Abstract

Background

The health of the population serves as a cornerstone for sustainable economic development and stands as a vital indicator of national prosperity and strength. Based on the theory of health human capital, this study explores the practical issues of enhancing residents' health levels through the synergy of social, ecological, and medical factors.

Methods

Utilizing provincial panel data from China spanning 2011 to 2020, this research employs entropy methods and dynamic Qualitative Comparative Analysis (QCA) to measure the health levels of Chinese residents. The study analyzes the synergistic roles of social, ecological, and medical factors in improving health from both a temporal perspective and a configurational viewpoint.

Results

No single factor is a necessary condition for achieving high health levels among residents. However, the necessity of per capita health expenditure has been increasing yearly, showing a temporal effect. There are four distinct pathways that can lead to high health levels, which can be further categorized into social-ecological-medical synergistic, social-ecological, and social-medical driven configurations. Temporally, the consistency of configurations 1 and 3 collectively decreased in 2020, likely due to the outbreak of COVID-19, which temporarily shifted the core factors affecting residents' health. Spatially, there are no significant regional effects among the four configurations, indicating that their explanatory power does not significantly differ across provinces.

Conclusions

The collaborative efforts of social, ecological, and medical factors are found to significantly enhance the health levels of residents in China. Drawing on existing research in the field of public health, the study proposes an analytical framework for understanding the factors influencing health levels. By integrating a configurational perspective into the study of health human capital theory, this research broadens the analytical lens used to examine health levels, offering important theoretical insights and policy implications.

Background

The enhancement of residents' health levels can significantly propel social development [1]. However, due to the vast geographic expanse and the varying levels of economic development across China, there is a pronounced disparity in health levels among different provinces. This disparity hinders the achievement of the cohesive objective of a 鈥淗ealthy China鈥�. Numerous scholars have extensively explored the interconnection between health and social development.

Grossman [2] has proposed a groundbreaking theoretical framework for the health production function. In Grossman's model, health fulfills three roles: firstly, it is a consumer good; secondly, it serves as an input (a factor of production); and thirdly, it is a form of capital. Grossman's theoretical model posits that individual health is influenced by income, education, and medical resources. Building on Grossman's model, scholars have conducted a plethora of empirical studies and applications, and the veracity of the model鈥檚 conclusions has been largely substantiated, providing a scientific theoretical foundation for the study of health issues.

Socio-economic determinants, including income level and educational attainment, are intricately associated with health outcomes [3]. Economic fluctuations can adversely impact health levels [4], and the health status of low-income individuals is, generally, poorer compared to their high-income counterparts [5]. Education is one of the determinants of life expectancy [6, 7], with individuals possessing higher income and education levels typically enjoying better health [8, 9].

Environmental factors constitute another crucial aspect affecting the health levels of Chinese residents. The accelerated pace of industrialization and urbanization has magnified the problem of environmental pollution. Contamination of air, water, and soil not only disrupts ecological equilibrium but also constitutes a direct threat to human health. Empirical research has demonstrated that respiratory diseases resulting from air pollution represent a substantial risk to public health [10]. The improvement in air quality brought about by clean heating policies greatly benefits residents' health [11]. Reducing environmental pollution can help mitigate income-related health inequalities [12]. Concurrently, there is growing evidence that air pollution detrimentally affects residents' mental health [13].

Medical factors [14] represent the third pivotal element influencing the health standards of Chinese residents. Research indicates that increased health investment can enhance residents' health levels. Intensive medical interventions resulting from such investments can notably reduce mortality rates. Moreover, for developing countries, the efficiency and equity of government health investments hold more significance than the scale of investment [15].

In this context, examining the determinants and pathways to enhancing health levels is crucial for China to achieve objectives such as the ongoing improvement of public health by 2030 and the attainment of the comprehensive strategic goal of establishing a healthy nation in harmony with a modernized socialist society by 2050. Consequently, a comprehensive analysis will be undertaken to examine the determinants affecting the health levels of Chinese residents, including socio-economic, environmental, and medical factors, from a group perspective. It will explore effective pathways and strategies to improve health levels based on these factors, providing references and lessons for the development of national and regional healthcare initiatives.

This study aims to investigate how various combinations of social, ecological, and medical factors synergistically impact the health of residents across different provinces in China. This investigation will adopt a group perspective, utilizing entropy and dynamic Qualitative Comparative Analysis (QCA) methods. The prospective contributions of this paper are delineated as follows: (1) Drawing upon the theory of health human capital, this paper proposes a theoretical framework for understanding the synergistic effects of different driving factors on the health levels of residents, thereby providing a reference for related empirical research. (2) It offers a comprehensive examination of the impacts of social, ecological, and medical factors on residents' health levels, thereby enhancing the scholarly discourse on the external factors contributing to health improvement. (3) By employing an advanced dynamic QCA method, this study bridges the methodological gap between panel data and the QCA approach. Furthermore, it explores, from a configurational perspective, the collaborative pathways through which social, ecological, and medical factors drive improvements in residents' health levels over time.

The study investigates the potential pathways for enhancing the health levels of residents by examining the interplay among societal, ecological, and medical factors within a configurational framework, as outlined in Table听1. The pathways are categorized into three distinct models: single leading, dual-combination, and multiple collaborative paths.

• (1) Single Dominant Lifting Path: This perspective explores the potential for a dominant influence among social, ecological, and medical factors. Although these factors collectively contribute to residents' health, their influence varies across different provinces. It remains essential to investigate the optimal approach, whether focusing on social leadership in providing economic and educational resources, emphasizing ecological environmental protection, or enhancing the accessibility of healthcare resources, to ultimately improve residents' health.

• (2) Dual-Combination Enhancement Path: Here, the focus is on the potential for complementary and symbiotic relationships among social, ecological, and medical factors. For example, effective medical care depends on a robust social foundation, where higher economic levels, resident income, and education contribute to improved material conditions and health awareness, thereby promoting better health level.

• (3) Diversified Synergistic Enhancement Path: This approach considers multiple synergistic pathways, wherein social, ecological, and medical factors collectively advance population health. Given China's vast and diverse landscape, regional differences in natural resource endowments must be accounted for. By integrating and coordinating social, ecological, and medical resources, regions can drive improvements in health levels, forming a comprehensive 鈥渟ocial-ecological- medical鈥� synergistic model.

Table 1 Possible pathways for improving the health of the population

This paper considers the combined effects of social, ecological, and medical factors from a health configuration perspective. Key factors such as economic status, income, education, ecological conditions, availability of healthcare professionals, and health expenditure per capita were identified as precursors. The study employs the entropy weight method to integrate these indices into a composite measure of health, used as the outcome variable. A theoretical framework outlines the synergistic impacts of various driver combinations on residents' health, as depicted in Fig.听1.

Fig.1

Theoretical framework

From the perspective of configurational analysis and employing the dynamic Qualitative Comparative Analysis method, a theoretical framework is developed to examine the multifactorial collaborative influences on residents' health levels. It breaks through the traditional single-factor perspective in existing research and extends the scope of the health human capital theory. The introduction of the dynamic Qualitative Comparative Analysis method overcomes the barriers between panel data and Qualitative Comparative Analysis, providing a novel tool for exploring the factors influencing health levels within a temporal dimension. The proposed framework not only uncovers the pathways of multi-factor synergy but also offers policymakers concrete strategies for optimizing residents' health levels, thereby holding significant practical implications.

Methods

Data construction

This research examines the synergistic impacts of six conditional variables, including economic level, residents' income, education level, ecological level, supply of health technicians, and per capita health cost on population health levels. The analysis is grounded in the human capital theory of health and is conducted from three perspectives: social, ecological, and medical. The outcome variables, along with the six conditional variables, were operationalized as follows:

Outcome variables

Health Level: The primary health indicators as outlined in the 鈥淗ealthy China 2030鈥� initiative include per capita life expectancy, infant mortality rate, mortality rate of children under five, maternal mortality rate, and the proportion of urban and rural residents meeting or exceeding the National Physical Fitness Standards. Based on the availability of data and specific measurement criteria, the study ultimately selects average life expectancy, incidence rate of infectious diseases, perinatal mortality rate, and maternal mortality rate to assess the population's health level. Life expectancy is a core indicator for measuring the overall health level of residents, as it comprehensively reflects a region's healthcare standards, living conditions, socioeconomic development, and disease control capacity. As a globally recognized measure of health, life expectancy facilitates cross-regional and international comparisons. The selection of infectious disease incidence is due to the direct threat posed by the prevalence of infectious diseases to residents' health, particularly in regions with limited medical resources, where the incidence rate significantly impacts the overall health status of the population. Perinatal mortality, which includes late fetal deaths and early neonatal deaths, is an important indicator of maternal and infant health. It is widely used internationally as a health metric, enabling cross-regional and international comparisons with high applicability. Compared to other related indicators (such as neonatal mortality or infant mortality), perinatal mortality has broader coverage and stronger explanatory power, making it a reasonable and scientific choice as one of the indicators for assessing residents' health levels. Maternal mortality directly reflects women's health during pregnancy, childbirth, and the postpartum period, as well as the quality of medical services available. The level of maternal mortality is often closely associated with a country or region's socioeconomic development, healthcare resource allocation, and the social status of women. Therefore, it serves as a critical indicator for evaluating the health level of residents in a given country or region. This research adopts methodologies from Feng [16] and Luo [17], initially utilizing the entropy value method to determine the weights of these four indicators. Subsequently, a weighted sum is computed to derive a comprehensive measure of the population's health level. Average life expectancy is treated as a positive indicator, while the incidence rate of infectious diseases and perinatal mortality rate are considered negative indicators. These comprehensively measured indicators are utilized as proxy variables to represent the population's health level.

Conditional variables

• (1) Economic Level: Economic level provides the foundation for ensuring population health. Following the methodology of Wang et al. [18], GDP per capita is utilized as a logarithmic measure of regional economic status.

• (2) Population Income: Income serves as a significant material foundation for health enhancement. In alignment with the methodology of Liu et al. [19], the logarithm of per capita disposable income is utilized as a proxy variable for assessing income level.

• (3) Educational Attainment: Education significantly enhances population health; generally, individuals with higher educational attainment tend to have longer life spans. This may be attributed to the propensity of educated individuals to invest in long-term health-related assets and to possess greater health awareness, leading to more timely and suitable medical interventions. Average years of schooling reflects the overall educational attainment of a region or country, encompassing the coverage of education from primary to higher education rather than being confined to a specific stage. This measure better captures differences in educational levels across regions and exhibits greater applicability. Data on average years of schooling are generally more readily available, and the statistical methods used are relatively standardized both domestically and internationally, facilitating cross-regional comparisons. Accordingly, drawing on Zhang's [20] methodology, the average years of schooling is adopted as a measure of educational attainment.

• (4) Ecological Environment Level: Residents' health is intrinsically linked to the quality of the ecological environment. Poor environmental quality, high pollution levels, and increased mortality rates negatively impact residents' health. Informed by Zhou's [21] research, the entropy value method is utilized to assess environmental pollution, encompassing wastewater (measured by per capita chemical oxygen demand and ammonia nitrogen emissions), exhaust gas (measured by per capita SO₂ emissions), and solid waste (measured by per capita industrial waste and urban garbage removal). The ecological environment level, ensuring a positive index, is calculated using Formula (1).

  $$Ecological\,environment\,level=1-Environment\,pollution\,level$$

  (1)

• (5) Supply of Health Technicians: An adequate supply of health personnel is essential to deliver quality healthcare services, thereby addressing the health consumption needs of the population and enhancing health outcomes [22]. As per Zhang's [23] practice, the number of health technicians per 1,000 population is used as a metric.

• (6) Per Capita Cost of Health: Health expenditure is a critical input in the health production system. It is quantified as the ratio of total health expenditure for a specific year to the average population during that period, reflecting the overall societal investment in healthcare services over that timeframe, typically one year. Based on Kaladharan's [24] study, per capita health cost is utilized as the measurement.

The data sources for each variable are presented in Table听2.

Table 2 Variable data sources

Descriptive statistical analysis

To analyze the descriptive statistics of each variable and explore the differences in outcome variables and condition variables across provinces, this analysis provides a foundation for subsequent research. The descriptive statistics of the outcome and condition variables are shown in Table听3.

• (1) Outcome Variables: From the perspective of outcome variables, the average health level of residents is 0.686, which indicates that there is still significant room for improvement in the health levels of Chinese residents. The standard deviation is less than 2, suggesting that the data exhibits relatively low variability overall.

• (2) Condition Variables: From the perspective of condition variables, economic level, population income, educational attainment, ecological environment level, supply of health technicians, and per capita cost of health are 1.470, 2.952, 9.694, 0.280, 6.234, and 3134.400, respectively. These figures reflect that the indicators for condition variables in China are generally at a relatively high level. Except for per capita cost of health, the standard deviations of other variables are all less than 2, indicating relatively low variability in the data overall. However, due to differences in regional economic development, the disparity in per capita cost of health is significant, with a gap of 10,609.472 between the maximum and minimum values. This highlights substantial regional differences in healthcare funding allocation.

Table 3 Descriptive statistical analysis

Research methodology

Dynamic QCA

Dynamic QCA is an analytical method that combines a configurational perspective with a temporal dimension, allowing for the examination of interactions among complex social, ecological, and medical factors. It is particularly well-suited for studying complex systems such as population health levels, which are influenced by multiple factors. Fixed effects models analyze the linear impact of a single factor on the outcome variable, assuming linear relationships between variables, and are thus unable to capture the nonlinear combinatorial effects of multiple factors. Random effects models fail to address the complex interactions between variables and tend to overlook the synergistic effects of multiple factors. Dynamic panel models typically focus on the dynamic effects of a single factor, making it difficult to simultaneously analyze the combinatorial effects of multiple factors. Traditional QCA considers only spatial factors鈥攏amely, the configurational patterns derived from cross-sectional data at a single point in time鈥攚ithout accounting for temporal factors, thereby limiting its ability to depict the evolutionary patterns of configurations. Compared to commonly used data analysis methods, dynamic QCA can identify the synergistic pathways of multiple factors and reveal the impact of different factor combinations on health levels, rather than merely examining the independent effects of single factors. By integrating the temporal dimension with a configurational perspective, dynamic QCA enables the analysis of the dynamic changes in multi-factor combinations across different time points, providing a more comprehensive causal explanation. Therefore, this study will draw on the relevant theories and methods proposed by Castro et al. [25] and Chang [26]. Using panel data and leveraging the SetMethods package in R 4.2.3, it will employ the dynamic QCA approach to systematically analyze the influencing factors and pathways for improving the health levels of Chinese residents.

Entropy value method

To address the subjectivity and randomness associated with the subjective assignment of weights, the Entropy Value Method is employed to ascertain the weights for evaluating residents' health levels. This method systematically determines the weight of each evaluation index. Among them, the weights of average life expectancy, incidence rate of infectious diseases, perinatal mortality rate and maternal mortality rate determined by entropy value method are 0.483, 0.274, 0.149 and 0.095 respectively.

Numerical calibration

Calibration is a critical step in Qualitative Comparative Analysis (QCA), enabling the transformation of outcome and condition variables into set relations, thereby establishing connections with fuzzy set membership. This process facilitates software recognition and computation. In the calibration process, establishing thresholds for full non-membership, crossover points, and full membership is crucial to ensure consistent data calibration. A fully non-membership point represents the critical value at which a variable completely does not belong to a set, with a membership score of 0. A crossover point represents the critical value at the boundary of the set, with a membership score of 0.5. A fully membership point represents the critical value at which a variable completely belongs to a set, with a membership score of 1. The core of calibration lies in mapping the raw data to the fuzzy set membership range (between 0 and 1) by setting the above three thresholds. Based on the distribution characteristics of the variable, specific values for the fully non-membership point, crossover point, and fully membership point are determined to map the raw data to fuzzy set membership scores. Through the direct calibration method, the raw data are transformed into fuzzy set membership scores using the predefined thresholds. The direct calibration method is adopted, utilizing the 95th percentile, 50th percentile, and 5th percentile as calibration points for full membership, crossover, and full non-membership, respectively. The calibration outcomes for each variable are detailed in Table听4.

Table 4 Variable Calibration

Results

Necessity analysis of individual conditions

Before performing a grouping analysis, it is crucial to evaluate whether each conditional variable acts as a necessary condition for the outcome, which involves examining the consistency level. If a single factor exhibits a consistency level exceeding 0.9, it is considered a necessary condition for the outcome. Within the context of QCA panel data analysis, assessing the magnitude of the adjustment distance is also key to evaluating the accuracy of pooled consistency. A high level of accuracy, indicative of more robust results, is reflected by an adjustment distance below 0.1. Conversely, if the adjustment distance surpasses 0.1, further inquiry into the condition's necessity is warranted.

R software is employed to analyze the necessary conditions for both high and low levels of resident health across 31 provinces in China. Table 5 presents the results of this necessity analysis. Notably, the adjustment distances for economic level, educational level, and ecological environment level are each below 0.1. However, the aggregate consistency level for each individual factor falls below 0.9, with coverage also being less than 1. This indicates that these three factors do not constitute necessary conditions for residents' health levels. Conversely, certain variables, such as population income, the supply of health technicians, and per capita health expenditures, exhibit an adjusted distance greater than 0.1, warranting further investigation into their necessity.

Table 5 Analysis of Necessary Conditions

The between-group consistency and coverage of variables with an adjustment distance exceeding 0.1 are presented in Table听6. In Case I, the level of between-group consistency remains below 0.9 across all years, suggesting an absence of a necessary relationship between high population income and low population health levels. In Case II, while the level of between-group consistency exceeds 0.9 in 2019 and 2020, the between-group coverage is consistently below 0.5 across all years, suggesting no necessary relationship between a high supply of health technicians and a low level of population health. In Case III, the between-group consistency level surpasses 0.9 for the years 2018, 2019, and 2020, and the between-group coverage exceeds 0.5 in 2018. However, further analysis of the scatter plot of between-group consistency reveals that the points are concentrated on the right side, as shown in Fig.听2. Consequently, there is no necessary relationship between high per capita health expenditure and low population health. Furthermore, although the condition variable is not a necessary condition for the outcome variable, there is a significant temporal effect, with its necessity increasing over time. In underdeveloped regions or areas with uneven distribution of medical resources, high per capita healthcare expenditure reflects the substantial costs residents bear to access basic medical services. This phenomenon is increasingly associated with low health levels. The excessive concentration of medical resources in large cities or high-end medical institutions has resulted in insufficient capacity for primary healthcare services. Consequently, residents are compelled to pay higher costs to obtain medical services, yet these expenditures have not significantly improved health outcomes. Therefore, policymakers need to place greater emphasis on the rationality and efficiency of healthcare expenditures, optimize resource allocation, and enhance investment in preventive healthcare services, particularly in underdeveloped regions. This ensures that the growth in healthcare expenditures effectively translates into improvements in residents' health levels.

Table 6 Data between groups with adjusted distances greater than 0.1

Fig.2

Case III between consistency

In summary, necessary conditions for the health level of the population do not exist. This finding implies that independent social, ecological, and medical explanations for the outcome variables are weak. It is essential to further investigate the synergistic driving effects of these three factors and identify combinations of various conditions that influence residents' health levels.

Sufficiency analysis of conditional configurations

In the construction of the truth table, the sample consistency threshold, PRI threshold, and frequency threshold were set at 0.85 [27], 0.7, and 1 respectively, to avoid contradictory configurations. A total of 310 sample cases were gathered for this analysis. Furthermore, the direction of the antecedent variables is not pre-assumed, instead opting for 鈥減resence or absence.鈥� Due to the significant disparities in resource endowments among Chinese provinces, it is not feasible to uniformly assess the effects of antecedent conditions on outcome variables.

In analyzing the results of valid configurations, R software offers three types of solutions: complex, intermediate, and simple. The intermediate solution is primarily used as the reference for adequacy analysis in QCA, with the simple solution serving as a supplement. The nested configuration of both intermediate and simple solutions is utilized. Core conditions are identified as variables present in both the simple and intermediate solutions, whereas marginal conditions are those variables unique to the intermediate solutions. Table 7 presents the configuration analysis results for each province, detailing their impact on the population's health level.

Table 7 Results of the cohort analysis for the realization of a high level of health of the population

Conditional groupings producing high levels of population health

The overall consistency of the intermediate solution for high residents' health level is 0.938, which exceeds the threshold of 0.75. This indicates that 93.8% of the provinces meeting the four conditions exhibit a high level of health. The overall coverage is 0.746, suggesting that 74.6% of cases with high resident health levels can be explained by these four types of conditional configurations. Both the overall consistency and overall coverage of the solutions surpass critical values, demonstrating strong explanatory power and the ability to identify the differentiated adaptive relationships of each condition variable in promoting residents' health levels based on the conditional configuration.

From the standpoint of individual configurations, the consistencies are 0.93, 0.937, 0.965, and 0.947, with coverages of 0.473, 0.449, 0.346, and 0.616, respectively, meeting the QCA analysis consistency standards. The high resident health levels achieved by each configuration are analyzed in detail below. To better highlight distinctions among configurations, the four types can be individually named: economy-ecology-medical synergetic-driven type, economy-education- environment synergetic driven type, economy-environment two wheel drive type, and economy- education-medical synergetic driven type.

These results include four groupings that can be further refined into three models: socio-ecological-medical synergy driven, socio-ecological driven, and socio-medical driven models.

• (1) Socio-ecological-medical synergy-driven model

Configuration 1 indicates that high household income and a high level of ecological environment protection serve as core conditions, while a complementary high supply of health technicians acts as a peripheral condition, leading to high resident health levels. The synergistic effect of high household income and high ecological environment protection as core conditions is reflected in the following: high income provides residents with greater economic support, enabling them to access better living conditions and healthcare services, while a well-protected ecological environment offers a fundamental guarantee for residents' health. Together, these factors lay a solid economic and environmental foundation for improving health levels. The supplementary role of a high supply of health technicians is evident in that, on the basis of high income and strong ecological environment protection, increasing the supply of health technicians significantly enhances the accessibility and quality of medical services, further improving residents' health levels. This pathway explains 47.3% of the cases of high resident health levels. Typical provinces include Zhejiang, Jiangsu, Hunan, and Guangdong.

Taking Zhejiang Province as an example, in recent years, household income in Zhejiang has steadily increased, with per capita disposable income ranking among the highest nationwide. The province has actively implemented the 鈥淭wo Mountains鈥� theory to promote ecological protection and introduced the Zhejiang Atmospheric Pollution Prevention and Control Regulations, effectively improving environmental quality. Additionally, the number of health technicians in Zhejiang increased from 306,900 in 2011 to 548,000 in 2020, ensuring the adequacy of medical resources to support residents' health. The coordinated effects of household income, ecological environment protection, and medical resources have positioned Zhejiang's residents' health levels among the highest in the country.

• (2) Socio-ecologically driven model

Configuration 2 indicates that high economic level, high household income, high educational attainment, and high ecological environment quality serve as core conditions, collectively leading to high resident health levels. The synergistic effects of these core conditions are reflected in the following: improvements in economic levels drive increases in household income, enabling residents to enjoy better living conditions and healthcare services. Higher levels of education enhance residents' health awareness and behaviors, further improving health outcomes. In the context of strong economic and educational conditions, ecological environment protection provides essential external support for residents' health. This pathway explains 44.9% of the cases of high resident health levels, with typical provinces including Guangdong, Henan, Shanxi, Inner Mongolia, Liaoning, and Hunan.

Taking Guangdong Province as an example, Guangdong's leading economic level and household income (with a per capita GDP of 88,521 RMB and a per capita disposable income of 41,029 RMB in 2020) provide ample economic support for residents' health. Furthermore, Guangdong has consistently maintained high ecological environment quality, with air quality predominantly rated as excellent. The improvement in educational attainment has further enhanced residents' health awareness. These factors work together, ensuring that Guangdong's residents maintain a high level of health.

Configuration 3 indicates that high household income and high ecological environment protection serve as core conditions, while non-high economic level, non-high educational attainment, and non-high per capita health expenditure act as peripheral conditions, collectively leading to high resident health levels. The synergistic effect of high household income and high ecological environment protection as core conditions is reflected in the following: despite relatively low economic and educational levels, as well as lower per capita health expenditure, high household income provides residents with a certain level of economic support, while strong ecological environment protection offers external guarantees for residents' health. The limiting role of the peripheral conditions is evident in that non-high economic level, non-high educational attainment, and non-high per capita health expenditure indicate that these provinces face relatively limited social and healthcare resources. However, through the combined effects of household income and ecological environment protection, high health levels can still be achieved. This pathway explains 34.6% of the cases of high resident health levels, with a typical example being Guangxi Province.

Taking Guangxi Province as an example, despite its relatively low economic and educational levels, household income in Guangxi increased significantly from 11,054 RMB in 2011 to 24,562 RMB in 2020, providing economic support for healthier living. At the same time, Guangxi implemented a series of environmental protection regulations, such as the Guangxi Zhuang Autonomous Region Environmental Protection Regulations and the Guangxi Zhuang Autonomous Region Atmospheric Pollution Prevention and Control Regulations, which effectively improved ecological environment quality. The synergy between income growth and ecological protection has enabled Guangxi to achieve relatively high levels of resident health, even with limited social and healthcare resources.

• (3) Socio-medically driven model

Configuration 4 indicates that high per capita GDP, high household income, and high educational attainment serve as core conditions, while high per capita health expenditure acts as a complementary peripheral condition, collectively leading to high resident health levels. The synergistic effect of high per capita GDP, high household income, and high educational attainment as core conditions is reflected in the following: improvements in economic development directly drive increases in household income, providing residents with better living conditions and greater ability to afford healthcare. Generally, regions with higher income levels also tend to have higher educational attainment among residents. Increased education enhances health awareness and promotes the adoption of healthy behaviors. The supplementary role of high per capita health expenditure as a peripheral condition lies in the fact that, in regions with favorable social conditions, higher per capita health expenditure further enhances the accessibility and quality of healthcare resources, thereby improving residents' health levels. This pathway explains 61.6% of the cases of high resident health levels, with 22.2% of provinces exclusively explained by this pathway. Typical provinces include Beijing, Tianjin, Shanghai, Chongqing, and Hebei.

Taking Beijing as an example: Beijing鈥檚 advanced economic conditions (with a per capita GDP of 164,158 RMB in 2020) and high household income (with a per capita disposable income of 69,434 RMB in 2020) provide robust economic support for residents' health. As the national center for education and healthcare resources, Beijing is home to numerous prestigious universities, and its residents' educational attainment ranks among the highest in the country. The city鈥檚 developed economy lays the foundation for high per capita health expenditure, enabling residents to access high-quality healthcare resources. The synergy between social and healthcare conditions has ensured that Beijing鈥檚 residents maintain some of the highest health levels nationwide.

Conditional groupings producing low levels of population health

There are two configurations leading to low population health levels, both of which share the same core conditions, as shown in Table听3. By examining the raw coverage and unique coverage values for configurations F₁ and F₂, it is evident that both indicators are higher for configuration F₁ than for F₂. This indicates that configuration F₁ has the strongest empirical relevance among the two configurations. The overall consistency of the two configurations is 0.947, which is significantly higher than the acceptable consistency threshold, suggesting that the combinations of conditions above have strong explanatory power for the outcome variable. Meanwhile, the overall coverage is approximately 0.6, with the raw coverage of the individual solutions being 0.55 and 0.393, respectively. This indicates that both configurations explain a certain proportion of the cases of low population health levels, and together, they account for a substantial portion of the regional low population health outcomes.

Configuration F₁ indicates that low economic development, low household income, low levels of education, weak ecological protection, and a low number of health technicians per 1,000 people lead to low population health levels. Among these, low economic development, low household income, low levels of education, and weak ecological protection are core conditions, while a low number of health technicians per 1,000 people is a peripheral condition. This pathway explains 55% of the cases of low population health levels, with 20.8% of such cases being uniquely explained by this configuration. Typical provinces and cities include Ningxia, Hainan, Guizhou, and Yunnan. Taking Ningxia as an example, the region's low level of economic development, with per capita GDP and household income ranking among the lowest in the country, limits residents' access to health resources. The low level of education among residents results in weak health awareness and behaviors, further exacerbating health issues. In addition, the fragile ecological environment in some areas, characterized by water scarcity and desertification, poses significant challenges to public health. The limited number of health technicians and the uneven distribution of medical resources further weaken the capacity of primary healthcare services.

Configuration F₂ differs from F₁ in that high per capita healthcare expenditure is a peripheral condition contributing to low population health levels, while the core conditions remain the same. This pathway explains 39.3% of the cases of low population health levels, with 5.1% of such cases being uniquely explained by this configuration. In summary, it can be observed that a lack of social, ecological, and medical conditions collectively leads to low population health levels. Typical provinces and cities include Qinghai, Gansu, Tibet, and Xinjiang. Using Qinghai as an example, the region's limited economic development and low household income restrict residents' access to healthcare resources. The low level of education among residents results in weak health awareness. Qinghai's fragile ecological environment, with issues such as desertification and water scarcity in some areas, further exacerbates health challenges. Although per capita healthcare expenditure is relatively high, the uneven distribution of medical resources and insufficient healthcare capacity in remote areas hinder improvements in population health.

Inter-group analysis

The adjusted inter-group consistency distances for the four configurations are all less than 0.1, indicating that there is no significant temporal effect among these configurations. Further examination of the temporal variation in consistency for the four configurations reveals that, from 2011 to 2020, the consistency levels fluctuated between 0.9 and 1. This suggests that the findings of this study provide strong explanatory power for the factors influencing residents' health levels. Over the time dimension, it was observed that the consistency of Configuration 1 and Configuration 3 collectively declined in 2020. During the pandemic, other factors鈥攕uch as epidemic prevention measures and the emergency allocation of medical resources鈥攖emporarily became core determinants of residents' health levels, thereby weakening the explanatory power of the original configurations based on social, ecological, and medical factors.

The decline in Configuration 1 can be attributed to the instability in the supply of healthcare personnel caused by the emergency allocation of medical resources during the pandemic, which weakened the explanatory power of this configuration. Additionally, lockdown measures negatively impacted household income and economic activity, further diminishing the contribution of social factors to health outcomes. The decline in Configuration 3, on the other hand, is likely due to the challenges faced by provinces with lower economic and educational levels in responding to the pandemic. While these provinces maintained relatively high levels of ecological environment protection, the prioritization of epidemic prevention efforts may have undermined the positive impact of ecological factors on health outcomes.

Furthermore, the varying measures adopted by provinces to address the COVID-19 pandemic likely exacerbated the decline in inter-group consistency. For instance, Hubei Province, as the epicenter of the outbreak, experienced severe disruptions to its medical resources and socioeconomic activities, which may have caused the drivers of health levels in Hubei to diverge significantly from those in other provinces. Some provinces, such as Beijing and Shanghai, implemented strict lockdown measures, while others adopted relatively lenient policies. These policy differences influenced the factors driving residents' health levels. Strict lockdowns may have amplified the role of medical resources while diminishing the contributions of social and ecological factors. Meanwhile, economically developed provinces like Guangdong and Zhejiang demonstrated greater capacity to respond to the pandemic compared to resource-constrained provinces like Tibet and Qinghai, leading to regional disparities in the pandemic's impact on residents' health levels.

These findings suggest that public health emergencies can alter the pathways driving health outcomes, amplifying or diminishing the effects of certain factors. The changes in inter-group consistency are illustrated in Fig.听3.

Fig.3

Changes in BECONS

Intra-group analysis

The within-group consistency of Configuration 1 is 0.930, with six provinces falling below this threshold: Inner Mongolia, Liaoning, Guizhou, Yunnan, Qinghai, and Xinjiang. This indicates that the explanatory power of Configuration 1 is relatively weak in these regions, as the synergistic effects of social, ecological, and medical factors have not been fully realized. The within-group consistency of Configuration 2 is 0.931, with four provinces falling below this threshold: Inner Mongolia, Liaoning, Qinghai, and Xinjiang. This suggests that Configuration 2 is less applicable in these regions, possibly due to the lack of effective synergy between social and ecological conditions. The within-group consistency of Configuration 3 is 0.937, with six provinces falling below this threshold: Inner Mongolia, Guizhou, Yunnan, Tibet, Qinghai, and Xinjiang. This indicates that the explanatory power of Configuration 3 is relatively weak in these regions. Configuration 4 has a within-group consistency of 0.945, with four provinces falling below this threshold: Inner Mongolia, Liaoning, Qinghai, and Xinjiang. This suggests that Configuration 4 is less effective in these regions, potentially due to the lack of effective synergy between social and medical conditions. For Configuration F₁, the within-group consistency is 0.955, with five provinces falling below this threshold: Anhui, Jiangxi, Shandong, Guangxi, and Hainan. The relatively low consistency values in these provinces indicate that the explanatory power of Configuration F₁ is weaker in these areas. Similarly, Configuration F₂ has a within-group consistency of 0.947, with six provinces falling below this threshold: Anhui, Jiangxi, Shandong, Guangxi, Hainan, and Ningxia. The low consistency values in these provinces suggest that Configuration F₂ is less effective in explaining the health outcomes in these regions.

Furthermore, it can be observed that the provinces with low within-group consistency in the four configurations leading to high levels of residents' health can all be adequately explained by the two configurations associated with low levels of residents' health. This indicates a complementary relationship between the two sets of configurations: the deficiencies in health outcomes that cannot be fully explained by the high-health configurations can be supplemented by the factors identified in the low-health configurations. The within-group consistency results are presented in Table听8 and Figs.听4, 5, 6, 7, 8 and 9.

Table 8 WICONS of provinces

Fig.4

WICONS for configuration 1

Fig.5

WICONS for configuration 2

Fig.6

WICONS for configuration 3

Fig.7

WICONS for configuration 4

Fig.8

WICONS for configuration F₁

Fig.9

WICONS for configuration F₂

Robustness tests

During the robustness testing process, the configurations leading to high levels of resident health were subjected to robustness checks by adjusting the consistency threshold and enhancing the Proportional Reduction in Inconsistency (PRI) consistency. The consistency threshold was adjusted from 0.85 to 0.9, and the PRI was increased from 0.7 to 0.75, while maintaining the frequency threshold at 1. The results indicated that the newly derived configuration paths were largely consistent with the existing configurations. Therefore, it can be concluded that the analysis results exhibit strong robustness.

Discussion

Through a configurational perspective, the analysis complements Grossman's health capital model in examining the determinants of population health levels. The Grossman model conceptualizes health as a stock of capital, emphasizing that individuals improve their health through investments such as healthcare expenditures and lifestyle choices. However, utilizing a dynamic QCA approach, the analysis reveals that health outcomes result from the combined effects of multiple external factors, thereby supplementing the health capital model's assumption of linear relationships between single variables. This research not only extends the applicability of the health capital theory but also provides a more comprehensive theoretical framework for understanding the complex mechanisms determining health outcomes, offering significant theoretical and practical implications.

Theoretical contribution

Building on existing research in the field of public health, an analytical framework is proposed to identify the factors affecting residents' health levels. Unlike previous studies, which often employ a singular or specialized classification method, this framework integrates social, ecological, and medical factors organically. By combining this with the theory of health human capital, the framework is expanded and refined to include six secondary conditions influencing health levels, forming the basis for qualitatively comparing configurational relationships. This research not only examines the impact of social factors (such as economic level, income, and education) on health levels but also explores the roles of ecological and medical factors (such as environmental protection, healthcare personnel supply, and per capita health expenditures). The findings reveal a joint impact of these three types of factors on residents' health levels, broadening research on external influences on health and providing support for the health human capital theory regarding the joint action of social, ecological, and medical factors.

This research introduces a configurational perspective into the research of health human capital theory, broadening the research lens on residents' health levels. Previous literature often relies on regression analysis to explore factors affecting health levels, typically focusing on individual factors like income or air pollution, while neglecting the interplay between multiple factors. This research, from a configurational perspective of health human capital theory, simultaneously considers six different systemic factors鈥攅conomic level, income, education, environmental protection, healthcare personnel supply, and per capita health expenditures鈥攁nd their combined impact on health levels. This exploration not only offers a valuable examination of the synergistic effects of these factors but also paves new avenues for future research in this field.

The application of the dynamic Qualitative Comparative Analysis (QCA) methodology in studying residents' health levels allows exploration of configurational effects over a longitudinal temporal dimension, thereby enhancing the precision of the research. Previous studies have shown limitations in using the QCA method, particularly in addressing temporal dimensions. By employing a dynamic QCA approach, temporality is incorporated into configurational research, utilizing panel data from 2011 to 2020 to analyze the temporal drivers of health levels among Chinese residents.

Practical implications

The research findings of this study propose the following policy recommendations for improving residents' health levels across different provinces in China:

• 鈥淪ocio-Ecological-Medical Synergy Driven鈥� Pathway Explanation: Through the positive coupling of social support, ecological protection, and medical supply, residents' health levels can be effectively improved. Therefore, in less developed western provinces, efforts should focus on strengthening social support, enhancing ecological protection, and increasing medical service supply to promote residents' health. This includes ensuring sustained and stable high-quality economic development to increase residents' income, strengthening ecological governance to improve regional environmental quality, and increasing investment in medical resources to ensure accessibility to healthcare services.

• 鈥淪ocio-Ecological Drive Type鈥� Pathway Explanation: When medical resource supply is insufficient, relying on the dual drivers of social and ecological factors can result in high levels of residents' health. Therefore, in central provinces with relatively low medical supply, efforts should focus on strengthening ecological protection to compensate for the lack of medical resources, achieving a positive coupling of social and ecological protection to jointly improve residents' health. This involves enhancing environmental protection efforts, comprehensively addressing urban and rural sanitation, developing sustainable and renewable energy sources [28], and strengthening ecological protection. Special attention should be given to monitoring and evaluating environmental factors closely related to public health, such as drinking water, food safety, air pollution (e.g., smog), and soil quality. Establishing systems for environmental and health surveys, monitoring, and risk assessment, and maintaining the 鈥渞ed line鈥� of a healthy environment are primary tasks for building a health-supportive environment, which can drive improvements in residents' health levels.

• 鈥淪ocio-Medically Driven鈥� Pathway Explanation: The positive coupling of social support and medical supply can enhance residents' health levels. Therefore, in the more developed eastern regions, to compensate for relatively weaker ecological protection, medical resource supply can be increased to promote residents' health through the collaborative effects of social support. This includes improving the efficiency of medical resource utilization, strengthening resource-constrained medical investments, optimizing health service processes, and fully leveraging the role of internet hospitals. Building comprehensive medical cloud services, streamlining online medical insurance payment systems, and improving the efficiency of medical supply utilization can further enhance residents' health levels. Moreover, research indicates that the development of digital finance can significantly enhance both the overall and mental health of residents [29], while simultaneously driving socio-economic development and providing critical support for public health. Recent studies further highlight that advancements in online network infrastructure can substantially benefit population health [30, 31].

Research limitations and future directions

The limitations of this study are primarily threefold: First, due to constraints in data availability, this research utilizes publicly available secondary data and employs objective evaluation metrics to assess residents' health levels. Future research should integrate subjective health indicators to further explore the mechanisms influencing health levels, which remains an area for improvement. Second, the study examines the factors affecting health levels at a macro level across provinces. Future studies could delve deeper into micro-level factors influencing health. Third, while this research focuses on the impact of social, ecological, and medical factors on health, these are external influences. It does not account for internal individual factors such as mental health, alcohol consumption, and obesity. Internal individual factors also play a crucial role in health outcomes. Mental health issues, such as depression and anxiety, may directly or indirectly affect an individual's physical health and utilization of healthcare services, while unhealthy lifestyles, such as smoking, excessive alcohol consumption, and lack of physical activity, are significant risk factors for many chronic diseases. These internal factors not only interact with external factors but may also, to some extent, moderate the impact of external factors on health outcomes. Future research could analyze how both internal and external factors jointly affect health levels across different provinces.

Conclusion

In the current era of rapid economic development, the importance of residents' health has become increasingly prominent. It is not only a crucial component of human capital but also a fundamental condition for economic development and an essential requirement for comprehensive social advancement. Social, ecological, and medical factors all play pivotal roles in enhancing people's health. Effectively coordinating these factors and promoting the synergy of various institutional elements is crucial for enhancing residents' health. Grounded in the theory of healthy human capital and employing both the entropy method and dynamic Qualitative Comparative Analysis (QCA), the research aims to measure the health level of residents across Chinese provinces and analyze the synergistic relationships between multiple factors from a group perspective. The following conclusions are drawn:

(1) Utilizing the dynamic QCA method, the study reveals that no individual factor is essential for high population health. However, population income consistently appears in all groupings associated with high health outcomes. Moreover, the necessity of per capita health expenditure has shown an increasing trend over the years, indicating a significant temporal effect. (2) In the sufficiency analysis of condition combinations, the study identifies factor combinations that enhance high population health levels. These combinations fall into four main paths, which can be categorized into three models. First, the 鈥渟ocial-ecological-medical synergy-driven鈥� model leverages the combined effects of these factors to elevate residents' health. Second, the 鈥渟ocial-ecological-driven鈥� model relies on a multi-factor combination when medical conditions are underdeveloped, using social and ecological factors to maintain high health levels. Third, the 鈥渟ocial-medical driven鈥� model emphasizes the synergistic enhancement of population health through the integration of social and medical care. These models reflect diverse approaches to achieving high health standards across different Chinese provinces, given their varying social, ecological, and medical contexts. (3) Although the overall consistency did not exhibit a temporal effect, the consistency of Grouping State 1 and Grouping State 3 collectively declined in 2020. This may be attributed to the outbreak of the COVID-19 pandemic, where certain conditions temporarily emerged as central factors in health promotion, thereby diminishing the explanatory power of other factors. (4) The relatively weak explanatory power for provinces falling below the Within Consistency threshold across group states may be attributed to other influencing factors affecting health levels in these regions, which are marked by lower economic development. These factors warrant further in-depth investigation.