Research on the Spatiotemporal Characteristics of Health Collaboration Development in China's Urban Agglomerations

——An Empirical Analysis Based on Four Major Urban Agglomerations

Wentao Zhu¹, Xiangfei Li¹* (Corresponding Author), Xumin Zhu²

(Note: Both of us are co-first authors)

1. School of Economics and Management, Tianjin Polytechnic University

2. School of Public Administration and Policy, Renmin University of China

AbstractObjective: This study focuses on China's four most politically significant and economically dynamic urban agglomerations, namely the Beijing-Tianjin-Hebei (BTH), Yangtze River Delta (YRD), Pearl River Delta (PRD), and Chengdu-Chongqing (CY) urban agglomerations, as research units. It explores their health development trends between 2005 and 2022 and compares their spatiotemporal characteristics in coordinated health development. Methods: An integrated index construction method and an improved urban gravity model were used to build urban health network models within these urban agglomerations, investigating the coordinated relationships and features in their health development. Conclusions: Over the 17-year period, China's four urban agglomerations exhibited both commonalities and unique differences in health coordination outcomes. The degree of coordinated development showed a significant strengthening trend but was noticeably influenced by major macro policies and public health events. Core cities within each urban agglomeration had substantial radiating effects on regional health coordination. The PRD and YRD demonstrated more pronounced growth in coordination intensity over time compared to BTH and CY. In terms of coordination patterns: BTH displayed a unipolar radiation structure, YRD exhibited a polycentric network structure, PRD showed a core-periphery structure, and CY presented a dual-core segmented policy-driven coordination structure.

Keywords: urban agglomerations; health; coordinated development; regional integration

Introduction

Against the backdrop of deep integration between globalization and urbanization, the connections between regional cities have become increasingly close.Globally,several highly integrated development regions have gradually emerged,represented by metropolitan areas and urban clusters¹.^{Theseregions typically revolve around one or more megacities, encompassing at least three major cities,and rely on efficient infrastructure networksto form urban spatial clusters characterized by compact spatial structures, tight economic linkages, and converging governance mechanisms23.Such urban clusters have not only become crucial platforms for internal resource integration and public policy coordination within nations but are also transforming into new spatial units of global competition4.WithChina's accelerated urbanization process leading to population concentration, economic restructuring, and social relationship reconfiguration,issues such as rising infectious disease transmission risks, imbalanced healthcare resource distribution, and lack of health equityhave all imposedhigherdemands on public health collaboration across large regions.In China'surban cluster regional integration process,the coordinated development in the health sector faces numerous complex challenges.Spatially,resource allocation imbalance is the primary challenge, with high-quality medical resources overly concentrated in core cities, while surrounding cities and rural areas suffer from outdated medical facilities and talent shortages,leading to excessive patient influx into major cities and exacerbating medical resource strain.Additionally, varying levels of economic and social development result in significant disparities among cities within urban clusters in terms of health financing, health resources, and resident health levels.}

The Chinese government has made numerous efforts to promote coordinated development in urban health clusters. For instance, the Beijing-Tianjin-Hebei urban cluster issued the "2025 Work Plan for High-Quality Integrated Development of Healthcare in Tongzhou-Wuqing-Langfang," aiming to establish unified standards for disease diagnosis codes, clinical medical terminology, and inspection norms, thereby facilitating information sharing and operational coordination among medical institutions. The Yangtze River Delta urban cluster has encouraged large public hospitals to alleviate high-quality medical resources to under-resourced areas like Anhui through the formation of medical alliances and groups, cooperative construction, or trusteeship. The Pearl River Delta urban cluster has focused on opening up and the health industry, closely promoting cooperation in health among Guangdong, Hong Kong, and Macao. Leveraging the technological and industrial advantages of the Greater Bay Area, it has fostered innovative development in the health industry and established international centers for technological innovation, among other initiatives. Althoughthe Chinese government's effortshave achieved significant accomplishments, it must be acknowledged that within urban clusters,different cities exhibitnotabledisparitiesin areas such as healthcare policies, medical management, and health standards.These administrative barriersalsohinder the free flow and sharing of medical resources, placingcertain pressureson the coordinated development of regional health.

The past two decades have witnessed the rapid rise and development of urban clusters in China. Systematically reviewing and summarizing the coordinated development processes of representative urban clusters in the health sector is crucial for gaining deeper insights into the characteristics of China's urban cluster development. It also provides valuable references for optimizing regional medical planning and constructing efficient and balanced regional health systems. To this end,this studyselects China's most representative and economically vibrantBeijing-Tianjin-Hebei, Yangtze River Delta, Pearl River Delta, and Chengdu-Chongqing urban clusters as research areas.By constructing a coordinated development index and a health gravity model between cities, the study explores and compares the processes and characteristics of coordinated health development in these four urban clusters from temporal and spatial dimensions.It revealsthe evolutionary patterns and mechanismsof coordinated health development in China's major urban clusters, offering theoretical support for optimizing regional medical planning and building efficient and balanced regional health systems.

Literature Review

2.1 Regional IntegrationDevelopmentand Public HealthDisparities

Within urban clusters, core cities often concentrate high-quality public health and medical resources, not only meeting local demands but also serving as regional hubs for resource distribution. The cross-regional flow and functional spillover of resources are key indicators of enhanced collaborative health security capabilities in urban clusters. Rational allocation of resources within the region will significantly improve the accessibility of medical and health resources across the entire area, making it necessary to further consider regional medical resource security capabilities from a holistic perspective under urban collaboration⁵.In these urban regions, medical resources are highly networked and coordinated among cities⁶. Core cities in these areas hold overwhelming advantages in policy, talent, economic strength, and medical standards, benefiting not only themselves but also neighboring cities. For example, among China's 125 national regional medical center construction projects, Beijing and Shanghai have respectively exported 50 and 20 projects to surrounding cities⁷, achieving a certain degree of radiation effect for high-quality medical services. However, public health services within urban clusters still exhibit structural imbalances. On one hand, core cities possess significant advantages in policy, talent, financial investment, and medical standards; on the other hand, non-core cities and peripheral regions face substantial gaps in resource supply, health facilities, emergency response capabilities, and more. The inter-regional coordination mechanisms remain underdeveloped, forming an asymmetric "single-core—multi-periphery" structure.

Existing research has foundthat under the context of cross-departmental and cross-regional collaboration,theambiguity of responsibilities and the absence of performance evaluation lead to inefficiencies in coordination. For example, the imbalanced development of urban functions in China (emphasizing economy and infrastructure while neglecting social services) has resulted in lagging public health⁸. China's grassroots emergency response capabilities in the medical and public health systems are weak⁹. In Beijing, the accessibility of secondary and tertiary hospitals for low-income communities is only one-fourth that of high-income communities, and the time required to reach primary healthcare institutions is twice as long for low-income groups¹⁰. Moreover, primary healthcare institutions provide limited medical services. Shanghai's central urban areas also face insufficient coverage of walkable primary healthcare services, with 22.8% of communities experiencing resource shortages. Elderly populations and low-income families face exacerbated healthcare access difficulties due to socioeconomic barriers¹¹¹². A significant portion of healthcare resources is concentrated in developed regions, such as Beijing and Zhejiang, leading to resource surpluses, while areas like Anhui and Jiangxi have a Gini coefficient as high as 0.88 when measured by service area¹³.

Overall, although urban agglomeration integration has promoted resource consolidation and institutional coordination, significant imbalances and fragmentation persist in public health services. Existing literature primarily focuses on the macro-level structure of urban agglomerations, with insufficient comparative analysis of different types of cities within them,especially non-core cities. There is a lack of systematic research on the coordination mechanisms ofnon-corecities. This paper will adopt a micro-level perspective, focusing on the role differences of various cities in coordinated health development, aiming to provide theoretical support and empirical analysis for achieving balanced resource allocation and health equity within urban agglomerations.

2.2 Factors Influencing Regional Public Health Collaborative Development

At the regional level, the collaborative development of public health is not a one-dimensional "expansion of quantity" but a multidimensional networked evolution process influenced by various internal and external factors. Different elements interact and couple, driving the deepening of cross-regional public health partnerships and the integration of service systems, thereby forming a more cohesive and dynamically evolving regionalhealthcollaboration network.

In recent years, existing studies have thoroughly revealed the multidimensional factors influencing the spatiotemporal evolution of public health collaboration in urban agglomerations or regions, particularly highlighting the critical role of transportation, technology, and policy interventions in resource flows and functional spillovers.Table 1 summarizes these influencing factors.

Table 1 xxxxx

From the perspective of transportation factors, transportation infrastructure and population mobility directly affect cross-city medical visits and the spatial accessibility of healthcare resources. Wang & Nie (2024), based on cross-city medical visit data from Chinese urban clusters, pointed out significant disparities in the distribution of critical care medical resources among different cities, with more developed transportation networks leading to more pronounced patient mobility. Dalen (2019), through analyzing medical tourism populations in the United States, revealed the spatiotemporal characteristics of international medical mobility. Zhang F. et al. (2023) further linked transportation density to the spread of epidemics, emphasizing the high correlation between population mobility and the diffusion of infectious diseases.

At the technological level, the digital economy, internet usage frequency, and the level of healthcare informatization have significantly improved the efficiency and equity of public health governance. Zhao X. et al. (2024) validated the positive effect of the digital economy on public health efficiency through large-scale data. Yang X. (2024) compared the differential impact of internet usage on health levels between urban and rural residents, further noting that uneven informatization could exacerbate the fragmentation of regional health coordination. Additionally, Haft & Allen (2024)'s study on the CRISP health information exchange platform provided empirical evidence that information sharing promotes coordination in primary healthcare.

From the perspective of policy intervention, institutional arrangements and the effectiveness of collaborative governance networks play an indispensable role in ensuring resource integration and service equalization. For example, Rangachari et al. (2024) evaluated the pathways for social determinant interventions under the U.S. Affordable Care Act; Jiao & Zhang (2025) analyzed the cross-regional, multi-sectoral collaborative governance mechanisms in Japan's nuclear-contaminated water incident; and Guo T. et al. (2025) demonstrated through empirical research on the Beijing-Tianjin-Hebei integrated healthcare policy that institutional innovation can effectively promote resource sharing and optimize supply-demand dynamics within the region.

However, existing research on regionalpublic healthcollaboration under current spatiotemporal dimensionstends tofocus narrowly on single elements, with most studies approaching the topic from singular perspectives such as transportation, information dissemination, or policy intervention, lacking a systematic description of themulti-element coupling mechanismsin inter-city healthcare resource collaboration. Therefore, this paper will construct ananalytical framework integrating multiple elementsto provide a more explanatory theoretical model for the precise allocation of regional medical resources and cross-city collaborative governance.

2.3 Research Methods forRegionalPublic HealthCollaborative Development

In recent years, spatial measurement technology, as a crucial tool for revealing the patterns of public health collaborative development, has achieved significant progress both theoretically and practically.

Spatial measurementas a crucial analytical toolhas been widely appliedin recent yearsin the field of public health.For example,Deng et al. studied the Guangdong-Hong Kong-Macao Greater Bay Area, using the space-time cube (STC) model and emerging hotspot analysis to confirm the stress effect of high-temperature environments on public health security, emphasizing the need to embed heat island mitigation strategies in urban planning to reduce population health risks²⁷. Li employed the Dagum Gini coefficient and spatial econometric models to analyze medical service supply across 31 Chinese provinces from 2012 to 2020, revealing regional disparities characterized by "convergence in the east and west, lagging in the central region"²⁸;Tang and Tan's four-dimensional indicator system found that factors such as soil erosion control and population density directly influence the risk resilience of public health systems²⁹;spatiotemporal evolution models play a key role in simulating epidemic spread and allocating medical resources. For instance, case distribution predictions based on kernel density estimation can improve emergency supply delivery efficiency by 35%,whilespatial autocorrelation analysis highlightsthe need to strengthen vaccination coverage in key areas³⁰.

In regional public health research, spatial scale measurement has become a key methodological approach for revealing the patterns of health resource allocation and service coordination. Existing studies primarily focus on using hotspot detection, spatial clustering analysis,regional disparity indices,spatial regression modeling, and other techniques to explore the spatial heterogeneity of medical resources, population health risks, and health service utilization.

In existing research, hotspot detection and spatial clustering identification are among the common analytical methods. Deng et al. (2023) took the Guangdong-Hong Kong-Macao Greater Bay Area as an example, combining the Space-Time Cube model with Emerging Hot Spot Analysis to reveal the spatial distribution characteristics of high-temperature stress and residents' health risks, providing spatial guidance for regional health adaptation policies. Similarly, Li et al. (2020) used Kernel Density Estimation to simulate the distribution of COVID-19 cases and identified spatially disadvantaged areas with insufficient vaccination coverage through Local Moran's I, improving the efficiency of resource allocation in pandemic emergency responses (Li et al., 2020). In terms of regional disparities and imbalance measurement, the spatial Gini coefficient, Theil index, and Dagum Gini decomposition method are widely used to characterize the regional differentiation patterns of medical resource allocation. Li (2024) measured the medical service supply capacity of China's 31 provinces and found a long-standing pattern of "high in the east, low in the west, and lagging in the central region," suggesting that service equity should be improved from the perspective of regional coordination mechanisms.

Composite indicator construction and spatial regression modeling have also become research hotspots. Tang and Tan (2022) proposed an urban resilience indicator system integrating ecology, population, governance, and services, and used the Spatial Durbin Model to analyze its impact on public health risk resilience, finding that variables such as soil and water conservation and green space coverage play a structural role in regional health resilience. Additionally, GIS-supported spatial visualization and multilevel regression analysis have been used to model the spatial transmission pathways between macro policies and micro health behaviors.

However, most current studies tend to adopt single-scale static measurements, making it difficult to reveal the dynamic synergistic relationships among cities at different levels within urban agglomerations and the transmission pathways between macro policies and micro health risks. Therefore, this study utilizes geographic information systems to construct a multi-scale spatial network model, combining spatial autocorrelation analysis to identify agglomeration and spillover effects at different scales.

2.4 Research Gaps

In summary, currentresearch predominantly focuses on the spatiotemporal changes inhealthcarecapacity within individual regions or cities, overlooking the impact of inter-city collaborativedevelopmentonregionaldevelopment as a whole,andneglecting the connections between large and small cities within the region. This studyconstructsa city network model for public health to describethe collaborative status and characteristics of public health among cities in China's four majorurban agglomerations, providing theoretical support for optimizingregional healthcare planning.Compared to existing research,themaincontributions of this study are as follows:

(1) Focusing on China's four most dynamic urban agglomerations, this study constructs a spatial gravity structure among cities in the field of public health, exploring for the first time the characteristics of collaborative healthcare development in China's major urban agglomerations from a spatiotemporal perspective.

(2) Fully considering China's political characteristics, the study examines the leading role of major events, key policies, and politically influential core cities in regional healthcare collaboration.

(3) It also investigates the role of non-core cities in healthcare collaboration, focusing on internal structural differences within urban agglomerations, and reveals the spatiotemporal network characteristics exhibited during the collaborative development process of China's four major urban agglomerations.

Research Methodology

3.1 Study Area

In recent years, with the deepening of urbanization in China, urban agglomerations such as the Yangtze River Delta (YRD), Pearl River Delta (PRD), and Beijing-Tianjin-Hebei (BTH) have emerged as core engines of national regional development and crucial carriers of regional integration strategies. The coordinated development within these urban agglomerations is gradually shifting from a singular focus on economic agglomeration to a multidimensional collaborative governance model encompassing social security and public health³¹. This study covers health data from four key regions: the Beijing-Tianjin-Hebei urban agglomeration (BTH), Yangtze River Delta urban agglomeration (YRD), Pearl River Delta urban agglomeration (PRD), and Chengdu-Chongqing metropolitan area (CY). These four study areas represent large and medium-sized urban agglomerations in China, including municipalities directly under the central government, cities with independent planning status, provincial capitals, and other major cities. As shown in Figure 1, these regions have large populations and substantial economic scales. Their achievements in coordinated healthcare development have played a vital role in advancing China's overall health sector.

Fig.1. China's economy distribution

Further selecting the most economically developed prefecture-level cities within each urban agglomeration as research samples, including 13 cities in the Beijing-Tianjin-Hebei urban agglomeration, 26 cities in the Yangtze River Delta urban agglomeration, 9 cities in the Pearl River Delta urban agglomeration, and 16 cities in the Chengdu-Chongqing urban agglomeration, totaling 64 cities. According to the official administrative hierarchy standards for Chinese cities³²³³, the cities in the study area are categorized. The classification criteria are represented by the city symbols in Table 2, and the types of sample cities in the four urban agglomerations of this study are shown in Table 3. (The table labeling order is incorrect; please verify.)

Table 2. The classification criteria for Chinese cities

Table 3. The basic information of the cities in the studied regions

3.2Constructinga Healthy Urban Network Model

This study attempts to build a network model between cities based ondifferenturban health development levels to analyze their coordinated development.

First, construct an Urban Health Index (UHI) that represents the health development level of each citytoanalyze the developmental changes of different cities from 2005 to 2022;

Next, the study analyzes the synergy level of health development among different cities within the urban agglomeration and constructs a Regional Health Collaborative Development Index (HCDI) to reflect the internal synergy level of the urban agglomeration.

On this basis, the synergy consistency of health development levels among different cities is analyzed using an improved urban gravity model;

Finally, an in-depth analysis is conducted on the constructed urban network model, extracting and summarizing four structural characteristics of the spatiotemporal collaborative development of urban agglomeration health.

This study, referencing Evangelista et al.³⁴, Al-Ghamdi et al. (2023)³⁵, Korir (2024)³⁶, selects representative health development indicators, including metrics reflecting urban health development levels such as the quantity of health resources, health development status, and local economic development levels. Additionally, indicators influencing inter-city synergy are chosen, such as distance between cities, travel time, and policy support for collaborative development, as shown in Table4:

Table 4. Required indicators for the study

This study utilizes data primarily sourced from the China Health Statistical Yearbook, China City Statistical Yearbook, as well as statistical yearbooks and publicly available data from individual cities and their respective provinces; policy documents are mainly obtained from the official websites of local health commissions and medical security bureaus. The health-related policies in this study refer to government-issued policy documents closely related to coordinated health development. The selection criteria for policy samples are as follows: first, ensuring the policy text content is closely related to health, with systematic searches using keywords such as "health coordination," "medical coordination," and "healthcare cooperation"; second, filtering government-issued laws, regulations, and measures, primarily selecting local normative documents while excluding those irrelevant to the research objectives.

3.2.1 Urban HealthLevel Measurement

To better measure the urban health development level in this study, this research constructs the Urban Health Index (UHI) based on the indicator data inTable3.Following the method of Sun et al. (2020)⁴², positive and negative indicators were standardized separatelyto obtain $$ .After determining the weights , the Urban Health Index for each city was calculated. For ease of subsequent calculations, the final result is expressed as:

3.2.2 Regional Health Collaborative Development Index

Based on the previous calculations and analysis of various indicators, and considering the impact of each indicator on regional health collaboration, the Analytic Hierarchy Process(AHP)method was used to score each indicator and construct a judgment matrix⁴³. On this basis,the Health Collaborative Development Index (HCDI) for each region from 2014 to 2022 was derived:

In the equation $$ ,theith region'sjth standardized indicator value, isthejth indicator's weight for each year, the number of indicators involved in calculatingUHIfor each city,andn=7; isthe element in theith row andjth column of the AHP judgment matrix A, used to calculate weights, representsthe coefficient related to theith indicator.

3.2.3 Construction of the Health-Based Urban Gravity Model

Based on Newton's law of gravitation and complex network theory, the health-based urban gravity model is constructed, with the formula as:

Among them, , represents the health index of cityiandj; represents the spatial distance between cities; represents the travel time distance; is the gravitational constant (set to 1), =1.8is the distance decay coefficient, is the time decay threshold. This model breaks through the limitation of traditional gravity models that only consider spatial distance by introducing the time dimension and an exponential decay function, making it more aligned with real-world collaborative mechanisms.Drawing on the research ofWeiet al.⁴⁴, we introduce  as theresource complementarity adjustment factor, where:  reflects the similarity in health levels between two cities,when the UHI of two cities is similar (e.g., Beijing and Tianjin, both with relatively highUHI), this ratio approaches 1, and the adjustment factor tends toward , suppressing the gravitational value; when the UHI of two cities differs significantly, this ratio approaches 0, and the adjustment factor tends toward 1, leaving the gravitational value unsuppressed. is the competition coefficient, representing the weakening effect of resource similarity on collaborative gravity,with a value of 0.3.

3.2.4 Urban Health Network Model

Using Pearson correlation coefficients and social network analysis (SNA), we construct a dynamic collaborative network,calculating the Pearson correlation coefficient of gravitational values between cities , with the formula:

Among them, $$ represents the gravitational value of city i to j at time t; , is the mean value. Subsequently, degree centrality is calculated to reveal the characteristics of the network structure. The formula for degree centrality is: . Here, indicates the connection strength between cityiandj(with values of 0 or 1).In summary,the indicators and their meanings involved in the construction of the urban health network modelare shown in Table 4:

Table. 4.Variable Symbol Definition

3.4 Model Validation and Robustness Analysis

This study employed the leave-one-out cross-validation (LOOCV) method to verify model stability. Specifically, data from one urban agglomeration was excluded each time, while the model was trained using data from the other three urban agglomerations to predict the collaborative gravitational values of the excluded agglomeration. The root mean square error (RMSE) and coefficient of determination ( ) were calculated. Results showed that all RMSE values were below 0.05, and values exceeded 0.92, indicating strong model generalization capability. Sensitivity analysis in this study involved varying the distance decay coefficient (1.5-2.0) and time threshold T (±20%) to observe changes in gravitational values. Results demonstrated that when changed by ±0.1, gravitational value fluctuations remained below 3%; when changed by ±20%, fluctuations were less than 5%, indicating low sensitivity to parameter variations and strong model robustness.

Research Findings

4.1Regional Health Development and Coordination Level

From Figure2, it can be observed that the public healthlevelsof the four urban clusters exhibit similar trends—maintaining an upward trajectory before 2020, followed by a significant decline after 2020 due to the impact of amajorpublic health emergency.Figure 3 reveals thatduringthe period from 2015 to 2019, theRegional Health Collaborative Development Index (HCDI)(HCDI)of the four study areas showed a notable upward trend. This is closely linked to the successive introduction of collaborative development policies inChinaduring this period, such as the "Beijing-Tianjin-Hebei Collaborative Development Plan Outline" in 2015, the "Yangtze River Delta Regional Integration Development Plan Outline" in 2019, the "Pan-Pearl River Delta Regional Deepening Cooperation Joint Declaration (2015-2025)" signed in 2014, and the "Chengdu-Chongqing Urban Cluster Development Plan" in April 2016. Following the implementation of these collaborative policies, the public health collaboration indices of the four major urban clusters experienced substantial growth compared to previous years, driven by robust national-level coordination policies. This demonstrates that national policy support has played a powerful role in advancing public health development within specific regions.

Fig.2. Changes in UHI across urban agglomerations

Fig.3. Spatiotemporal changes and coordination index of health development levels in urban agglomerations

4.1Network Structure of Coordinated Health Development in Urban Agglomerations

Based on the calculation results obtained from the previously constructed health collaboration gravity model,using ArcGIS Map and Origin 2024 software,the urban network visualization is performed, as shown inFigure4.(I've said countless times to provide some explanation for the figure! The meaning of urban nodes, gravity lines, and regional colors—add a few sentences.)

The overall spatial distribution of each urban agglomeration reflects the development status of health services among cities within the region, as well as the coordinated health development among cities before and after the implementation of collaborative policies. It can be observed that before and after policy implementation, the UHI index of key cities within the urban agglomeration is higher than that of other cities in the region, indicating that the public health development of these key cities has consistently maintained a leading position in the region. With the implementation of collaborative policies and improvements in areas such as technology and transportation, the influence of these key cities on other cities in the region has grown increasingly stronger, exerting a certain radiating effect on the public health development of surrounding cities.

However, it is equally important not to overlook the role of geographical distance in regional health collaboration. The collaborative pull between most cities tends to exhibit lower intensity levels when the distance between them is greater, whereas relatively closer cities—especially between major cities and others within the region—demonstrate higher levels of collaborative intensity, such as Beijing and Langfang.

Within urban agglomerations, the higher the transportation accessibility between cities, the greater the level of synergistic intensity exhibited by the urban gravitational lines. Additionally, in terms of temporal evolution, the gravitational pull among cities within urban clusters continues to strengthen, indicating a gradual enhancement of regional healthcare synergy over time. Compared to the period from 2014 to 2018, the synergistic intensity of medical resources across all urban agglomerations grew more significantly between 2018 and 2022, demonstrating rapid development in inter-city medical collaboration and further strengthening of interaction and coordination in medical resources during this period by 45%.

Fig.4.The Status of Health Coordination in Four Urban Agglomerations

Table 4illustrates the fundamental characteristics of regional public health collaboration across four urban clusters. The comparison reveals that the Pearl River Delta (PRD) exhibits relatively higher intensity of medical resource coordination among its cities, while other regions show smaller disparities in inter-city medical resource coordination—a pattern evident through the comparison of mean and median gravity values across the studied areas.

Table 4. The basic characteristics of each region in HCGM

It can be observed that the largest city group in the Beijing-Tianjin-Hebei urban agglomeration is Beijing-Langfang, while the largest city group in the Yangtze River Delta urban agglomeration is Shanghai-Ningbo; the largest city group in the Pearl River Delta urban agglomeration is Dongguan-Zhongshan; and the largest city group in the Chengdu-Chongqing urban agglomeration is Chengdu-Ya'an.

4.2 The Role of Urban Nodes in Coordinated Health Development

Figure5 displays the proportion of each city's gravity value relative to the total regional value, revealing that key cities in each region account for a significant share. In the YRD, Shanghai, as the most central city in the region, holds the highest proportion, but the gap with the second to fourth cities is relatively small, indicating its influence in coordinated public health development is not particularly prominent. In other urban clusters, the proportion of the most central city exceeds 20%, suggesting these cities play a more crucial role in regional health coordination. However, in the PRD, Guangzhou and Shenzhen dominate in public health coordination, both exceeding 30%, with Guangzhou's proportion surpassing 50% at its peak.

Figure 6 illustrates theurban clusters' average gravity index changes from 2005 to 2022. It shows that key cities like Shanghai and Beijing have significantly higher gravity indices than others in their regions, underscoring their unshakable position in regional public health coordination. Thus, these cities play a more substantial role in intra-regional collaborative development.

Fig.5. The proportion of key cities in spatiotemporal evolution

Fig.6. The average gravitational value in Four Urban Agglomerations

4.3 Spatial Characteristics of Health Collaborative Development in Urban Agglomerations

Combining the previously calculated inter-city , and plotting Figure 7, it more intuitively demonstrates the relationship of public health collaboration within each study area—particularly between non-core cities in the region. Figure 7 displays the correlation of inter-city health collaboration gravitational values ( ) across different urban agglomerations. Red indicates positive correlation, with darker red representing higher values, while blue indicates negative correlation, with darker blue representing lower values. The color intensity reflects the magnitude of the gravitational value and the strength of correlation. From the inter-city correlation characteristics, the YRD exhibits the strongest synergy with a globally high-concentration positive correlation network, while the CY, with its mix of red and blue blocks and low synergy among core city pairs, emerges as the region with the weakest collaboration. By integrating Figure 7 with Figure 4, Figure 8 was generated, revealing the internal public health collaborative development structure of different urban agglomerations.

Fig.7. Heat Maps of Correlation in

Fig.8. The Collaborative Structure

ThroughFigures 7 and 8,analysisrevealsthe collaborative development structures and characteristics of the four major urban clusters between 2014 and 2022.The details are as follows:

The Beijing-Tianjin-Hebei (BTH) region exhibits a unipolar radiating collaborative structure. Beijing forms strong synergies (dark red) with neighboring cities such as Langfang, Baoding, and Zhangjiakou, while cities farther from Beijing in the region also maintain notable connections (light red). This indicates high gravitational values between Beijing and other regional cities, forming a strong linkage axis centered on Beijing, reflecting the spillover effects of capital resources. Meanwhile, Tianjin, another major city in the region, is marginalized. Its weak collaboration (blue) with surrounding cities, such as Tangshan and Cangzhou, reflects functional overlap with Beijing, leading to suppressed influence over neighboring cities. As the capital of Hebei Province, Shijiazhuang faces similar challenges as Tianjin. Additionally, its economic level is far below that of Beijing and Tianjin, and due to industrial homogeneity and intra-provincial competition, it lacks coordination, resulting in a collaborative vacuum with eastern Hebei. This creates a pronounced "strong around Beijing" spatial differentiation, with the BTH region overall displaying a pattern of "one dominant core, weak secondary cores, and fragmented peripheries."

The Yangtze River Delta (YRD) regionexhibitsa polycentric networked collaborative structure.The deep red coloration in the YRD indicates a high overall synergy, reflecting the region's strong collaborative dynamics.Within the region, a robust tri-core collaborative network has formed between Shanghai-Nanjing-Hangzhou, with core cities demonstrating significant radiating effects on neighboring areas. Intra-provincial collaboration outperforms cross-regional coordination, while remote cities show weaker synergy. Centered on Shanghai, a deep red triangular zone emerges with Nanjing and Hangzhou, where gravitational values reach 14.8 (Shanghai-Nanjing) and 8.6 (Shanghai-Hangzhou), with Nanjing and Hangzhou also displaying red connectivity. Among neighboring cities, Suzhou and Shanghai (100 km apart) show deep red (high gravitational value), while distant cities like Chizhou (over 300 km) appear light blue. Intra-provincial city pairs such as Suzhou-Wuxiexhibit red linkages, indicating strong collaboration. This demonstrates that in the YRD region, beyond traditional major cities, economically advanced cities can also become focal points for regional public health collaboration.

The Pearl River Delta region (PRD) exhibits a core-periphery collaborative structure. The dual cores of Guangzhou and Shenzhen form ultra-strong collaborative nodes, while neighboring cities display a "core-periphery" gradient radiation structure due to geographical proximity. Collaboration among other cities within the same province is active, but the gap in collaboration between distant cities and the core areas is significant. The connection between Guangzhou and Shenzhen appears in deep red (gravitational value 56.4), representing the strongest regional collaboration pair. Neighboring cities like Shenzhen-Dongguan and Guangzhou-Foshan also appear in deep red, while distant cities such as Zhaoqing and Jiangmen show noticeably lighter colors, indicating weaker collaboration with core cities.

The Chengdu-Chongqing region (CY) exhibits a dual-core segmentation-policy-driven collaborative structure. Compared to other urban clusters, CY appears lighter in color with more blue hues, indicating weaker collaboration relative to other regions. The collaboration between the Chengdu-Chongqing dual cores is weak. While Chengdu shows localized red areas with policy-bound neighboring cities (e.g., Ya'an), market-driven collaboration (such as Chengdu- Ziyang) is insufficient; non-core cities generally appear blue, showing clear lack of collaboration. The grid cells for Chengdu and Chongqing are notably lighter, significantly lower than other regional core city pairs. Chengdu and Ya'an appear red (gravitational value 9.68), but Chengdu and Ziyang appear light blue (gravitational value < 10). Secondary cities within Sichuan, such as Nanchong and Luzhou, appear blue with almost no collaborative signals.

Conclusion

Based on the above research findings, this paper summarizes some correlations between regional public health collaborative development and key regional cities, which may have certain implications for studying regional public health collaborative development.

The PRD exhibits a higher degree of public health collaborative development, while the YRD demonstrates more balanced public health collaboration.Research on the public health collaborative development across China's four major urban clusters reveals that the PRD has the highest level of synergy, likely due to the close integration and relatively short distances between its cities. According to Amap data, the average straight-line distance between cities within the PRD is 55.51 km, significantly shorter than the BTH's 183.68 km, the YRD's 240.61 km, and the CY's 102.21 km. Additionally, the PRD's cities boast robust economic development, and with further improvements in transportation infrastructure, intercity connections within the cluster will strengthen, thereby enhancing regional public health collaboration. Meanwhile, the YRD shows smaller disparities in public health collaboration among its cities, presenting a more balanced state compared to other urban clusters. As the largest and most economically advanced among the four studied regions, the YRD includes multiple cities like Shanghai, Nanjing, and Suzhou, each with GDP exceeding one trillion yuan, fostering more equitable development. This relative balance—particularly in contrast to the PRD and BTH—better equips the region to withstand shocks from public health emergencies⁴⁶, as ripple effects from crises in core cities would have severe negative impacts on non-core cities within the region.

Beijing and Chengdu exert significant influence over public health collaboration within their respective urban clusters.Figure 3 shows that both Beijing and Chengdu had the highest urban health indices in their clusters even before collaborative policies were implemented, with their gravity indices accounting for over 20% of their regions' totals—far surpassing the second-ranked cities. The implementation of collaborative policies will further amplify their regional influence, given Beijing's role as the political center and Chengdu's status as a key regional city. However, when such cities face major public health risks, other cities in the BTH and CY will experience severe ripple effects. Despite the presence of municipalities like Tianjin and Chongqing, administrative boundaries and jurisdictional constraints limit their positive regional impact⁴⁷, underscoring the importance of decentralizing Beijing's non-capital functions and promoting regional collaborative development.

The coordinated development of public health in the YRD region exhibits a multi-point pattern, while other regions show a trend of concentration around major cities.As mentioned earlier, the YRD region has numerous economically strong cities. Beyond Shanghai as the core-level city, other key cities within the region also boast excellent healthcare standards, thereby driving the development of surrounding smaller cities. In other urban clusters, there is a tendency to further enhance the proportion of core cities in coordinated development, thereby promoting the growth of the entire urban agglomeration. This is particularly evident in the BTH and CY regions, which is somewhat related to their relatively uneven development.

The collaborative health relationships between cities of different scales exhibit distinct structures.The collaborative health relationships between cities of different scales are influenced by multiple factors. The differences in collaboration among large, medium, and small cities are shaped by administrative barriers, geographical distance, economic capacity, and policy supply. Optimizing regional coordination requires breaking cross-regional institutional obstacles, strengthening urban resource allocation, and leveraging policy adjustments to enhance market dynamics, thereby improving coordination balance and resilience. The closeness of collaboration between large cities is directly related to administrative barriers and geographical distance. Core cities in the Yangtze River Delta and Pearl River Delta regions, due to tight policy linkages and geographical proximity, have formed highly efficient collaboration models, such as SH-NB and GZ-SZ, with gravitational values reaching 14.8 and 56.4, respectively. In contrast, cross-provincial city pairs like CD-CQ and BJ-SJZ face limitations in medical resource allocation due to differences in healthcare policies and greater distances, with gravitational values around only 27%, highlighting the constraints of administrative fragmentation.Corecitiesand non-core citiesare influenced by the "inverse square law of distance," policies, and economic levels in their collaboration, while non-core citiesrely on administrative hierarchies and economic linkages for coordination. Cities within the same province or industrial cluster achieve resource sharing through administrative coordination or industrial demand, whereasHD-HS and NC-LS, due to resource scarcity and lack of specialized planning, exhibit weak collaborative momentum, with low or even negative gravitational values.

Limitations and Prospects

However, the current research inevitably faces certain limitations when exploring the coordinated development of medical resources among the four major urban agglomerations. The primary issue lies in data constraints, which led this study to focus primarily on the coordination of medical resources at the prefecture-level city scale. As the integration process within urban agglomerations accelerates, medical interactions between smaller regions are becoming increasingly frequent, and public demand for nearby, efficient medical services is growing more urgent. This suggests that future research needs to delve into smaller regional scales to capture more nuanced dynamics of medical coordination. Secondly, this study defined regional medical resources using only two core indicators: the number of hospital beds and the number of practicing (assistant) physicians, representing a broad conceptualization of medical resources. In reality, within China's urban medical system, hospitals of different types and levels play distinct roles, particularly tertiary Grade A hospitals, which serve not only as central forces in regional medical resource coordination but also as key factors influencing the distribution and utilization of medical resources. Additionally, in the retrieval and collection of relevant documents from different regions, varying interpretations of regional coordination were observed, necessitating further refinement of this indicator in future research. Moreover, during the implementation of coordination policies, different regions may issue their own related policies, and these policy orientations directly impact the findings of this study. Finally, it is noteworthy that in the quantitative analysis of the balance between supply and demand for regional medical resources, significant disparities exist in the demand for medical resources among different cities within the same region or among different social groups within the same city. This complexity was not fully reflected in the current study, presenting new perspectives and challenges for future in-depth research.

Author contributions ：XL proposed the topic and designed the overall article. WZ processed and analyzed the data. XZ summarized the data. All authors contributed to the article and approved the submitted version.

Funding ：This work was supported by the National Social Science Fund of China (Grant Number 22BGL223).