions, GPs, and beds are 0.022, 0.333, and 0.464, respectively. Substantial disparities exist among municipalities regarding PHC resource accessibility. Lishui boasts the highest SA for PHC institutions at 0.094, while Jiaxing has the lowest at 0.013. Hangzhou, Jiaxing, Ningbo, Shaoxing, Taizhou, and Wenzhou all fall below the provincial average in institutional SA. In terms of GP accessibility, Lishui again ranks highest with an SA of 0.603, contrasting with Wenzhou’s lowest SA of 0.189. Huzhou, Jiaxing, Taizhou, and Wenzhou are below the provincial average for SA of GP. For SA of bed, Shaoxing leads with an SA of 1.062, whereas Zhoushan trails at 0.103. Hangzhou, Huzhou, Ningbo, Taizhou, Wenzhou, and Zhoushan all exhibit bed SA scores below the provincial average.

Fig. 4

Spatial accessibility of different primary healthcare resources in 11 municipalities. A primary healthcare institutions; B general practitioners; C beds

Equity assessment

The Gini coefficients for the SA of PHC resources across various municipalities in Zhejiang Province are presented in the Table 2. At the provincial level, the equity of institutions, GPs and beds varies considerably. The Gini coefficients are 0.5534, 0.3322, and 0.4973, respectively, indicating high inequity for institutions, proper equity for GPs, and significant disparity for beds. Specifically, Lishui, Wenzhou, and Zhoushan exhibit the highest inequity for institutions, GPs, and beds, with Gini coefficients of 0.559, 0.403, and 0.793, respectively. In contrast, Jiaxing demonstrates the highest equity for institutions and beds, with Gini coefficients of 0.195 and 0.273, respectively, while Jinhua shows the highest equity for GPs, with a Gini coefficient of 0.213.

Table 2 Equity assessment of spatial of different primary healthcare resources at the provincial and municipal levels

For institutions, 6 out of 11 municipalities exceed the warning line, with 2 municipalities reaching the “high inequity” threshold. In terms of GPs, only 1 out of 11 municipalities surpasses the warning line, and 8 municipalities fall within the “relative equity” range. For beds, 8 out of 11 municipalities exceed the warning line, with 3 municipalities classified under “high inequity.”

Result of Getis-Ord Gi*

This study builds upon equity assessments to conduct significance tests on the SA of residential points and visualizes the results spatially. Figure 5 presents the findings from the Getis-Ord Gi* analysis. For the three indicators, while most residential points exhibit a random distribution, there are quite variations in the patterns of hot and cold spots. Specifically, cold spots for PHC institutions are primarily located in a few residential points in the northern and southern regions of the study area, while hot spots are concentrated in areas distant from city centers. GP cold spots cover more extensive areas, with some even present in the central urban area of Jinhua; hot spots are similarly situated in peripheral regions. The distribution of bed hotspots is more extensive, showing a clear clustering trend, with some hotspots even appearing in the urban centers of certain municipality.

Fig. 5

Getis-Ord Gi* statistics of spatial accessibility of primary healthcare resources. A primary healthcare institutions; B general practitioners; C beds

Discussion

Enhancing PHC accessibility is a crucial task in China’s healthcare reform and a foundational goal for the “Healthy China 2030” strategy [37]. This study employs spatial analysis methods to investigate the SA and equity of PHC in a pilot healthcare reform province. The study reveals three main findings. Firstly, more than 90% of residents in the study area can reach the nearest primary healthcare institution within a 15-minute drive, although significant urban-rural disparities are observed, with urban residents reaching the nearest institution 2.15 min faster on average than rural residents. Secondly, there are substantial differences in the SA of PHC resources among the various municipalities within the province. Notably, Quzhou and Lishui, which have lower population and economic levels, exhibit higher SA levels. Lastly, the equity of SA for PHC institutions and beds is at levels of significant disparity or high inequity, with these inequities primarily driven by the extreme SA levels in areas far from urban centers.

On average, urban residents in Zhejiang Province can reach PHC institutions by car 2.12 min faster than rural residents. However, this urban advantage was not observed in the results of the SA under Getis-Ord Gi* analysis. This finding contrasts with previous studies, which have suggested that PHC accessibility is significantly higher in city centers compared to suburban areas [26, 27]. The discrepancy may be attributed to differences in local population distribution, resource allocation, and methodological approaches. This phenomenon in this study can be attributed to the fact that more than 70% of the population is concentrated in urban areas [38]. When population serves as the denominator in accessibility calculations, the relative accessibility of PHC institutions for urban residents is not markedly superior to that of rural residents. Consequently, no significant hotspots are observed within urban areas. Contrarily, certain rural regions with superior transportation infrastructure and lower population density display markedly higher PHC accessibility, resulting in distinct hotspot concentrations.

In a comparative analysis across different municipalities, most exhibit higher population coverage rates than residential point coverage rates. This discrepancy suggests that, within a 30-minute driving range, some sparsely populated communities or villages remain unable to access nearby PHC institutions. Getis-Ord Gi* analysis reveals that these residential points (cold spots) predominantly lie near county borders, where transportation infrastructure is generally inadequate. Furthermore, substantial disparities in SA are evident among different municipalities, with these variances manifesting differently across resource indicators. According to equity assessments, SA of GP demonstrates superior equity compared to beds and institutions. This implies that, despite the suboptimal equity in institutional SA (institutions are carriers of GPs and beds), effective GP resource allocation can enhance equity in SA to a relatively reasonable level. Specifically, Lishui and Quzhou exhibit higher accessibility for all three resource indicators than the provincial average, as does Zhoushan for institutional and GP resources. This may be attributed to the higher per capita resource provision in these areas. Similarly, studies conducted in other regions of China have also identified disparities in the spatial distribution of PHC resources [25, 28, 30]. Some of these studies have further quantified disparities using analysis tools such as the Gini coefficient, highlighting significant inequalities or excessive variations in the SA of certain PHC indicators [28, 30]. Although the choice of indicators and methodologies in these studies differs from those employed in our research, the evidence consistently underscores the urgent need to address inequalities in PHC accessibility across regions in China.

Intriguingly, Lishui, Quzhou, and Zhoushan are among the economically least developed municipalities in Zhejiang Province, suggesting that SA of PHC does not positively correlate with economic status [39]. However, internal equity within Lishui and Zhoushan is concerning, with the equity of institutions and beds reaching levels of high inequity. The natural terrain and transportation network maps of Zhejiang Province may provide an explanation for this. Western Quzhou and Lishui’s extensive mountainous terrain and Zhoushan’s archipelagic geography create natural barriers, exacerbating travel difficulties for residents accessing PHC institutions, thereby contributing to these disparities. Nonetheless, such severe disparities are not observed in SA of GP, indicating that strategic resource allocation optimization can ameliorate equity in these regions. In contrast, the SA equity of PHC in Huzhou and Jiaxing, located in the northern part of Zhejiang Province, is relatively favorable, with both municipalities exhibiting proper to high levels of equity. This can be attributed to the relatively dense transportation network and the concentrated distribution of residential points in these areas. However, the overall resource accessibility in both municipalities falls below the provincial average, indicating a mismatch between the available resources and the local population size. This underscores the need for an overall increase in resource allocation to better align with the population demands. The hotspot analysis results for these two municipalities reveal that the majority of hotspots are still concentrated in areas far from urban centers.

The study’s findings have significant policy implications for enhancing the SA of PHC. First, the convenience for rural residents in accessing nearby PHC institutions is lower than in urban areas. The need for improved planning and renovation of PHC facilities in villages to address this gap. Additionally, some urban areas exhibit inadequate SA due to mismatched resource allocation relative to local population sizes. Increasing resource distribution in these regions is essential to enhance overall SA equity. Furthermore, the relatively higher equity observed in GP accessibility suggests that the allocation of PHC beds should be guided by the distribution of general practitioners to ensure balanced resource availability. Lastly, adopting comprehensive evaluation methods for the regular monitoring of PHC resource distribution will enable timely and effective policy adjustments.

This study demonstrates several strengths and limitations. It enhances accuracy in measuring SA by employing the navigation function of map websites, which integrates real-time traffic data to account for factors such as traffic signals and congestion—considerations often overlooked by traditional transportation network models. Additionally, the analysis is refined to the community or village level, providing more precise results through a comprehensive methodological approach that includes accessibility evaluation, equity assessment, and spatial analysis. Despite these strengths, the study is not without limitations. The potential overestimation of SA in low car-usage areas due to the exclusive use of driving as the transport mode. The use of village and neighborhood committee locations as proxies for demand points provides a relatively accurate estimate, but may not fully capture the true distribution of residential points. Moreover, while travel times were averaged across different periods, discrepancies with actual conditions may still exist due to variations in the timing of healthcare visits. Finally, the findings, derived from a pilot reform region with favorable resource allocation and economic conditions, may have limited generalizability to other provinces.

Conclusion

In conclusion, this study underscores the complex dynamics of SA of PHC in Zhejiang Province. While urban residents generally enjoy shorter travel times to PHC institutions, the significant concentration of the population in urban areas neutralizes this advantage, resulting in comparable accessibility levels between urban and rural regions. The analysis reveals substantial disparities in PHC resource accessibility across different municipalities, with economically disadvantaged areas like Lishui, Quzhou, and Zhoushan demonstrating higher accessibility due to more favorable resource allocation. However, internal inequities, particularly in institutional and bed accessibility, remain a concern. These findings highlight the critical need for strategic resource distribution to enhance overall and equitable access to PHC services, considering the unique geographic and infrastructural challenges of each region. Future studies should incorporate diverse transportation modes and more precise demand point data to provide a comprehensive understanding of accessibility dynamics.