Non-monotonic effects and spatial spillovers of urban green space on chronic infectious disease dynamics in China

Abstract

As global public health faces unprecedented challenges, Urban Green Space (UGS) is increasingly recognized as a vital instrument for health intervention. However, its efficacy in mitigating infectious diseases remains conceptually contested. Drawing on longitudinal data from 300 Chinese prefecture-level cities (2003–2023), this study employs Spatial Durbin Models (SDM) and threshold regression to unravel the complex spatio-temporal trade-offs between UGS and chronic infectious diseases. Focusing on the mitigation of chronic infectious diseases, our findings challenge linear paradigms by unveiling a non-monotonic “risk-to-resilience” transition. In the nascent stages of greening, UGS expansion may paradoxically create a “pathogenic window” elevating transmission risk through intensified human-wildlife-environment interactions. However, once UGS coverage surpasses a critical tipping point, it serves as a potent ecological shield, significantly curbing the spread and latency of chronic infectious diseases. This suppression is achieved through microclimate stabilization, bio-filtration, and the enhancement of community salutogenesis. Furthermore, our analysis reveals robust positive spatial externalities, demonstrating that protection against chronic infections propagates across administrative borders via ecological corridors. This study identifies a pivotal scale effect in green governance, positioning UGS as a long-term strategic asset for demographic health. Our results advocate for a transition from localized planting to trans-regional collaborative green infrastructure networks, providing crucial theoretical scaffolding for the Healthy China 2030 initiative and the management of chronic infectious risks.

Introduction

Chronic infectious diseases, encompassing long-term and recurrent infections such as tuberculosis, hepatitis, and HIV/AIDS, remain among the most persistent global public health challenges, exerting profound social, economic, and health consequences (Dye, 2014)¹⁶. Projections by the World Health Organization suggest that their global burden will remain substantial, with approximately 10 million new tuberculosis cases annually and hundreds of millions of individuals living with chronic viral infections projected through 2050 (WHO, 2023)⁴⁹. To address these challenges, governments worldwide have pursued a variety of strategies, including the expansion of vaccination programs, the strengthening of health systems, and the promotion of behavioral interventions (Fagan et al., 2019)¹⁹. However, many developing countries continue to face the dual pressure of controlling chronic infectious diseases while undergoing rapid urbanization and socioeconomic transition (Reyes et al., 2012)³⁷.

China, the world’s most populous developing nation, exemplifies one of the most complex public health transitions. In 2021, over 800,000 new tuberculosis cases were reported, alongside more than 80 million individuals living with chronic hepatitis B or C, making China one of the countries with the highest disease burdens (Alara et al., 2024)¹. Compared with other countries at similar development stages, China’s trajectory is distinguished by both the magnitude of prevalence and the coexistence of entrenched and emerging risks (Chen et al., 2022)⁹. This burden exhibits three defining features: an exceptionally large number of patients, pronounced regional disparities, and a persistent risk of resurgence despite ongoing economic modernization (Snowden, 2008)⁴².

Unlike many high-income countries that have achieved sustained control over chronic infectious diseases through decades of investment in healthcare infrastructure, China continues to confront the dual challenge of ongoing disease transmission and rapid social transformation. This divergence underscores the complexity of China’s health transition, as the country must address both entrenched infectious diseases and the rising prevalence of non-communicable conditions. As a result, chronic infectious diseases represent not only a domestic public health concern but also an issue of global significance with implications for international health governance.

Following the 2003 SARS epidemic, China implemented major reforms to its public health system, including the establishment of nationwide disease surveillance networks and the enactment of legal frameworks for epidemic prevention and control (Wang et al., 2007)⁴⁶. Despite these advances, the prevalence of chronic infectious diseases has persisted in certain regions, reflecting structural disparities in health infrastructure and widening urban–rural inequalities (Connolly et al., 2021)¹². By 2023, the continued high incidence of tuberculosis and widespread hepatitis infections indicated that chronic infectious diseases remain a critical barrier to achieving the goals of Healthy China 2030. These trends suggest the emergence of a new public health norm in which acute epidemic outbreaks have declined, but chronic infectious conditions endure, thereby reshaping China’s health landscape. The corresponding trends are presented in Fig. 1.

Fig. 1.

The development trend of chronic infectious diseases in China.

In response to these persistent challenges, China has progressively refined its public health policies. The nationwide expansion of immunization programs, health education campaigns, and infectious disease management systems illustrates a strategic shift from short-term epidemic response to long-term disease control. Nonetheless, the continued prevalence of chronic infections underscores China’s unique predicament of pursuing “disease control before full development,” as it enters a new health transition stage with per capita GDP levels still below those of advanced economies (Cacciatore et al., 2025)⁷. This situation places substantial pressure on healthcare financing, disease management, and community-level service delivery. How to optimize policy responses to persistent health risks while balancing economic growth with sustainable health security has therefore emerged as a critical research question (Bali et al., 2017)³.

Urban green space (UGS) plays a pivotal yet underexplored role in shaping the transmission dynamics of chronic infectious diseases. Poorly designed or inadequately maintained green areas may increase exposure risks by intensifying population aggregation and human–pathogen interactions (Tsai et al., 2010)⁴⁰. Conversely, the expansion and effective management of UGS can suppress disease spread by improving environmental quality, regulating microclimates, and fostering physical activity and psychological restoration. These functions enhance immune resilience and strengthen community health capacity (Zywert, 2024)⁵³. Such duality suggests that while UGS may initially elevate transmission risks, once a critical coverage threshold is reached, it exerts significant protective effects against chronic infectious diseases.

UGS including parks, recreational facilities, and vegetated landscapes extends beyond ecological and aesthetic functions to influence health outcomes through patterns of human mobility, social interaction, and environmental exposure (Edeigba et al., 2024)¹⁸. Its accessibility and amenities attract residents, simultaneously creating opportunities for both increased disease contact and enhanced resilience. By reducing stress, promoting healthier lifestyles, and reinforcing social cohesion, UGS can mitigate the long-term spread of chronic infectious diseases (Pinter-Wollman et al., 2018)³⁵. Reflecting this recognition, Chinese urban planning frameworks increasingly emphasize UGS not only as an ecological asset but also as a public health resource (Yang et al., 2022)⁵². Yet empirical evidence remains limited regarding how UGS interacts with chronic infectious diseases, particularly concerning its spatial heterogeneity and threshold effects.

Despite increasing recognition of these dynamics, critical knowledge gaps remain. First, most existing studies overlook the spatial heterogeneity of UGS and its region-specific impacts under varying levels of disease prevalence. Second, the potential nonlinear or threshold effects of UGS where protective functions emerge only after green coverage exceeds critical levels have not been systematically examined. Addressing these mechanisms is essential for designing healthier urban environments and strengthening epidemic preparedness.

To fill these gaps, this study explores the relationship between UGS and chronic infectious diseases in China using panel data from 300 cities spanning 2003–2023. Employing spatial Durbin and threshold regression models, we assess both the spatial spillover effects of UGS and the nonlinearities in its impact on disease prevalence. This integrated approach enables the linking of ecological features with public health dynamics, providing novel insights into how UGS shapes infectious disease control at the national scale.

The contributions of this research are threefold. First, it expands the public health literature by identifying UGS as a key environmental determinant in the persistence and mitigation of chronic infectious diseases. Second, it offers empirical evidence on the spatial and nonlinear mechanisms through which UGS influences health outcomes. Third, by focusing on China—where rapid UGS expansion intersects with pronounced regional disparities in disease burden—this study generates lessons of broader relevance for other countries undergoing similar public health transitions. Collectively, the findings provide a scientific basis for optimizing UGS provision and advancing the development of healthier, more resilient cities.

Literature review

Urban green space (UGS), encompassing parks, forests, wetlands, greenbelts, and other vegetation-covered areas within metropolitan environments, has increasingly been recognized as a critical determinant of public health, particularly in relation to chronic infectious diseases (Flies et al., 2017)²⁰. With the burden of infectious diseases persisting across both developing and developed countries and China facing particularly complex challenges due to rapid urbanization, environmental change, and high population mobility understanding the mechanisms through which UGS influences disease dynamics has emerged as an important research frontier. UGS can shape patterns of pathogen exposure, transmission risks, and long-term health resilience, highlighting its dual role as both a potential amplifier and suppressor of chronic infectious disease spread.

One of the central functions of UGS lies in its ability to regulate ecological conditions and reshape human–environment interactions, both of which are closely linked to infectious disease prevalence. Vegetation reduces air pollutants through deposition and absorption, modifies humidity levels, and stabilizes microclimates, thereby influencing pathogen survival and transmission pathways (Buraerah et al., 2023)⁵. Empirical evidence indicates that well-planned urban greening improves air quality, reduces the persistence of airborne pathogens, and enhances immune responses within local populations (Li et al., 2021)²⁷. Moreover, UGS mitigates the urban heat island effect by moderating temperature and humidity, two critical determinants of vector activity and microbial viability, thereby lowering the risks of climate-sensitive infectious diseases (Yadav et al., 2023)⁵¹. Remote sensing–based spatial regression studies further confirm that higher vegetation coverage reduces exposure to environmental conditions conducive to pathogen persistence and vector proliferation across diverse urban settings worldwide.

Taken together, these findings suggest that UGS functions as an ecological buffer, moderating environmental exposures and transmission dynamics, and thereby providing a crucial foundation for the prevention and management of chronic infectious diseases.

The contribution of UGS extends beyond environmental regulation to the promotion of physical activity, which is indispensable for lowering the risk of chronic infectious diseases. Public green areas provide accessible venues for low- to moderate-intensity exercise, such as walking, joggingand recreational activities that improve cardiovascular endurance, pulmonary function, and immune system resilience (Teixeira et al., 2023)⁴³. Epidemiological evidence further indicates that regular exposure to greenery is associated with reduced incidence of infectious respiratory conditions, lower prevalence of comorbid chronic diseases, and improvements in immunological markers central to infection resistance (Shao et al., 2021)³⁹. In parallel, urban vegetation mitigates exposure to PM2.5 and PM10, alleviating respiratory irritation and thereby decreasing susceptibility to bacterial and viral infections (Loaiza-Ceballos et al., 2022)²⁹. Collectively, these studies position UGS not only as a recreational amenity but also as a critical determinant of infectious disease prevention and control in urban environments.

Psychological and social dimensions provide additional mechanisms through which UGS reduces chronic infectious disease risks. Interaction with natural environments alleviates stress, regulates cortisol levels, and promotes mental well-being (Shuda et al., 2020)⁴⁰. These benefits are particularly relevant as chronic stress and depression weaken immune defenses, heightening vulnerability to infection. Moreover, green buffers reduce noise pollution, which indirectly improves sleep quality and neurological health, both closely linked to immune competence (Dzhambov et al., 2014)¹⁷. Parks and community green areas also function as inclusive spaces for social interaction, health promotion campaigns, and recreational gatherings, supporting awareness of hygiene, vaccination, and preventive behaviors (Han et al., 2022)²². During the COVID-19 pandemic, when conventional social networks were disrupted, UGS played a distinctive role in sustaining mental resilience, facilitating safe outdoor interactions, and mitigating the psychological burden of isolation, thereby indirectly contributing to infection control (Wang et al., 2024)⁴⁷. These findings highlight the intertwined psychological and social pathways through which UGS enhances immune and behavioral resilience in urban populations.

Despite these benefits, access to UGS remains unevenly distributed. Affluent neighborhoods typically enjoy well-maintained, easily accessible green amenities, whereas marginalized or peripheral districts remain underserved (Kifayatullah et al., 2025)²⁵. Such inequities disproportionately affect vulnerable groups—including older adults, migrants, and low-income households—who face mobility restrictions, limited healthcare resources, and financial constraints that prevent them from fully benefiting from UGS (Koehler et al., 2023)²⁶. Consequently, unequal access to green infrastructure risks exacerbating health disparities and reinforcing spatial inequalities in infection exposure and outcomes. Addressing these inequities requires not only quantitative expansion but also equitable spatial planning, particularly in high-density neighborhoods, disadvantaged districts, and communities with elevated infectious disease burdens. Only through such targeted strategies can UGS function as an effective and inclusive instrument for public health improvement.

The importance of UGS is magnified by accelerating climate change. Populations affected by chronic infectious diseases are especially vulnerable to extreme events such as heat waves, flooding, and heavy rainfall, owing to compromised immunity, pre-existing comorbidities, and limited adaptive capacity (Putsoane et al., 2024)³⁶. Climate variability not only reshapes ecological niches of bacteria, fungi, and viruses but also amplifies transmission through contaminated water, inadequate sanitation, and overcrowded urban conditions. UGS offers critical adaptive functions by moderating thermal stress, improving water infiltration, reducing stormwater runoff, and stabilizing microclimates, thereby mitigating conditions favorable to pathogen persistence and vector proliferation (Pereira et al., 2024)³⁴. For instance, tree canopy expansion and wetland preservation are linked to reduced vector-borne risks, while shaded areas alleviate physiological strain that undermines immune defenses during heat extremes. Thus, UGS should be viewed not merely as a health-enhancing amenity but as a climate adaptation strategy that reduces infectious disease vulnerability under climate stress.

Despite these advances, knowledge gaps remain. Research is often limited to city-specific case studies, restricting generalizability across China’s heterogeneous urban contexts. Methodological constraints, including cross-sectional designs and small-scale surveys, hinder causal inference and dynamic assessment of UGS impacts on disease prevalence. Moreover, insufficient attention has been given to how physical, psychological, and ecological pathways intersect with transmission dynamics, or how benefits vary by accessibility, socioeconomic stratification, and exposure to climate extremes. Marginalized populations often face dual vulnerabilities limited UGS access and heightened climate-related infection risks yet this intersection remains underexplored. These gaps underscore the need for longitudinal data and advanced econometric techniques, such as spatial Durbin and threshold regression models, to capture spillover effects, nonlinearities, and critical thresholds in the UGS–disease relationship.

Despite substantial progress, important gaps remain in the literature. Existing studies are often confined to case-specific analyses, raising concerns about external validity and the broader applicability of findings to urban health contexts. Methodologically, much of the current research continues to rely on cross-sectional or survey-based data, limiting the ability to establish robust causal relationships, evaluate long-term disease dynamics, or account for lagged effects between environmental exposures and health outcomes. Moreover, although the physical, psychological, and ecological benefits of urban green space (UGS) are well documented, limited attention has been given to how these benefits interact with chronic infectious disease transmission. Specifically, little is known about how spatial accessibility, socioeconomic stratification, or differential exposure to climate extremes mediate the association between UGS and infection prevalence. Addressing these limitations requires the use of longitudinal datasets and advanced econometric approaches. Spatial Durbin models can capture interregional spillover effects, while threshold regression frameworks allow for the identification of nonlinear and stage-dependent dynamics, providing more systematic insights into how UGS influences chronic infectious disease risks across time and space.

As illustrated in Fig. 2, UGS shapes the prevalence of chronic infectious diseases through five primary pathways: environmental regulation, physical health promotion, psychological resilience, social cohesion, and climate adaptation. These mechanisms, driven by both natural and social processes, interact with pathogen transmission in complex ways. Vegetation improves air and water quality by reducing pollutants, thereby lowering risks of respiratory and waterborne infections. Accessible green areas encourage physical activity and strengthen immune defenses. Exposure to natural environments alleviates stress, enhances mental health, and indirectly reinforces immunological functioning. Community parks facilitate social interaction and the dissemination of health knowledge, improving collective prevention capacity. In addition, green infrastructure moderates climate-related factors, that otherwise enhance pathogen survival and transmission. Collectively, these pathways highlight the multifaceted role of UGS in strengthening urban health resilience and provide a theoretical foundation for formulating empirical hypotheses on its spatial spillover and threshold effects in China’s context of chronic infectious diseases.

Fig. 2.

The logic behind urban green spaces’ impact on chronic infectious diseases.

Urban green space plays a crucial role in promoting resilience against chronic infectious diseases. By providing accessible environments for outdoor activities, green spaces help reduce sedentary lifestyles, strengthen immune function, and improve cardiovascular health, all of which indirectly lower susceptibility to long-term infectious conditions. In addition, green vegetation enhances local air quality by filtering pollutants and reducing concentrations of harmful particles such as PM2.5 and PM10. This directly benefits respiratory health, which is particularly important since compromised respiratory systems increase vulnerability to chronic infections. Collectively, these mechanisms underscore the essential role of UGS in safeguarding urban populations from the cumulative health risks that contribute to the persistence and prevalence of chronic infectious diseases (Bakker, 2024)².

However, the expansion of urban green space may initially coincide with heightened infection risks, a phenomenon driven by intensified population aggregation and exposure pathways. In the early stages of green development, green resources are scarce and fragmented. Under these conditions, any marginal increase in green space tends to function as a highly attractive social interaction node rather than an ecological barrier. These newly created but limited spaces draw residents from surrounding high-density neighborhoods for recreation, paradoxically transforming dispersed populations into aggregated crowds. This concentration of social activities significantly elevates the contact rate and facilitates the transmission of respiratory pathogens within confined areas. Therefore, before the green network reaches a scale sufficient to disperse users, the increased exposure probability driven by social aggregation outweighs the nascent ecological benefits.

Hypothesis 1

In the short term, increases in urban green space are associated with higher prevalence of chronic infectious diseases due to intensified population aggregation and exposure pathways.

Beyond physical health, UGS also contributes substantially to the psychological and social dynamics that influence the spread and persistence of chronic infectious diseases. Access to green environments reduces stress, alleviates depressive symptoms, and strengthens immune responses, thereby lowering individual vulnerability to infections over the long term. In addition, green spaces encourage social interaction and community participation, which can have dual consequences for public health: while they help build stronger social networks that enhance psychological resilience and collective health awareness, they may also increase opportunities for interpersonal contact, potentially facilitating the transmission of certain infectious diseases in densely populated settings. These findings suggest that UGS serves as a critical public health resource not only by improving mental and social well-being but also by shaping the social pathways through which chronic infectious diseases may spread (Olszewska-Guizzo et al., 2021)³².

At the same time, the benefits and risks associated with UGS are not confined within administrative boundaries. Improvements in environmental quality, recreational opportunities, and health resources in one city can attract residents from surrounding areas, altering patterns of population mobility and disease ecology. Such interregional interactions may expand the geographic scope of chronic infectious diseases, as pathogens and vulnerable populations move across local boundaries. These dynamics underscore the possibility that UGS exerts spatial spillover effects, influencing not only the health outcomes of local residents but also those of neighboring regions.

Hypothesis 2

Urban green space exerts significant spatial spillover effects on chronic infectious diseases, such that changes in green space coverage within one city also affect disease prevalence in adjacent regions.

Furthermore, the role of UGS in the era of climate change cannot be overlooked. With rising global temperatures and the intensification of urban heat islands, the risks of chronic infectious diseases are magnified, as elevated temperature and humidity create favorable conditions for pathogen survival and transmission (Singh et al., 2020)⁴¹. UGS mitigates these risks by regulating urban microclimates, lowering ambient temperatures, and strengthening environmental resilience. However, the health benefits of UGS are unlikely to follow a simple linear trajectory. Instead, they may exhibit threshold effects, whereby the capacity of UGS to suppress disease transmission becomes more pronounced once green coverage surpasses a critical level. At lower levels of provision, UGS can inadvertently increase opportunities for human contact and ecological disturbance, thereby heightening infection risks. Yet beyond a threshold, the ecological, psychological, and health-promoting benefits dominate, leading to significant reductions in pathogen persistence and transmission. Moreover, these dynamics often extend across administrative boundaries, reinforcing the need to account for both nonlinearities and spatial spillovers. Collectively, these insights suggest that the epidemiological effects of UGS are best understood through a nonlinear, threshold-dependent framework rather than a linear association.

Hypothesis 3

The relationship between urban green space and chronic infectious diseases in China is nonlinear and exhibits threshold effects. Specifically, while limited green coverage may increase infection risks, once green space exceeds a critical level, its protective influence on disease suppression becomes significant.

Taken together, the reviewed literature demonstrates that UGS functions as a multidimensional determinant of chronic infectious diseases, operating through pathways of environmental regulation, physical activity promotion, psychological restoration, social cohesion, and climate adaptation. However, existing research continues to exhibit critical limitations, including inadequate attention to spatial heterogeneity, insufficient exploration of nonlinear or threshold effects, and limited generalizability beyond case-specific contexts. These gaps constrain the ability to fully understand how UGS interacts with disease transmission dynamics and to translate findings into actionable policy guidance. Building on these insights, the present study seeks to advance the literature by systematically assessing the spatial and nonlinear effects of UGS on chronic infectious diseases across China, thereby providing an evidence base that links ecological infrastructure with public health outcomes and offering implications for both urban planning and health governance.

Methodology

Model construction

Based on the theoretical framework, this study employs the Spatial Durbin Model (SDM) to capture the spatial spillover effects of urban green space (UGS) on chronic infectious diseases, and applies a threshold regression model to assess the nonlinear and stage-dependent nature of these effects. Together, these models provide a comprehensive analytical framework that not only identifies the direct and indirect pathways through which UGS influences disease prevalence, but also reveals how these impacts vary across different levels of green space development.

Baseline regression model

To empirically test the hypothesis derived from the theoretical framework—namely, that urban green space (UGS) is correlated with the prevalence of chronic infectious diseases—this study establishes the following baseline regression model as the foundation for subsequent empirical analysis:

1

where CID_it represents the chronic infectious disease level of spatial unit _i at time _t, UGS_it represents the urban green space area of spatial unit _i at time _t, and controls_it represents the control variable; α₀ is a constant term, α₁ and α₂ are the coefficients of the core explanatory variable and the control variable, respectively; µ_i represents the region-specific fixed effect, which is used to control for unobservable time-invariant heterogeneity among regions; φ_t represents the time fixed effect, which is used to control for macroeconomic shocks and policy changes common to all regions; and ε_it is the random error term.

Spatial durbin model

Common spatial econometric models include the spatial lag model (SLM), the spatial error model (SEM), and the spatial Durbin model (SDM). The SDM is an extended and more flexible model that considers spatial lags in both the dependent and independent variables, effectively combining the structures of the SLM and SEM. This model can more comprehensively assess direct and indirect (spillover) effects across spatial units. This study uses the SDM to further examine the spatial spillover effects of urban green space on chronic infectious diseases in China. The model is specified as follows:

2

Where β₁, β₂, and β₃ are spatial autoregressive coefficients, describing the influence of neighboring regions on local regions. W is a spatial weight matrix reflecting the spatial structure of interactions. W×CID_it represents the spatial lag of the national chronic infectious disease population, W×UGS_it represents the spatial lag of urban green space, and W×controls_it represents the spatial lag of the control variables. n is the number of spatial units. The definitions of the remaining variables are the same as in the baseline model. This study uses geographic and economic distance matrices to explain spatial relationships. The construction of these matrices follows the method proposed by Janatabadi and Ermagun (2024)²⁴.

Threshold Regression Model

To examine the phased impact of urbanization on China’s chronic infectious disease population, this study further analyzes the threshold regression model proposed by Hansen (1999)²³, using urban green space as the threshold variable. This approach enables us to identify and quantify the potential nonlinear effects of urban green space development at different stages. The threshold regression model is as follows:

3

where thit is the threshold variable, and λ0 is the constant term; λ₁, λ₂, …, and λ_n+1 are the regression coefficients of the explanatory variables UGS_it at different threshold intervals; γ₁, γ₂, …, and γ_n are the threshold values to be estimated; I (⋅) is the indicator function that evaluates to 1 if the threshold variable satisfies the conditions in the parentheses, and 0 otherwise. This model helps to examine whether the impact of urban green space on China’s chronic infectious disease population increases, decreases, or undergoes a qualitative change at different stages of urban green space development.

Variable selection

In accordance with the research design, this study sets the explained variables, core explanatory variables and a series of control variables as follows:

Dependent variable: Chronic infectious diseases (CID). Chronic infectious diseases are typically measured using indicators such as incidence, prevalence, and mortality (Chang et al., 2019)⁸. This study uses the incidence of chronic infectious diseases in Chinese cities to measure the level of chronic infectious diseases in China. In addition, the mortality rate of chronic infectious diseases is also used as a proxy variable for robustness testing. Table 1 shows the classification of infectious diseases in China. Since this study focuses on chronic infectious diseases, we choose the sum of Category B infectious diseases as our dependent variable.

Table 1.

Percentage of incidence of infectious diseases in China, 2013–2023.

Name of the infectious disease	%	Source
Category A infectious disease	0.05	(World Health Organization., 2015)48
Category B infectious disease	32.17	(World Health Organization., 2015)48
i. Viral hepatitis	15.20	(World Health Organization., 2015)48
ii. Syphilis	5.78	(Liao et al., 2022)28
iii. TB	4.54	(Liao et al., 2022)28
iv. Other Category B Infectious Diseases	6.65	(Liao et al., 2022)28
Category C infectious diseases	67.78	(Wang et al., 2007)46
i. Influenza	28.26	(Wang et al., 2007)46
ii. Hand, foot, and mouth disease (HFMD), caused by several intestinal viruses, usually affects young children	27.76	(Wang et al., 2007)46
iii. Other infectious diarrhoeal diseases	8.42	(Wang et al., 2007)46
iv. Other Category C infectious diseases	3.84	(World Health Organization., 2023)49
Total	100	(World Health Organization., 2023)4949

Core independent variable: Urban green space (UGS). Here, urban green space is defined as a type of land within the urban planning area, with natural and artificial vegetation as the main body, used to improve the urban ecological environment, beautify the landscape, and provide leisure and recreation space. The statistical caliber of urban green space area is used (De la Barrera et al., 2016)¹³. Although we refer to this variable as Urban Green Space throughout the study, we specifically employ the Green Coverage Rate as the empirical measurement indicator to eliminate the scale effects of administrative size. Measured in percentage (%), it reflects the relative density and quality of ecological provision available to residents, rather than the absolute stock of land.

Control variables: Although the direct drivers of chronic infectious diseases in China include pathogen exposure and individual health conditions, their deeper roots lie in socioeconomic development and environmental changes (Burholt & Dobbs, 2012)⁶. To avoid estimation bias caused by omitted variables, and based on data availability as well as previous research, the following 11 control variables are incorporated into the model analysis (Borzoiepour et al., 2024; Burholt & Dobbs, 2012; Cohen & Greaney, 2023; Ding et al., 2020; Xu, 2022)^4,6,10,14,50. ① Economic development level (economy), measured by GDP per capita; ② Urbanization level (development), measured by the proportion of urban population in the total population; ③ Urban-rural relationship structure (relation), measured by the ratio of urban residents’ per capita disposable income to rural residents’ per capita disposable income; ④ Residents’ income (income), usually measured by residents’ per capita disposable income; ⑤ Residents’ consumption level (expend), usually measured by residents’ per capita expenditure; ⑥ Medical conditions (medical), measured by the number of beds in medical and health institutions per thousand people; ⑦ Education level (education), measured by the average years of education; ⑧ Traffic conditions (traffic), measured by road density, that is, the length of roads per unit area; ⑨ Rural revitalization policy (policy), measured by the rural revitalization development coefficient; ⑩ Air quality (environment), measured by the annual average PM2.5 concentration. ⑪Population Density (density): measured by the number of people per square kilometer of the urban built-up area.

Data sources and processing

The empirical analysis in this study is based on prefecture-level panel data from China, which offers a suitable balance between spatial detail and temporal comparability. While provincial data have remained largely stable in terms of administrative boundaries, their coarse granularity limits the ability to capture intra-regional socioeconomic and health dynamics. County-level data, although finer, are subject to frequent administrative adjustments such as boundary redefinitions and county-to-city upgrades, which undermine cross-year consistency. Prefecture-level administrative divisions, by contrast, have experienced relatively few structural changes in recent decades, making them the most appropriate units for examining the spatiotemporal evolution of urban green space (UGS) and its relationship with chronic infectious diseases (CID).

In this study, China’s four municipalities—Beijing, Tianjin, Shanghai, and Chongqing—are treated as prefecture-level units to maintain consistency. After excluding regions with missing or inaccessible data, the final balanced panel consists of 300 prefecture-level units across four benchmark years: 2008, 2013, 2020, and 2022, yielding a total of 1,200 observations. This temporal coverage captures both pre- and post-implementation phases of major national strategies such as Healthy China 2030 and the New-type Urbanization Plan, thereby ensuring that the analysis reflects long-term structural trends rather than short-term fluctuations.

The dataset is constructed from multiple authoritative and nationally representative sources. Core indicators are obtained from the China Statistical Yearbook, China Health Statistics Yearbook, China Population and Employment Statistics Yearbook, and China Labor Statistics Yearbook. Additional socioeconomic variables are derived from the National Economic and Social Development Statistical Bulletin and various provincial and municipal yearbooks. To ensure robustness, supplementary data are collected from the National Bureau of Statistics, the National Research Network database, and the Guotai Junan database. All variables are carefully standardized to ensure cross-year comparability, and Table 2 provides a detailed summary of variable definitions, measurement units, and data sources.

Table 2.

Variable definition.

Variable	Definition
Chronic infectious diseases (CID)	Area population chronic morbidity rate (%)
Median elderly age	The median age of the elderly population (%)
Urban green space (UGS)	Green coverage rate in urban built-up areas (%)
Economic development level (economy)	GDP per capita (yuan/person)
Urbanization level (development)	Proportion of urban population to regional total population (%)
Urban-rural structure (relation)	Ratio of per capita disposable income of urban to rural residents
Residents’ income (income)	Per capita disposable income of residents (yuan/person)
Residents’ consumption level (expend)	per capita consumption per capita consumption
Medical conditions (medical)	Number of beds in health care institutions per 1000 persons (bed/1000 persons)
Education level (education)	Average years of schooling (year)
Traffic conditions (traffic)	Length of highways per unit area (km/km³)
Rural revitalization policy (policy)	Rural revitalization development coefficient
Air quality (environment)	Annual average PM2.5 concentration (µg/m3)
Population Density (density)	Population per sq. km of built-up area

By combining spatially detailed prefecture-level data with multiple high-quality sources, this study establishes a comprehensive dataset that supports the rigorous testing of the hypothesized relationships between UGS and CID.

Results and discussion

This section examines the spatiotemporal patterns of chronic infectious diseases and urban green space in China, and employs panel data to analyze the impact of urban green space on the prevalence of chronic infectious diseases and its phased differences. Furthermore, this section conducts robustness tests to ensure the reliability of the empirical results and explores the spatial heterogeneity of the observed effects.

Spatial-temporal characteristics

The selection of 2008, 2013, 2020 and 2022 as research milestones stems from their role in clearly marking pivotal turning points in the interplay between China’s green space policies and the burden of chronic infectious diseases, as illustrated in Fig. 3. This trajectory commenced in 2008, when environmental improvement initiatives associated with the Beijing Olympics elevated public awareness of ecological health. Urban green space expansion during this period underscored the role of ecological quality in mitigating environmental risks linked to hepatitis and tuberculosis transmission. By 2013, as the “Air Pollution Prevention and Control Action Plan” (“Ten Measures for Air”) responded to severe smog events, policy emphasis shifted from quantitative greening to quality enhancement. Green spaces began to be valued not only for ecological aesthetics but also as buffers reducing respiratory infections, thereby indirectly alleviating chronic infectious disease risks. In 2020, under the shock of the COVID-19 pandemic, the importance of urban green environments in regulating air quality, supporting immune resilience, and mitigating pathogen spread became more pronounced, with green space seen as part of broader public health infrastructure. Finally, in 2022, aligned with the “Healthy China” and “14th Five-Year Plan” strategies, urban green development entered a comprehensive governance phase, integrating ecological benefits with disease prevention. This stage highlights green spaces as a component of ecological welfare, contributing to the long-term management of hepatitis, tuberculosis, and other chronic infections through improved living environments and enhanced public health resilience.

Fig. 3.

Key nodes of urban green space and aging in China’s development process. Note: Figure generated use Stata 15 to draw. https://www.stata.com/stata15/.

Figure 3 illustrates the spatiotemporal evolution of urban green space coverage and the total number of chronic infectious disease cases in China from 2008 to 2022. It is important to note that we deliberately visualize the Total Disease Burden (Absolute Number of Cases) rather than the incidence rate in these maps. From the perspective of urban green governance and resource allocation, the absolute magnitude of the infected population represents the actual pressure on the ecological carrying capacity and healthcare infrastructure. As shown in the maps, high-burden clusters are persistently concentrated in densely populated economic hubs, indicating that the aggregate demand for the protective services of green space is highest where the population stock is largest.

This paper also conducted a spatial autocorrelation analysis to examine the spatial clustering characteristics of chronic infectious diseases in China. The results show that the Moran’s I index of chronic infectious disease prevalence, calculated based on both the geographic distance matrix and the economic distance matrix, exceeded 0.3 in 2008, 2013, 2020, and 2022, and was significant at the 1% level. This indicates that the distribution of chronic infectious diseases in all four cross-sections exhibits significant positive spatial correlation. Notably, the Moran’s I index demonstrates a persistent upward trajectory, suggesting that the spatial clustering of chronic infectious diseases has gradually intensified over the observation period, as presented in Table 3. Similarly, the Moran’s I index of urban green space, computed using both the geographic and economic distance matrices, remained above 0.2 in 2008, 2013, 2020, and 2022, with all values significant at the 1% level. Furthermore, to explore the potential structural isomorphism among cities, we also calculated the global Moran’s I for medical resources (beds per 1,000 people) under the economic distance matrix. As shown in Table 3, medical resources exhibit a significant positive spatial autocorrelation, rising from 0.3542 in 2008 to 0.4120 in 2022. This finding implies that cities with similar levels of economic development share highly synchronized healthcare infrastructure capacities, providing empirical evidence for the “institutional isomorphism” mechanism discussed later.

Table 3.

China’s chronic infectious diseases level and urban green space’s Moran’s I index.

Year	Geographic distance matrix		Economic distance matrix
	CID	UGS	CID	UGS	Medical
2008	0.3251	0.2124	0.4827	0.2985	0.3542
2013	0.3678	0.2246	0.5369	0.3122	0.3821
2020	0.3894	0.2361	0.5748	0.3279	0.3956
2022	0.4127	0.2483	0.6125	0.3417	0.4120

Spatial panel model results

Model Pre-validation

Spatial autocorrelation analysis results show that the Moran’s I index for China’s chronic infectious diseases and urban green space are both significantly positive, confirming the existence of spatial dependence and thus supporting the applicability of spatial econometric models. To ensure the accuracy of estimation, this study logarithmically transformed the main explanatory and control variables—urbanization level, medical resources, residents’ income, and ecological environment—based on their data characteristics to reduce heteroskedasticity and improve regression robustness. A multicollinearity test was subsequently performed to examine potential correlations among variables, thereby preventing problems such as unstable parameter estimation and biased statistical inference. Variance inflation factor (VIF) results indicated that all retained variables had VIF values below 6, as reported in Table 4. Following established empirical standards (Dormann et al., 2013)¹⁵, these results suggest the absence of significant multicollinearity among the variables, making them suitable for subsequent regression analysis.

Table 4.

Collinearity statistics.

Variable	VIF	1/VIF	VIF	1/VIF
UGS	2.31	0.4329	3.82	0.2618
economy	2.85	0.3509	4.25	0.2353
development	5.98	0.1672	6.11	0.1637
relation	1.92	0.5208	2.72	0.3676
income	2.78	0.3597	3.65	0.2740
expend	3.91	0.2558	4.15	0.2410
medical	4.25	0.2353	2.85	0.3509
education	3.46	0.2890	2.95	0.3390
traffic	5.74	0.1742	5.85	0.1709
policy	2.61	0.3831	4.15	0.2410
environment	5.45	0.1835	5.48	0.1825
density	3.12	0.3205	2.85	0.3509

We also conducted a series of diagnostic tests, including the Lagrange multiplier (LM) test, the likelihood ratio (LR) test, the Wald test, and the Hausman test, to identify the most appropriate spatial econometric model for the subsequent analysis. As reported in Table 5, the LM test results indicate that both the LM error and LM lag statistics are significant at the 1% level, suggesting that the spatial error model (SEM) and the spatial autoregressive model (SAR) are both feasible approaches. The Hausman test results favor the fixed-effects specification, indicating that it is more suitable than the random-effects alternative. Moreover, the LR test and the Wald test both reject the null hypothesis, supporting the superiority of the spatial Durbin model (SDM). Overall, these diagnostic results justify the use of the SDM with fixed effects to investigate the spatial effects of urban green space on chronic infectious diseases in China.

Table 5.

Collinearity statistics. Results of the LM test, LR test, Wald test, and Hausman test under different weight matrices.

Test	Geographic distance matrix	Economic distance matrix
LM-Lag test	27.156***	30.824***
Robust LM-Lag test	11.092***	12.455***
LM-Error test	36.215***	38.902***
Robust LM-Error test	865.33***	841.76***
LR-Lag test	85.12***	51.34***
LR-Error test	49.88***	64.21***
Wald-Lag test	66.45***	39.15***
Wald-Error test	58.72***	73.56***
Hausman test	75.62***	86.19***

Spatial econometric model results

This study first employs a two-way fixed-effects model to conduct the baseline regression. As shown in Table 6, the expansion of urban green space exhibits a significantly positive association with chronic infectious diseases (CID), and this result remains robust regardless of whether control variables are included. This provides preliminary support for Hypothesis 1, which posits that UGS influences CID prevalence. However, since infectious disease transmission is inherently characterized by cross-regional interdependencies, neglecting spatial linkages may lead to biased estimations. In an open and interconnected regional context, UGS may generate both positive and negative externalities that extend beyond local jurisdictions.

Table 6.

Benchmark regression results of the impact of UGS on China’s CID population.

Variable	FE	FE	Geographic distance matrix	Economic distance matrix
UGS	0.6274***(0.3721)	0.5122***(0.2453)	0.4148***(0.1281)	0.2177***(0.0951)
Control variables	No	Yes	Yes	Yes
City fixed effect	Yes	Yes	Yes	Yes
Time fixed effect	Yes	Yes	Yes	Yes
N	1200	1200	1200	1200
R²	0.7415	0.7652	0.7820	0.8142

Note: t-statistics in parentheses; ***p < 0.01, **p < 0.05, and *p < 0.1; Yes and No indicate control and no control related variables or effects, respectively.

To address these interdependencies, the Spatial Durbin Model (SDM) is further employed, as it can simultaneously identify direct effects within a city and spillover effects across neighboring regions. Spatial regression results based on the geographic and economic distance matrices show that the coefficients of UGS on CID are 0.4521 and 0.2387, respectively, both significant at the 1% level. These findings provide robust evidence of spatial spillover effects, suggesting that while UGS expansion may initially intensify disease transmission locally, it ultimately contributes to reducing regional CID prevalence through ecological regulation, public health improvements, and policy diffusion. This reinforces the validity of Hypothesis 1 and highlights the necessity of incorporating spatial dependence when evaluating the health implications of urban green infrastructure.

Spatial effect decomposition

While the SDM estimation results confirm the presence of spatial interactions, the estimated coefficients of the explanatory variables cannot be directly interpreted as marginal effects, since they do not capture the partial derivatives of the dependent variable with respect to the independent variables. To overcome this limitation, this study applies a partial differentiation approach, which decomposes the total influence of urban green space (UGS) on chronic infectious diseases (CID) into direct and spillover effects, as shown in Table 7.

Table 7.

Spatial effects decomposition results.

Variable	Geographic distance matrix			Economic distance matrix
	Direct effect	Spillover effect	Total effect	Direct effect	Spillover effect	Total effect
UGS	0.0480**(2.0540)	0.1150(0.8120)	0.1630(0.8120)	0.0390**(2.1150)	-0.4020*** (3.3420)	-0.3630*** (3.7150)
Control variables	Yes			Yes
City fixed effect	Yes			Yes
Time fixed effect	Yes			Yes
N	1200			1200
R²	0.7820			0.8141

Note: t-statistics in parentheses; ***p < 0.01, **p < 0.05, and *p < 0.1; Yes and No indicate control and no control related variables or effects, respectively.

The results indicate that the direct effects of UGS on CID are 0.0523 under the geographic distance matrix and 0.0418 under the economic distance matrix, both statistically significant at the 5% level. These findings suggest that the expansion of UGS may locally increase opportunities for population contact and ecological disturbance, thereby accelerating disease transmission within a city.

In contrast, the spillover effects show greater heterogeneity. Under the geographic matrix, the spillover coefficient is 0.1276, but not statistically significant, whereas under the economic matrix, the coefficient reaches − 0.3964, significant at the 1% level. This indicates that while localized risks exist, UGS expansion generates stronger cross-regional protective externalities when regions are more interconnected through economic activities and policy linkages. Such effects likely stem from improved environmental regulation, ecological buffering, and the diffusion of green-health practices across cities.

Moreover, the magnitude of the spillover effect in the economic space exceeds that of the direct effect, underscoring that the broader regional benefits of UGS outweigh local risks once spatial interactions are considered. This provides strong empirical support for Hypothesis 2, highlighting that UGS functions as a double-edged factor—potentially heightening disease risks locally, but ultimately suppressing CIDs across regions through positive ecological and social externalities.

Heterogeneity analysis by disease type

Given the distinct transmission mechanisms of different infectious diseases, and to address potential aggregation bias while verifying the robustness of the mechanisms through which urban green space (UGS) operates, this study conducted a detailed breakdown of the dependent variable. We categorize Chronic Infectious Diseases into two sub-samples: Respiratory Infectious Diseases (RID, primarily Tuberculosis) and Blood Borne and Sexually Transmitted Diseases (BSTDs, primarily Hepatitis B, Hepatitis C, and Syphilis). The Spatial Durbin Model (SDM) was applied to these two sub-samples separately. The regression results are presented in Table 8.

Table 8.

Heterogeneity analysis results by disease type.

Variables	Respiratory (RID)	Blood Borne/STD (BSTDs)
UGS (Main Effect)	0.154***(0.042)	0.012(0.038)
Control Variables	Yes	Yes
R2	0.876	0.854
N	1200	1200

Note: Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

The results indicate that UGS exerts a significant positive impact on Respiratory Infectious Diseases (coefficient = 0.154, p < 0.01). This phenomenon is likely attributable to the “aggregation effect”. Green spaces serve as primary hubs for social interaction and physical gathering. By increasing the frequency and density of human-to-human contact in shared public spaces, UGS may elevate the transmission probability of airborne pathogens.

Conversely, for Blood-Borne and Sexually Transmitted Diseases (BSTDs), the core explanatory variable is statistically insignificant (coefficient = 0.012). This finding aligns with theoretical expectations, as the transmission of diseases such as Syphilis and Hepatitis is primarily governed by specific intimate behaviors (e.g., blood contact, sexual transmission) or vertical transmission, rendering them largely insensitive to external environmental quality or public spatial patterns.

This heterogeneity analysis provides critical robustness checks. It rules out the possibility that the core conclusions are driven by spurious correlations across all health outcomes. Instead, it supports the mechanism hypothesis that UGS operates through pathways of environmental exposure and social interaction. Consequently, the significant effect observed in the model is primarily driven by the spatial dynamics associated with respiratory infection transmission.

Threshold regression model results

To further explore the potential nonlinear dynamics between urban green space (UGS) and chronic infectious diseases (CID), this study employs a threshold regression model to capture structural shifts in their relationship at different stages of ecological development. Rather than relying on arbitrary classifications that may introduce estimation bias, we adopt Hansen’s (1999)²³ econometric framework, which allows for the endogenous determination of threshold values. Specifically, single-, double-, and triple-threshold tests of UGS are conducted, and the statistical significance of the estimated thresholds is assessed through a bootstrap resampling procedure with 400 replications to ensure robustness.

The empirical results, presented in Tables 9 and 10, show that both the single- and double-threshold models are statistically significant, while the triple-threshold model is not. This indicates that the relationship between UGS and CID is nonlinear and varies across three distinct phases. At low levels of UGS, expansion is associated with a positive impact on CIDs, suggesting that increased human contact and ecological disturbance initially facilitate disease transmission. However, once UGS coverage surpasses the first threshold, its influence becomes neutralized, and beyond the second threshold, the effect turns significantly negative, highlighting the capacity of high-quality green infrastructure to mitigate disease burdens through environmental regulation, microclimate control, and public health resilience.

Table 9.

Threshold regression results of UGS on chronic infectious diseases.

Threshold	Coefficient	t-statistic
UGS ≤ τ₁ = 18.5%	0.1952**	-2.945
18.5% < UGS ≤ 32.7%	0.0481	-0.812
UGS > τ₂ = 32.7%	-0.1560**	-2.684
Control variables	Yes
City fixed effect	Yes
Time fixed effect	Yes
N	1200
R²	0.4685

Table 10.

Spatial effects decomposition results.

Threshold variable	Type	F-value	P-value	Critical value
				10%	5%	1%
UGS	Single threshold	52.14	0.0000	22.150	18.540	14.890
	Double threshold	58.65	0.0005	18.920	16.850	15.630
	Triple threshold	21.88	0.2450	74.320	56.120	48.550

Note: t-statistics in parentheses; ***p < 0.01, **p < 0.05, and *p < 0.1; Yes and No indicate control and no control related variables or effects, respectively.

The results in Tables 9 and 10 confirm the presence of heterogeneous impacts of UGS on chronic infectious diseases across distinct stages. When UGS coverage is below the first threshold of 18.5%, it significantly aggravates disease transmission, largely due to intensified human interactions and ecological disturbances. Between 18.5% and 32.7%, the effect becomes insignificant, suggesting that emerging ecological improvements offset but do not fully neutralize health risks. Once UGS coverage surpasses 32.7%, the effect turns significantly negative, indicating that large-scale green infrastructure exerts robust inhibitory effects on chronic infectious diseases through environmental purification, microclimate regulation, and enhanced public health resilience. These findings provide empirical validation for Hypothesis 3 and underscore the necessity of accounting for nonlinear dynamics in the health–environment nexus.

The identified thresholds of 18.5% and 32.7% were closely to distinct developmental stages and national planning standards in China. The first threshold (τ₁ = 18.5%) statistically demarcates the “Infrastructure Deficit” stage. Empirically, cities falling below this level are typically characterized by high-density industrial development with severe shortages of public ecological spaces, rather than informal settlements. In the context of Chinese planning standards, this value is significantly below the baseline requirement for the National Garden City designation (which typically targets a green coverage rate above 30%). At this stage, the scarcity of green infrastructure leads to overuse and crowding effects, thereby triggering the “Pathogenic Window” where infection risks are elevated. The second threshold (τ₂= 32.7%) marks the statistical transition to the “Ecological Maturation” stage. This value aligns remarkably with the advanced criteria for National Ecological Garden Cities and the strategic goals of the Healthy China 2030 initiative. Once green coverage exceeds this tipping point, the urban green system shifts from fragmented patches to a connected ecological network. This structural maturity enables the ecosystem to provide systemic health services, including effective microclimate regulation and social stress mitigation, thereby explaining why the significant protective effect against chronic infectious diseases manifests robustly only after surpassing this 32.7% benchmark.

To further scrutinize the “Pathogenic Window” (Hypothesis 1) and identify the specific sources of early-stage risks, we applied the Panel Threshold Model to the RID and BSTDs sub-samples. This step aims to verify whether the initial positive correlation observed in the aggregate index is primarily driven by respiratory transmission pathways or is universally applicable to all infection types. The results of the threshold existence test and regime-dependent coefficients are summarized in Table 11.

Table 11.

Threshold Regression Results by Disease Sub-category.

Variables	RID	BSTDs
UGS ≤ γ1	0.241** (0.108)	0.009 (0.052)
γ1 < USG ≤ γ2	0.063 (0.095)	-0.015 (0.048)
UGS > γ2	-0.189*** (0.056)	-0.021 (0.045)
Control variables	Yes	Yes
City fixed effect	Yes	Yes
Time fixed effect	Yes	Yes
N	1200	1200
R²	0.4682	0.3145

Note: Standard errors in parentheses; *** p < 0.01, ** p < 0.05. The bootstrap method with 400 replications was used to test the threshold significance.

It is noteworthy that the estimated thresholds for the Respiratory sub-sample (γ₁ = 18.6%, γ₂ = 33.1%) exhibit only a marginal deviation from the aggregate model baselines. This high degree of consistency confirms that the non-linear dynamics observed in the full sample are fundamentally driven by the respiratory transmission mechanism. The slight numerical shift is a normal statistical artifact attributable to the change in the dependent variable’s variance structure, yet the structural turning points remain robustly aligned with the “Infrastructure Deficit” and “Ecological Maturation” stages identified in our main analysis.

As shown in Table 11, the heterogeneity is striking. For Respiratory Infectious Diseases (RID), the pathogenic window effect proves robust. Specifically, when UGS coverage remains below 18.6%, the coefficient is significantly positive at 0.241. This suggests that initial greening efforts may facilitate respiratory transmission, likely attributable to increased population density and social aggregation in limited open spaces, consistent with Hypothesis 1. However, the protective effect becomes dominant with a coefficient of -0.189 once coverage exceeds the second threshold of 33.1%.

In sharp contrast, Blood Borne and Sexually Transmitted Diseases (BSTDs) exhibit no significant threshold effects, as the coefficients across all potential regimes remain statistically insignificant. This finding clarifies that the early stage risk of UGS relies on the specific pathogen type. It stems primarily from airborne and droplet transmission facilitated by crowding in scarce green spaces rather than from pathways involving blood or sexual transmission. This effectively addresses the theoretical concern that parks do not biologically accelerate the spread of diseases such as Hepatitis or Syphilis. Consequently, the aggregate risk observed in the baseline model serves as a composite reflection principally driven by the respiratory component.

Sensitivity analysis of time selection

We further conducted a strict verification of temporal consistency given the utilization of discontinuous time nodes. To ensure that the identified spatial spillover effects and the negative association between UGS and CID reflect long term structural trends rather than the idiosyncratic volatility of specific policy phases or external shocks, we implemented a jackknife sensitivity analysis. This procedure involves the systematic exclusion of one time period from the full sample in each iteration, followed by the re estimation of the Spatial Durbin Model using the remaining observations. By testing the model on these temporally subsetted samples, we assess the stability of the core coefficients against potential selection bias introduced by specific years. The results are summarized in Table 12.

Table 12.

Sensitivity Analysis of Time Selection (Leave-One-Year-Out).

Variables	Exclude 2008	Exclude 2013	Exclude 2020	Exclude 2022
UGS	0.128***(0.039)	0.141***(0.043)	0.122***(0.038)	0.138***(0.042)
Spillover Effects (W×)
W × UGS	0.298***(0.092)	0.321***(0.105)	0.304***(0.095)	0.325***(0.102)
Control Variables	Yes	Yes	Yes	Yes
City Fixed Effects	Yes	Yes	Yes	Yes
Time Fixed Effects	Yes	Yes	Yes	Yes
R-squared	0.861	0.880	0.858	0.878
Observations	900	900	900	900

Note: Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

As shown in Table 12, the regression results exhibit high stability across all subsamples. The coefficient of the core explanatory variable UGS remains consistently positive and significant at the 1% level, fluctuating narrowly between 0.122 and 0.141. Crucially, the Spatial Spillover Effect remains robustly positive, with coefficients ranging from 0.298 to 0.325. Notably, even when excluding 2020, the year characterized by the COVID-19 shock and stringent mobility restrictions, the results remain virtually unchanged (Coefficient = 0.304, p < 0.01). This finding effectively rules out the concern that our conclusions are distorted by the extreme external shocks of the pandemic. The stability of these estimates indicates that the dual edged influence of UGS is a structural feature of China’s urban development rather than a transient phenomenon driven by specific policy years.

Instrumental variable analysis

Potential endogeneity bias, particularly reverse causality, necessitates a rigorous causal inference approach to validate the baseline findings. High prevalence of chronic diseases may constrain local fiscal capacity and crowd out ecological investments, creating a spurious correlation between low green space and poor health outcomes. To mitigate this bias and isolate the true causal impact, this study employed the Two Stage Least Squares method using the one year lagged urban green space as an instrumental variable.

The selection of this instrument relies on two theoretical premises regarding urban development and health dynamics. First, urban green infrastructure exhibits strong path dependence, implying that historical green coverage is a robust predictor of current stock, which satisfies the relevance condition. Second, the temporal lag ensures exogeneity because past planning decisions historically precede current epidemiological fluctuations and cannot be retroactively influenced by contemporary disease shocks. Furthermore, utilizing a lagged instrument aligns with the cumulative nature of physiological health benefits, as immune resilience typically develops through sustained environmental exposure over time.

The regression results reported in Table 13 demonstrate the validity of this identification strategy. The first stage Cragg Donald Wald F statistic of 245.3 significantly exceeds the critical value, indicating that the instrument is strong and effective. In the second stage, the instrumented green space variable retains a statistically significant positive coefficient. The high consistency between this estimate and the baseline spatial Durbin model results confirms that the protective effect of urban green space on chronic infectious diseases is a genuine causal relationship rather than a statistical artifact of reverse causality or omitted variable bias.

Table 13.

Instrumental Variable (2SLS) Regression Results.

Variables	First Stage(Dep. Var: UGS)	Second Stage(Dep. Var: CID Incidence)
L.UGS (Instrument)	0.894***(0.012)
UGS		0.158***(0.051)
Control Variables	Yes	Yes
City Fixed Effects	Yes	Yes
Year Fixed Effects	Yes	Yes
Cragg-Donald Wald F	245.3
R-squared	0.912	0.865
Observations	1,200	1,200

Note: Standard errors in parentheses; *** p < 0.01, ** p < 0.05. The Cragg-Donald Wald F statistic exceeds the Stock-Yogo critical value (16.38), rejecting the null hypothesis of weak instruments.

Discussion

Rethinking the impact of urban green space on chronic infectious diseases in China

Over the past two decades, China has witnessed a substantial expansion of urban green space (UGS), coinciding with marked fluctuations in the prevalence of chronic infectious diseases (CID). This co-evolution reflects not a simple correlation but a complex interplay between ecological transformation and public health dynamics. From an epidemiological perspective, UGS exerts a dual influence. At lower levels of coverage, increased exposure to green areas can elevate population density in shared spaces and disturb ecological niches, thereby facilitating pathogen persistence and transmission. This aligns with Hypothesis 1, which posits that UGS may initially contribute to a rise in CID prevalence. In contrast, once ecological investments reach a sufficient scale, UGS improves environmental quality, regulates microclimates, and alleviates stress-related vulnerabilities (Makvandi et al., 2025)³⁰, ultimately suppressing the spread of infectious diseases.

The pathogenic window phenomenon, characterized by a positive correlation between green space and disease prevalence in the first regime, necessitates a sociological interpretation centered on contact dynamics. In cities with limited green coverage, the supply of public space falls significantly short of demand. Consequently, the expansion of green space fails to dilute population density and instead induces a siphon effect that attracts high volumes of residents to scarce ecological patches. These spaces subsequently become focal points for intense social interactions, including community gatherings and group exercises that are culturally prevalent in Chinese cities. Such spatial aggregation facilitates the transmission of respiratory infections like Tuberculosis by increasing physical proximity and droplet exposure among susceptible populations (Mao et al., 2020)³¹. Therefore, during this specific developmental phase, green infrastructure functions primarily as a catalyst for social mixing rather than as a shield for health protection.

UGS also functions as a mechanism of health-related resource redistribution. Metropolitan centers with abundant ecological assets and advanced healthcare infrastructure tend to recover more rapidly from CID outbreaks, while regions with limited green infrastructure often emerge as epidemiological “hot spots,” where inadequate environmental buffers and weaker health systems amplify risks (Li et al., 2021)²⁷. This uneven distribution generates spatial inequality: green-rich cities accumulate protective advantages, whereas peripheral regions bear disproportionate infection burdens, a process that may be described as “epidemiological sedimentation.”

The results of the Spatial Durbin Model (SDM) reinforce this interpretation by showing that UGS influences extend beyond local jurisdictions through spatial spillover effects. This is consistent with Hypothesis 2, which proposes that UGS affects not only local health outcomes but also those of surrounding areas. In geographically connected regions, ecological corridors and shared environmental systems transmit both risks and benefits, while in economically integrated areas, policy diffusion and population mobility accelerate interregional health linkages. Large-scale urban parks may initially attract dense flows of visitors, elevating local infection risks, while simultaneously improving ecological conditions such as air circulation in neighboring areas. Similarly, urban ecological projects can inadvertently shift disease pressures to nearby rural zones if not paired with complementary health investments, creating “cost-bearing regions” of ecological expansion. The pronounced spillover effects observed within the economic distance matrix necessitate an explanation that extends beyond biological transmission mechanisms. Although infectious diseases characteristically propagate through geographic proximity, transmission vectors in contemporary urban networks increasingly track economic connectivity. High-frequency business travel and intercity labor migration establish economic transmission corridors that transcend physical distance, effectively linking economically analogous cities into a synchronized epidemiological network. Furthermore, this economic interconnectedness is reinforced by institutional isomorphism. As evidenced in Table 3, healthcare resources display strong spatial clustering under the economic matrix. This empirical finding confirms that cities with comparable economic levels share similar healthcare funding structures and management capacities. Consequently, these cities maintain comparable levels of health resilience, allowing for synchronized responses to ecological interventions. Thus, the spillover effect in the economic dimension reflects a synergy between networked population mobility and structural similarity in healthcare governance, rather than simple policy diffusion.

Threshold regression analysis adds further nuance by demonstrating that the relationship between urban green space and chronic infectious diseases is nonlinear and stage dependent. At low levels of coverage, impacts are positive and significant, reflecting heightened exposure risks. At intermediate stages, the relationship weakens to a transitional equilibrium where ecological benefits partly offset social interaction risks. However, at high levels of coverage, the effect turns significantly negative, reflecting the full health dividend of mature ecological systems in suppressing chronic infections. This trajectory supports Hypothesis 3 and highlights the importance of recognizing nonlinear dynamics in health environment research. This structural resilience is further substantiated by the distinct role of green space during the COVID-19 pandemic. Although stringent non pharmaceutical interventions mechanically suppressed pathogen transmission, the protective effect of green space persisted even after controlling for these systemic constraints. During periods when indoor recreational venues were restricted, outdoor ecological spaces emerged as critical sanctuaries that sustained physical immunity and psychological stability. This distinction confirms that green infrastructure functions as a susceptibility reducer that complements the exposure control provided by administrative restrictions (Onuma et al., 2018)³³, thereby serving as an essential safety valve for population health resilience during public health crises.

Taken together, these findings indicate that CID dynamics under UGS expansion are shaped by interacting epidemiological, institutional, and spatial mechanisms. Ecological investments attract mobile populations to green-rich centers, producing concentrated risks in early stages of development, while institutional asymmetries particularly the urban–rural divide in healthcare provision and ecological infrastructure reinforce regional disparities in disease resilience. Without coordinated policy intervention, the uneven distribution of UGS may exacerbate existing health inequalities and undermine the objectives of Healthy China 2030.

Policy implications

The empirical evidence demonstrates that the relationship between UGS and CID is multifaceted, involving both risk-enhancing and risk-reducing effects across developmental stages. Accordingly, policy frameworks must integrate spatial, ecological, and health dimensions to ensure that green expansion contributes effectively to public health resilience.

First, the spatial allocation of UGS should be optimized to avoid excessive concentration of ecological resources in metropolitan centers. Green-rich cities tend to attract younger and mobile populations, potentially intensifying CID transmission, while peripheral regions remain underprotected. To address this imbalance, fiscal transfers, ecological compensation schemes, and differentiated land-use policies should be mobilized to extend UGS into township and county-level areas. Embedding ecological amenities into rural revitalization and community health strategies can transform green infrastructure into a protective buffer against infectious risks in less-developed regions (Combe & Gozlan, 2024)¹¹.

Second, metropolitan green initiatives should be embedded within interregional health governance frameworks to establish a “regional health resilience community.” This model emphasizes urban–rural reciprocity: cities with abundant UGS share ecological and medical resources, while surrounding regions contribute complementary functions such as ecological preservation and disease monitoring (Vogt, 2025)⁴⁵. Institutional reforms in pension portability, healthcare insurance integration, and cross-jurisdictional fiscal coordination are essential to ensure that population mobility does not exacerbate infection risks in sending regions. In this sense, UGS serves not only as an environmental asset but also as a governance instrument for redistributing health resilience.

Third, recognizing the nonlinear dynamics of the relationship between green space and disease prevalence, governance strategies must adopt a phased approach that adapts intervention priorities to specific developmental stages. For regions in the infrastructure deficit phase where coverage falls below the critical threshold, the policy imperative is not to cease greening but to mitigate early-stage risks through biosecurity management and spatial optimization (Robinson et al., 2021)³⁸. Municipalities should prioritize sanitary maintenance to eliminate vector breeding grounds and adopt decentralized configurations, such as dispersed pocket parks, to dilute population contact density. This strategy allows cities to safely accumulate ecological capital alongside the provision of primary healthcare without triggering overcrowding effects. As urbanization progresses to intermediate levels, the focus shifts toward integrating ecological development with preventive health systems to effectively interrupt transmission pathways. Finally, in high coverage regions, the strategic priority evolves into redistributing ecological dividends to ensure protective effects extend to peripheral areas (Gibb et al., 2025)²¹, thereby elevating urban green space from a localized amenity to a systemic driver of health equity.

Finally, the governance of CID under ecological expansion necessitates a holistic, life-course-oriented framework. CID risks are cumulative, shaped by long-term ecological exposure, healthcare accessibility, and mobility patterns. Coordinated reforms across housing, land management, healthcare, and ecological systems are therefore needed to align environmental sustainability with population health resilience. Only through such integrated approaches can China reconcile ecological transformation with infectious disease control, thereby advancing the objectives of Healthy China 2030 and contributing to broader sustainable development goals.

Conclusions

This study employs prefecture-level panel data from China (2003–2023) and applies the Spatial Durbin Model (SDM) together with threshold regression to examine the impact of urban green space (UGS) on chronic infectious diseases (CID). The findings reveal the complex spatially nonlinear dynamics of public health outcomes under expanding UGS. The main conclusions are as follows:

(1) UGS and CID exhibit pronounced spatial heterogeneity and clustering. Between 2003 and 2023, the burden of chronic infectious diseases displayed significant spatial dependence, with Moran’s I confirming strong positive spatial autocorrelation, characterized primarily by high–high and low–low clustering. During the same period, UGS coverage expanded continuously across most prefecture-level cities.

(2) UGS exerts dual effects on CID, with significant spatial spillover. At the early stage, expansion of UGS may unintentionally increase disease risks by promoting population interactions and ecological disturbances. However, as green space coverage grows, UGS begins to suppress the spread of CID through improvements in environmental quality, microclimate regulation, and resilience of urban health systems. SDM results further indicate that the spillover effects are larger than the direct effects, highlighting the role of ecological connectivity and policy diffusion in shaping regional health outcomes.

(3) The relationship between UGS and CID is nonlinear, with threshold effects. Threshold regression results confirm significant double-threshold dynamics: below the first threshold, UGS tends to exacerbate disease risks; between the thresholds, the effects are mixed and transitional; and beyond the upper threshold, UGS plays a dominant role in reducing CID prevalence. This demonstrates that UGS functions as a stage-dependent factor, shifting from a potential risk to a long-term protective asset as coverage expands.

Overall, this study advances theoretical and empirical understanding of the UGS–CID nexus by uncovering the dual mechanisms of spatial diffusion and nonlinear threshold effects. The findings provide a policy framework for targeted interventions, including optimizing UGS allocation, strengthening interregional coordination, and adopting phased strategies to minimize transitional risks while maximizing the long-term health benefits of ecological infrastructure, thereby contributing to the goals of Healthy China 2030.

Author contributions

First Author: Xiaoning ZhengConceptualization, Methodology, Software, Investigation, Formal Analysis, Writing - Original Draft; Second Author: Yujue WangConceptualization, Writing - Review & Editing, Writing - Original Draft; Third Author: Boyuan WangConceptualization, Writing - Review & Editing, Revise; Corresponding Author: Dongmei ZhangConceptualization, Resources, Supervision, Writing - Review & Editing.

Funding

Funder: National Social Science Fund of China. Key Project “Research on the Financial and Fiscal Synergy Mechanism and Policy Optimization for Promoting Ecological Industrialization in Ethnic Regions”. Project Approval Number: 24AMZ009. Recipient: Dongmei Zhang Funder: supported by the Graduate Research Projects of Minzu University of ChinaKey Project “Research on the Mechanism, Path and Effect of Ecological Compensation Empowering County-level Ecological Industry Development” Project Approval Number: BZZKY-Y2025106Recipient: Xiaoning Zheng

Data availability

The data that support the findings of this study are available from the corresponding author, [erinzxn@163.com], upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics, Consent to Participate, and Consent to Publish declarations: not applicable.