2.1.1. Compression of total pollutant emissions driving sustained improvements in residents’ health outcomes.

The sustained decline in carbon emission intensity — measured by carbon emissions per unit of GDP and per capita carbon emissions — directly compresses the total volume of atmospheric pollutant emissions resulting from fossil fuel combustion. The reduction in the share of coal consumption within the energy structure further decreases the concentration of harmful substances such as PM2.5 and SO₂, thereby lowering the morbidity and mortality rates of pollution-related diseases, including respiratory diseases and cardiovascular and cerebrovascular diseases. Within the public health efficiency indicator framework, life expectancy at birth and the health management rate for children under seven years of age serve as core output indicators reflecting residents’ health outcomes. A reduction in the disease burden attributable to pollution will directly improve performance on these output indicators, thereby driving efficiency gains under conditions of unchanged input levels. Accordingly, improvements in emission intensity on the carbon governance side may indirectly act upon the output efficiency of the public health system through the structural optimization of health demand.

2.1.2. Rigid emission reduction constraints inducing competitive fiscal allocation between environmental expenditure and health investment.

The environmental pollution control investment intensity within the composite carbon emission governance indicator system reflects the magnitude of local government financial investment in pollution control. Under conditions of finite local fiscal capacity, the expansion of environmental expenditure driven by binding emission reduction targets will generate a competitive allocation relationship with per capita healthcare expenditure. For central and western provinces characterized by relatively low fiscal self-sufficiency, the crowding-out effect of environmental expenditure will directly compress the fiscal space available for public health service investment, thereby constraining improvements in output indicators such as bed occupancy rate and medical insurance enrollment rate. The existence of this pathway implies that the relationship between carbon governance and public health efficiency is not purely positive; the resource competition mechanism may produce stage-specific negative effects under particular fiscal conditions. This constitutes one of the core reasons why this study adopts the coupling coordination degree rather than simple correlation indicators to evaluate the relationship between the two systems.

2.1.3. Low-carbon industrial transition synergistically expanding the supply foundation and efficiency space of public health services.

The share of tertiary industry value-added in the low-carbon transition capacity indicators measures the degree to which regional industrial structure is upgrading toward a service-oriented direction. A service-sector-dominated industrial structure simultaneously reduces carbon emission intensity per unit of GDP and, by expanding the scale of modern service industries such as healthcare and health management, objectively broadens the supply foundation of public health services. This contributes to enhancing the resource density of input indicators such as the number of primary healthcare institutions per 10,000 population and promotes improvements in health service efficiency through market competition mechanisms. Concurrently, improvements in clean production levels — as represented by the increase in the comprehensive utilization rate of general industrial solid waste — also indirectly optimize the input-output structure of the public health system by reducing the “passive consumption” of medical resources attributable to industrial pollution.

2.2.1. Improvements in residents’ health levels consolidating the human capital foundation for green and low-carbon transition.

Residents’ health levels — as represented by output indicators such as life expectancy at birth and child health management rates — constitute an important dimension of regional human capital quality. Efficient public health services provide sustained and stable human capital support for low-carbon industrial transition and green technology research and development by extending healthy working years and reducing productivity losses attributable to illness. Within the carbon governance indicator system, R&D-driven green innovation capacity and the clean structural transformation of industry are both highly dependent on the continuous supply of healthy labor. Accordingly, improvements in public health efficiency create conditions for the deepening of carbon emission governance from the human capital supply side.

2.2.2. Optimization of health fiscal efficiency releasing resource space to support carbon emission governance investment capacity.

Improvements in public health service efficiency imply the achievement of higher levels of health output under existing input scales, thereby reducing medical resource waste and fiscal overexpenditure attributable to inefficiency. The increase in urban and rural residents’ medical insurance enrollment rates contributes to reducing the risk of poverty due to illness among residents, alleviating the fiscal pressure on government-funded medical assistance programs, and objectively creating additional fiscal space for local governments to direct toward environmental pollution control investment. This pathway is of particular importance for central and western provinces characterized by pronounced fiscal constraints, and represents a key mechanism for understanding how improvements in the coordination level of the two systems are mutually constitutive preconditions.

3.1.1. Entropy weight method.

As the first step of the analytical framework, this study requires the transformation of multidimensional carbon emission governance behaviors into a composite quantitative index that can be systematically compared with public health service efficiency. The comprehensive level of carbon emission governance encompasses eight indicators across four dimensions — emission intensity, low-carbon transition capacity, governance effectiveness, and carbon sink capacity. Given that these indicators differ substantially in units of measurement and their relative importance cannot be reliably determined through prior knowledge, the entropy weight method assigns objective weights based on the actual degree of variation in the indicator data, thereby effectively avoiding the bias inherent in subjective weighting approaches. This method is well-suited for the composite integration of multidimensional governance performance, and is therefore employed to calculate the comprehensive carbon emission governance level index U for each province, providing one of the core input variables for the subsequent construction of the coupling coordination degree model [32]. In this study, the entropy weight method is applied to measure the comprehensive level of carbon emission governance across Chinese provinces, where the composite index is constituted by indicators across multiple dimensions, including emission reduction intensity, technological investment, and policy effectiveness. The degree of dispersion among indicators is reflected by computing the information entropy of each indicator: the greater the degree of variation in an indicator — that is, the smaller its information entropy — the greater the amount of information it carries, and accordingly, the higher the weight assigned to it. The specific calculation procedure is as follows:

(1) Data Standardization: Given the differences in units of measurement and ranges of variation across indicators, the raw data must first be standardized to eliminate the influence of dimensional discrepancies. For positive and negative indicators, the following formulas are applied respectively. Let denote the raw value of province i on indicator j; the standardized value is computed as:

(1)

(2)

where and denote the maximum and minimum values of indicator j across all provinces, respectively. To avoid undefined logarithmic operations, a translation adjustment is applied to the standardized values: , where is a sufficiently small positive constant.

(2) Calculation of Indicator Proportions. The proportion of province i under indicator j is computed as:

(3)

(3) Calculation of Information Entropy. The information entropy of indicator j is calculated as:

(4)

where .

(4) Calculation of Indicator Weights. The objective weight of indicator j is derived from its information entropy as:

(5)

where m denotes the total number of indicators, satisfying .

3.1.2. DEA-SBM model.

Building upon the measurement of the comprehensive level of carbon emission governance, the second step of the analytical framework requires an independent assessment of public health service efficiency in order to obtain the other core input variable required for the coupling coordination degree model. The public health service system is characterized by a complex structure involving multiple inputs and multiple outputs, and it is difficult to obtain unified market price information for the monetary conversion of input and output factors. By directly incorporating input redundancy and output shortfall into the objective function, the SBM model is capable of more accurately reflecting the sources of efficiency loss within the public health service system [33], thereby providing a reliable efficiency measurement foundation for evaluating the coordinated development of carbon emission governance and public health efficiency.

(6)

where and represent the i-th input and r-th output of the K-th decision-making unit (DMU), respectively; and denote input redundancy and output shortfall, respectively; and m and q represent the numbers of input and output indicators, respectively. When = 1, the DMU lies on the production frontier and is DEA-efficient; when ρ < 1, efficiency loss is present. Based on the operational characteristics of the public health service system, an evaluation framework comprising 5 input indicators and 4 output indicators is constructed. The SBM model possesses desirable properties of units invariance and monotonicity, enabling effective efficiency evaluation of indicators with heterogeneous units of measurement, making it well-suited for the measurement requirements of this study.

3.1.3. Coupling coordination degree model.

Having obtained the comprehensive level of carbon emission governance and public health service efficiency independently, the analytical framework proceeds to its third step: integrating the two separate measures into a composite evaluation index that reflects the synergistic state of the two systems. Examining the absolute levels of either system in isolation is insufficient to determine whether the two systems are in a state of benign coordination. To address this, the present study introduces the coupling coordination degree model, which simultaneously captures the intensity of interaction between the two systems (coupling degree C) and the overall development level of both systems (composite evaluation index T), thereby transforming the complex bidirectional relationship into an objectively quantifiable and comparable dimension for evaluating coordinated development. This provides a unified analytical object for the subsequent regional disparity decomposition and spatial analysis [34].

(1) Coupling Degree

The coupling degree C reflects the intensity of interaction between the two systems, and is calculated using the following formula:

(7)

where U denotes the composite carbon emission governance level index derived from the entropy weight method, and E denotes the public health service efficiency value obtained from the DEA-SBM model; both are standardized to the interval [0, 1]. The value of C ranges from 0 to 1; a higher C value indicates stronger interaction between the two systems and closer systemic association.

(1) Comprehensive Evaluation Index. The coupling degree C alone is insufficient to distinguish whether the two systems are in a state of virtuous mutual promotion or low-level equilibrium. A comprehensive evaluation index T is therefore introduced to reflect the overall development level of the two systems:

(8)

where and represent the weight coefficients assigned to the composite carbon emission governance level and public health service efficiency, respectively. Given that this study regards both environmental governance and public health as equally important objectives for sustainable development, the weights are set as α = β = 0.5.

(2) Coupling Coordination Degree. Building upon the coupling degree C and the comprehensive evaluation index T, the coupling coordination degree D is constructed as:

(9)

The value of D ranges from 0 to 1; a higher D value indicates a higher level of synergistic development between the two systems and a greater tendency toward a virtuous coordination state. Following the classification scheme established in the existing literature [35], D values are divided into 10 grades (Table 1) to facilitate intuitive assessment of the coordinated development stage of each province during the study period.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Classification criteria for the coupling coordination degree between carbon emission governance and public health efficiency in China.

https://doi.org/10.1371/journal.pone.0350658.t001

3.2.1. Dagum Gini coefficient.

Taking the coupling coordination degree as the analytical object, the fourth step of the framework shifts toward a refined identification of the structural sources of regional disparities. The overall Gini coefficient can only measure the degree of aggregate inequality, and is incapable of addressing the policy-critical question of whether disparities primarily originate from differences between regions or from within regions [36]. By introducing the concept of hypervariable density, the Dagum Gini coefficient decomposes overall disparity into three sources — intra-regional difference, inter-regional net difference, and hypervariable density — thereby effectively resolving the problem of sub-sample cross-overlap. This approach enables a structural tracking of the sources of spatial inequality in the coupling coordination degree, and provides empirical evidence for the precise design of differentiated regional policies [37].

The Dagum Gini coefficient is defined as:

(10)

where G denotes the overall national Dagum Gini coefficient; k represents the number of regional divisions (i.e., eastern, central, and western regions, k = 3); n is the total number of provinces; and are the number of provinces in the j-th and h-th regions, respectively; and represent the coupling coordination degree of the i-th province in region j and the r-th province in region h, respectively; and denotes the national mean coupling coordination degree. According to Dagum’s theoretical framework, the overall Gini coefficient G is decomposed into three components:

(11)

(12)

(13)

where denotes the contribution of intra-regional differences (Intra-regional Difference Contribution), representing the weighted average of disparities among provinces within each region; represents the contribution of inter-regional net differences (Inter-regional Net Difference Contribution), reflecting the disparities attributable to differences in mean coupling coordination degrees across regions; and denotes the contribution of hypervariable density (Hypervariable Density Contribution), capturing the cross-regional overlap in coupling coordination degrees—that is, the disparities arising from cross-regional interaction effects.

3.2.2. Moran’s index.

The decomposition of the sources of regional disparities reveals the static distributional pattern of the coupling coordination degree. The fifth step of the framework further inquires whether there exists systematic spatial dependence in the level of coordinated development among adjacent provinces. To address this question, the present study employs the Global Moran’s Index to examine the spatial autocorrelation characteristics of the coupling coordination degree, dynamically tracking the evolutionary trends of spatial agglomeration patterns throughout the study period, and assessing whether provinces with high values and those with low values each exhibit tendencies toward geographic clustering [38,39]. This provides a basis for understanding the spatial reinforcement mechanism underlying regional differentiation, and lays the foundation for interpreting spatial spillover effects in the subsequent analysis of influencing factors.

The Global Moran’s Index is calculated as:

(14)

where I is the Global Moran’s Index; n is the total number of spatial units (30 provinces in this study); and represent the coupling coordination degree values of provinces i and j, respectively; is the mean coupling coordination degree across all provinces; and is the spatial weight between provinces i and j in the spatial weight matrix. This study adopts theGeographic Contiguity Matrix, whereby = 1 if province i and province j are geographically adjacent, and = 0 otherwise.

The Global Moran’s Index ranges from −1–1. When I > 0, the coupling coordination degree exhibits positive spatial autocorrelation, indicating that high-value regions tend to be adjacent to other high-value regions and low-value regions cluster together, reflecting spatial agglomeration. When I < 0, negative spatial autocorrelation is present, whereby high-value and low-value regions alternate in spatial distribution. When I approaches 0, the coupling coordination degree follows a random spatial distribution with no significant spatial association. By computing the Global Moran’s Index for each year from 2012 to 2023, the evolutionary trajectory of the spatial agglomeration pattern of the coupling coordination degree can be dynamically tracked.

3.4.1. Carbon emission governance indicators.

Drawing on a review of existing literature on frameworks for measuring the level of carbon emission governance, this study constructs a comprehensive indicator system for carbon emission governance from four dimensions: carbon emission intensity, low-carbon transition capacity, carbon governance effectiveness, and carbon sink capacity (see Table 2). On the emission side, carbon emissions per unit of GDP and per capita carbon emissions are selected as proxy indicators for carbon emission intensity; the former reflects the carbon efficiency of economic activities, while the latter captures the emission pressure from a demographic perspective, together characterizing the dual economic-demographic intensity of regional carbon emissions [43]. On the transition side, the share of coal consumption is used to measure the degree of clean energy structure transformation, and the share of value added by the tertiary industry is employed to represent the level of service-oriented upgrading of the industrial structure, both of which align with China’s low-carbon development strategy centered on the dual axes of energy structure adjustment and industrial transformation [44]. On the governance side, the intensity of investment in environmental pollution control and the comprehensive utilization rate of general industrial solid waste are introduced; the former measures the extent of financial commitment by local governments to pollution control, while the latter reflects the degree of cleaner production in industrial processes, and together they characterize the implementation effectiveness of carbon governance policies [45]. On the carbon sink side, forest coverage rate and the green coverage rate of built-up areas are incorporated to assess a region’s carbon absorption and sequestration capacity from the dual perspectives of natural carbon sinks and urban green spaces [46]. The indicator system systematically and objectively reflects the comprehensive level of each province across the full chain of carbon emission governance, encompassing four dimensions: emission pressure, transition pathways, governance investment, and ecological compensation.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Indicator system for carbon emission governance level.

https://doi.org/10.1371/journal.pone.0350658.t002

3.4.2. Public health service efficiency indicator system.

In accordance with the dual requirements of the DEA-SBM model for input and output indicators, this study constructs a public health service efficiency evaluation indicator system encompassing three categories of input dimensions — human resources, physical resources, and financial resources — and three categories of output dimensions — health outcomes, service efficiency, and service coverage (see Table 3). On the input side, the number of licensed (assistant) physicians per thousand population and the number of registered nurses per ten thousand population are used to characterize the supply density of human resources; the number of medical institution beds per ten thousand population and the number of primary healthcare institutions per ten thousand population are employed to measure the allocation level of physical resources; and per capita healthcare expenditure is incorporated to reflect the intensity of financial resource investment. The combination of these three categories of input indicators systematically covers the core elements of the supply side of public health services [47–49]. On the output side, life expectancy per capita is used to reflect the ultimate health outcome of healthcare services; bed utilization rate is employed to measure the actual conversion efficiency of physical resources; and the urban-rural residents’ basic medical insurance enrollment rate, neonatal visit rate, and health management rate for children under seven years of age are jointly used to characterize the breadth of population coverage and the degree of implementation of basic public health services. The inclusion of the latter three indicators is particularly consistent with China’s healthcare reform policy orientation of “strengthening primary-level institutions and emphasizing prevention,” and is capable of effectively capturing the efficiency improvements arising from the extension of public health services from the curative end to the preventive end [50–52]. Under the premise of balancing data availability and indicator representativeness, this input-output indicator system is able to comprehensively reflect the actual operational efficiency of provincial-level public health service systems, thereby providing a reliable efficiency measurement foundation for the subsequent coupling coordination analysis with the level of carbon emission governance.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Indicator system for public health service efficiency.

https://doi.org/10.1371/journal.pone.0350658.t003

3.4.3. Influencing factor indicators.

This study selects five indicators from four dimensions — economic development, social structure, demographic structure, and environmental regulation — as the influencing factors of the coupling coordination degree between carbon emission governance and public health efficiency (see Table 4). Per capita GDP, as the core proxy indicator of economic development, not only constitutes the material foundation upon which local governments advance carbon emission governance and healthcare service investment, but also directly determines the fiscal carrying capacity of both systems under conditions of resource competition [53]. The urbanization rate is introduced from the perspective of social structure; the urbanization process, on the one hand, drives adjustments in energy consumption structure and the spatial agglomeration of carbon emissions, and on the other hand, reshapes the demand scale and supply mode of public health services, making it a key social transformation variable affecting the coordinated development pattern of the two systems [54]. The proportion of the population aged 65 and above reflects the degree of aging in the regional demographic structure; the deepening of population aging intensifies the carrying pressure on the public health service system while also indirectly influencing carbon emission governance levels through its effects on labor force structure and industrial energy consumption patterns. Its inclusion in the model enables the identification of the compound influence pathways through which demographic structural change affects the coupling coordination degree [55]. R&D expenditure intensity, measured by the ratio of regional research and development expenditure to GDP, captures technological innovation capacity; the research, development, and diffusion of green and low-carbon technologies constitute the endogenous driving force behind the simultaneous improvement of carbon emission reduction and healthcare service efficiency, and this indicator is capable of effectively capturing the promotional effect of technological progress on the coordinated development of the two systems [56]. Urban population density, represented by the number of urban residents per unit of administrative area, characterizes the degree of spatial agglomeration; the spatial distribution pattern of the population both directly affects the supply radius and resource utilization efficiency of public health services, and influences carbon governance performance through changes in energy consumption intensity and transportation emission levels. Its incorporation into the spatial structure dimension facilitates the identification of structural constraints on regional coordinated development from the perspective of spatial carrying capacity.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Indicator system for influencing factors.

https://doi.org/10.1371/journal.pone.0350658.t004

4.6.1. Level of economic development.

The effect of per capita GDP on the coupling coordination degree exhibits significant regional heterogeneity (see Table 8). At the national level, the coefficient is 0.019, which fails to pass the significance test (p > 0.1). Disaggregated by region, the coefficient for the eastern region is 0.076 (p < 0.01), indicating a significant positive marginal effect of economic development level on the coupling coordination degree in the east; the coefficient for the central region is −0.079 (p < 0.01), representing a significant negative effect; and the coefficient for the western region is −0.012, which does not reach statistical significance. It is thus evident that the effect of economic development level on the coupling coordination degree is inconsistent across regions, with significant inter-regional differences.

Download:

• PNG
  larger image
• TIFF
  original image

Table 8. Results of the two-way fixed-effects Tobit model for influencing factors.

https://doi.org/10.1371/journal.pone.0350658.t008

4.6.2. Social structure.

The coefficient of urbanization rate at the national level is −0.306, which is significantly negative at the 10% significance level; however, the standard error reaches 0.180, indicating limited estimation precision. The region-specific coefficients are −0.000 for the eastern region, 0.063 for the central region, and −0.311 for the western region, none of which pass the significance test. This indicates that the estimation of this variable in the overall sample is subject to considerable uncertainty. Therefore, within the present analytical framework, no statistically significant evidence is found for a direct effect of urbanization rate on the coupling coordination degree, and its underlying mechanism warrants further investigation.

4.6.3. Demographic structure.

The coefficient of the aging population proportion (std_x3) at the national level is 0.019, which does not reach statistical significance. Disaggregated by region, the coefficients for the eastern region (−0.017) and the central region (−0.024) are both statistically insignificant, while the coefficient for the western region is 0.0481, which is significantly positive at the 10% significance level. Only in the western region does this variable exhibit a statistically significant positive association with the coupling coordination degree, with no significant effects observed in the remaining regions. This indicates that the effect of aging on the coupling coordination degree is not linearly consistent, but rather exhibits directional divergence due to differences in regional resource conditions and system carrying capacity. The significant positive effect in the western region stands in sharp contrast to the non-significant negative tendency observed in the eastern and central regions, and the specific underlying mechanism warrants further analysis.

4.6.4. Environmental regulation.

The coefficient of R&D expenditure intensity (x4) at the national level is 0.045 (p < 0.05), indicating that technological innovation investment exerts a significant positive promotional effect on the coupling coordination degree. Disaggregated by region, the coefficient is largest for the central region (0.185, p < 0.01), followed by the western region (0.110, p < 0.1), while the coefficient for the eastern region is −0.024, which does not reach statistical significance. The marginal effect of R&D investment on the coupling coordination degree exhibits a gradient characteristic of increasing magnitude from east to west, reflecting the diminishing marginal returns to technological innovation under the already high baseline governance level of the eastern region, as well as the reality that technological investment in the central and western regions remains at a critical stage of gap-filling.

4.6.5. Population density.

The coefficient of the logarithm of urban population density at the national level is −0.080 (p < 0.01), representing a significant negative effect. Disaggregated by region, the coefficients for the eastern region (−0.036) and the central region (−0.026) both fail to reach statistical significance; the coefficient for the western region is −0.139 (p < 0.01), exhibiting the most pronounced negative effect. This indicates that spatial population agglomeration in the western region has failed to generate agglomeration economic benefits; instead, the expansion of the public service supply radius and the rising per capita infrastructure costs have formed a significant inhibitory effect on the overall level of coordination, while this effect has yet to reach statistical significance in the eastern and central regions.

5.1.1. Sustained and gradual improvement in coordination level, with policy transition pains and institutional time lags constraining the deepening of synergistic progress.

During the period from 2012 to 2023, the coupling coordination degree between carbon emission governance and public health efficiency in China exhibited an overall evolutionary pattern of initial fluctuation followed by a sustained upward trend, with the coordination degree of most provinces rising from the range of 0.44–0.63 to the range of 0.54–0.72. The consistently directional improvement over 12 years indicates that the coordinated advancement of the “dual carbon” targets and the Healthy China strategy has preliminarily driven a virtuous interaction between the two systems. Nevertheless, the coordination degree of most provinces has yet to break through the threshold of intermediate coordination, the depth of inter-system synergy remains limited, and the pattern of interprovincial differentiation has not undergone fundamental restructuring. Examining the temporal trajectory, the coordination degree trough in 2016 — during which the mean public health efficiency value dropped sharply from 0.834 to 0.769, with the central region declining to 0.712 — concentrated the manifestation of three structural tensions. Drawing on existing theoretical frameworks, the literature, and plausible contributing factors [57,58], the study identifies the following three potential mechanisms: first, the input-output time lag arising from the intensive implementation of healthcare reform policies such as hierarchical diagnosis and treatment systems and family physician contracting; second, the competitive allocation between environmental protection and healthcare expenditure triggered by the transmission of the rigid carbon emission constraints of the 13th Five-Year Plan to local fiscal systems; and simultaneously, the paradoxical pattern of “increasing bed numbers accompanied by declining efficiency” formed by the concurrent outflow of medical personnel and persistently low facility utilization rates in the central region. The overall recovery in coordination degree after 2018 is closely associated with the policy synergy generated by the deepening of ecological civilization institutional reform and the comprehensive launch of the Healthy China strategy, indicating that the degree of integration of cross-sectoral policy frameworks constitutes a key institutional variable driving the sustained improvement of provincial coordination levels.

5.1.2. Inter-regional disparities dominate overall divergence, with factor endowment differences and fiscal system logic jointly entrenching the east-west divide.

Inter-regional disparities have consistently constituted the dominant source of overall disparities in the coupling coordination degree, with structural imbalances characterized by “deepening east-west divide and increasing intra-western dispersion” intensifying. The decomposition results of the Dagum Gini coefficient indicate that the contribution rate of inter-regional disparities remained stably within the range of 39.74%–42.26% throughout the observation period, and after 2018, the Gini coefficient between the eastern and western regions expanded continuously from 0.090 to 0.114, while the intra-western Gini coefficient surged from 0.110 to the elevated range of 0.114–0.123. This pattern is consistent with the inherent differences in factor endowments between the eastern and western regions: the comprehensive level of carbon emission governance in the eastern region persistently remained at a high level throughout the observation period and continued to improve, while the western region has long lingered in the low-value range, with the mean gap between the two systems providing direct empirical support for the continuous expansion of the inter-regional Gini coefficient. From the perspective of the fiscal system, the fiscal self-sufficiency rate of western provinces is relatively low, and the fragmented allocation of central transfer payments may be insufficient to effectively drive systematic improvements in public health service efficiency; however, given the limited availability of direct fiscal data, the precise estimation of the magnitude of this mechanism awaits further deepening in subsequent research. The intensification of carbon emission governance policies since 2018 has also produced markedly asymmetric shocks to the industrial structures of different regions — the eastern region completed industrial green transformation first, advancing carbon emission reduction and economic benefits in tandem; the central and western regions remain highly dependent on heavy industry and resource-based industries, experiencing more severe structural adjustment pains under emission reduction constraints, which indirectly weakens the resource allocation capacity of local governments in the healthcare sector and further entrenches the coordination degree gap between the eastern and western regions.

5.1.3. Sustained strengthening of positive spatial autocorrelation, with factor mobility mechanisms and institutional connectivity effects jointly driving agglomeration polarization.

The statistical characteristics of continuously strengthening spatial agglomeration are pronounced, with the Global Moran’s Index rising from 0.227 in 2012 to 0.497 in 2022, representing an increase of 118.9%. The local spatial autocorrelation results indicate that the high-high agglomeration zones expanded from 4 provinces in 2012 to 14 in 2023, the low-low agglomeration zones likewise expanded from 9 to 11 provinces, and transition-type provinces contracted substantially, with the bipolarization pattern clearly manifested at the data level. The continuous expansion of high-value agglomeration zones demonstrates a degree of temporal alignment with the advancement of cross-administrative-region collaborative policies in areas such as the Yangtze River Delta and Pearl River Delta, while the low-value locked zones have long been concentrated in the northwestern inland provinces, corresponding to the persistently low performance of these provinces in terms of carbon emission governance level and public health efficiency. The two stage-specific low values of the Moran’s Index in 2013 and 2017 can be interpreted in conjunction with major contemporaneous policy milestones: in 2013, the Third Plenary Session of the 18th Central Committee of the Communist Party of China deliberated and adopted the Decision on Several Major Issues Concerning Comprehensively Deepening Reform, incorporating ecological civilization construction into the “Five-in-One” overall layout for the first time; during the policy transition period in which the implementation details awaited clarification at the provincial level, the interprovincial coordination gap experienced a brief narrowing. In 2017, the comprehensive implementation of the 13th Five-Year Plan for Poverty Alleviation directed large-scale targeted fiscal resources toward underdeveloped county-level areas, which in the short term raised the coordination levels of low-value provinces, compressed the spatial gap among provinces, and produced a stage-specific homogenization effect. The foregoing patterns indicate that well-documented targeted policy interventions possess realistic potential to temporarily disrupt spatial polarization in the short term; however, the homogenization effect is difficult to sustain, and once the policy dividend period concludes, the market logic of factor mobility and the structural constraints of endowment differences will drive spatial agglomeration back onto a strengthening trajectory, underscoring the imperative of designing institutional frameworks for long-term regional coordination policies.

5.1.4. Significant regional heterogeneity in associated factors, with divergent economic development pathways and differential marginal returns to technological innovation jointly shaping the coordination pattern.

The region-specific results of the two-way fixed effects Tobit regression reveal that the direction and magnitude of the associations between various factors and the coupling coordination degree differ significantly across the eastern, central, and western regions, reflecting the fact that the associative logic between these factors and coordinated development is not uniform across different stages of development and resource endowment conditions. The level of economic development (per capita GDP) exhibits a significant positive association with the coupling coordination degree in the eastern region (β = 0.076, p < 0.01), yet a significant negative association in the central region (β = −0.079, p < 0.01), while the coefficient for the western region is −0.012 and statistically insignificant. Given that the aforementioned associations represent correlations rather than causal relationships and may partially reflect the confounding influence of omitted variables, an alternative explanation nonetheless warrants attention: provinces in the central region with higher per capita GDP tend also to be those with a more concentrated heavy industrial base and higher costs of industrial structural transformation, and the structural contradiction between industrial carbon emission intensity and healthcare resource allocation may jointly suppress the coupling coordination degree of these provinces, thereby producing a statistically negative association between economic development level and coordination degree. Furthermore, provincial-level omitted variables such as biases in fiscal expenditure structure and differences in the stringency of environmental regulation enforcement may also confound this association. R&D expenditure intensity is significantly positively associated with the coupling coordination degree at the national level (β = 0.045, p < 0.05); disaggregated by region, the association is most pronounced in the central region (β = 0.185, p < 0.01), followed by the western region (β = 0.110, p < 0.1), with no significant association observed in the eastern region, exhibiting a gradient characteristic of increasing magnitude from east to west. This indicates that a strong positive association persists between technological innovation investment and improvements in coordination degree in the central and western regions, making it one of the factors most closely associated with the coordination level of the two systems in the central and western regions at the present stage. Urban population density exhibits a significant negative association with the coupling coordination degree both nationally (β = −0.080, p < 0.01) and in the western region (β = −0.139, p < 0.01), with no significant associations observed in the eastern and central regions; this association is consistent with the relatively weak public service supply capacity in the western region. The urbanization rate and the aging population proportion fail to reach statistical significance in most regions, and their net associations with the coupling coordination degree remain indeterminate within the present research framework. The foregoing results collectively indicate that economic development level and technological innovation investment are the factors most prominently associated with coordinated development across regions, and that the magnitude of these associations varies significantly with regional development conditions, such that differentiated factor allocation strategies carry important reference value for enhancing regional coordination levels.

5.2.1. Establishing cross-sectoral policy sequencing and coordination mechanisms to alleviate resource misallocation induced by institutional time lags.

Strengthening cross-sectoral policy sequencing coordination and fiscal synergy management should be prioritized to break through the resource misallocation caused by institutional time lags and policy conflicts. The study finds that the coupling coordination degree exhibited a pronounced trough during periods of intensive concurrent implementation of carbon emission constraint policies and healthcare reform policies, most notably in the central region, and that the coordination degree promptly recovered following an increase in the degree of policy framework integration, indicating that temporal conflicts among cross-sectoral policies constitute an important associated factor constraining the coordinated advancement of the two systems. It is recommended that a linkage-based budget protection mechanism for environmental and healthcare expenditure be established at the provincial level, with a clearly defined rigid floor for public health expenditure, so as to prevent the transmission of emission reduction constraints to local fiscal systems from generating a crowding-out effect on healthcare investment. Incorporating reductions in carbon emission intensity and improvements in public health service efficiency into a unified official performance evaluation system would employ institutional constraints to incentivize local governments to proactively resolve the resource competition between the two systems. With respect to the output time lag of healthcare reform policies, it is recommended that efficiency evaluation mechanisms be embedded in newly added primary-level healthcare investment projects, with regular tracking of input-output changes and timely initiation of resource adjustment procedures for projects exhibiting pronounced time lags, so as to prevent sustained investment from depressing overall efficiency levels. From the perspective of institutional design, the coordinated planning of environmental governance and public health policies should be incorporated into top-level arrangements, replacing ex post remedial resource redistribution with ex ante institutional coordination, so as to fundamentally compress the interference space of institutional time lags on the coordination process.

5.2.2. Constructing a differentiated regional coordination policy framework to unblock institutional bottlenecks in the east-west divide.

Region-specific and category-specific policy implementation grounded in regional heterogeneity characteristics should be pursued, leveraging differentiated interventions to narrow the divergence in coordinated development across the three major regions. Inter-regional disparities have consistently been the dominant source of overall disparities, with the east-west divide continuing to widen after 2018 and intra-western differentiation simultaneously intensifying. The heterogeneous effects of influencing factors likewise indicate that the core constraints on coordinated development differ fundamentally across regions, making it imperative to construct differentiated precision intervention frameworks based on the core constraints of each region, replacing large-scale resource accumulation with structural policy reorientation. The policy focus of the eastern region should shift from incremental investment toward the optimization of existing stock efficiency, exploring market-oriented collaborative pathways between carbon governance and public health services to establish institutional models replicable in the central and western regions. The central region should prioritize the promotion of green structural adjustment in its industrial composition, establishing a linkage mechanism between emission reduction targets and improvements in healthcare efficiency, preventing singular emission reduction constraints from compressing local healthcare fiscal space, and gradually resolving the structural contradiction between economic expansion and improvements in coordination degree. Given the pronounced intra-regional differentiation problem in the western region, it is recommended that transfer payment funds be concentrated toward county-level units with weak coordination capacity, while simultaneously promoting institutional transplantation and technology transfer from the eastern to the western region, replacing mere financial transfers with capacity building so as to fundamentally enhance the endogenous coordination capacity of western provinces.

5.2.3. Activating regional spatial spillover effects and interrupting the self-reinforcing cycle of low-value agglomeration.

A regional collaborative policy system oriented toward the dual objectives of activating high-value spillovers and blocking low-value lock-in should be constructed, replacing isolated provincial-level policy efforts with collective action mechanisms. Spatial agglomeration continued to intensify throughout the observation period, with high-value and low-value agglomeration zones expanding simultaneously, transition-type provinces contracting substantially, and the bipolarization pattern trending toward entrenchment; moreover, the two targeted policy intervention nodes demonstrated a high degree of temporal alignment with the stage-specific low values of the Moran’s Index, indicating that targeted policy investment possesses the potential to temporarily disrupt spatial polarization in the short term, but the homogenization effect is difficult to sustain once the policy dividend period concludes, and spatial agglomeration promptly returns to a strengthening trajectory. With regard to activating high-value spillovers, high-coordination provinces in the Yangtze River Delta, Pearl River Delta, and similar regions should be encouraged to deepen cross-provincial coordination mechanisms, engaging in substantive cooperation on issues such as green technology sharing, the construction of regional environmental health data platforms, and cross-provincial public health collaborative governance, so as to convert the advantages of institutional connectivity into systemic public goods capable of radiating to surrounding areas and orderly expanding the spatial coverage of high-value agglomeration zones. With regard to blocking low-value reinforcement, precision-targeted assistance should be implemented for northwestern inland provinces that have long remained in low-low agglomeration zones, encouraging high-coordination provinces to continuously export healthcare service management experience and green technology support, replacing traditional aid-construction models with trustee-based cooperative arrangements so as to sustainably enhance the local operational capacity of recipient areas. Institutional capacity building and long-term technical support mechanisms should replace stage-specific resource injections, gradually improving the endogenous coordination capacity of low-value agglomeration zones and interrupting the self-reinforcing cycle of low-value agglomeration at the institutional level.

5.2.4. Implementing region-specific precision factor allocation and activating the maximum associative benefits of relevant factors through differentiated policy instruments.

Differentiated factor allocation priorities should be formulated, replacing nationally uniform deployment with precision policy instruments. The empirical results indicate that the direction of association between economic development level and coordination degree differs fundamentally between the eastern and central regions, that the positive association of R&D investment intensity increases progressively from east to west, and that urban population density exerts a significant inhibitory effect on the coordination degree of the western region; the core associated factors for coordinated development are not uniform across regions, making differentiated policy implementation a necessary prerequisite for enhancing policy effectiveness. At the level of policy implementation, given that the economic development level of the eastern region has already demonstrated a stable positive association with the coordination degree, the policy focus should shift toward the sustained optimization of economic growth quality, promoting the incorporation of healthcare institution energy consumption management into the carbon trading accounting framework, and using market-based incentives to guide the green and intensive utilization of healthcare resources, thereby forming a virtuous pattern in which economic growth and improvements in the coordination of the two systems advance in tandem. In the central region, the negative association between economic growth and coordination degree reflects the fact that economic expansion at the present stage has yet to effectively drive the simultaneous improvement in the efficiency of both systems; the priority should therefore be placed on promoting the clean transformation of the industrial structure, establishing a provincial industrial guidance fund doubly linked to emission reduction targets and improvements in healthcare efficiency, and giving priority credit and fiscal support to industrial projects meeting green standards, so as to gradually resolve the structural contradiction between economic growth and improvements in coordination degree. In the western region, where the positive association between R&D investment intensity and coordination degree is most prominent, technological innovation investment should be treated as the priority policy lever for driving improvements in coordination degree; national key science and technology projects should prioritize allocating grants to western research institutions, and eastern enterprises should be encouraged to establish green technology joint laboratories in the western region through market-based approaches. Simultaneously, in response to the significant inhibitory effect of urban population density on the coordination degree of the western region, the maintenance of public health service coverage rates should be established as a binding prerequisite for advancing urbanization, with the simultaneous promotion of grid-based layout of rural healthcare services and the construction of digital telemedicine access, so as to fundamentally alleviate the structural inhibition of population agglomeration on the coordination level of the western region.