Necessity of Analyzing the Korea Community Health Survey Using 7 Local Government Types
Article information
Abstract
Objectives
This study examined the potential of a new analytical framework for the Korea Community Health Survey (KCHS) with classification criteria for the sub-national governmental level, the degree of urbanization including an urban-rural multimodal category, and population size as a more effective tool to address local health problems and deduce practical implications.
Methods
Retrospective survey data from 2023 KCHS were obtained. Frequency analyses were performed for self-rated good health status, current smoking status, and unmet medical needs as proxies for health status, as well as health behavior and healthcare utilization, utilizing individual weights to represent national community residents.
Results
We established a new classification of local governments into 7 types to facilitate regional comparisons. These local government types are presumably composed of populations showing statistically significant differences in demographic characteristics. There were evident differences in health status, health behavior, and healthcare utilization in comparisons of groups categorized by local government types.
Conclusions
This study suggests that regional disparities can be analyzed using a new typology of local governments. This practically effective approach could be used in decision-making for community-centered health projects in terms of community health planning. Future research should conduct analyses of KCHS data that use these 7 local government types to comprehensively reflect regional characteristics.
INTRODUCTION
The Korea Community Health Survey (KCHS) plays an essential role in the evaluation and performance monitoring of community health and medical care by providing valuable insights for policy-making and healthcare delivery. Unlike other national-level statistics, KCHS is a nationwide survey designed to produce regionally representative health statistics in Korea [1–3]. By offering regional and even sub-regional data, KCHS has enabled the identification of regional disparities in health outcomes and the implementation of targeted interventions. In the domain of public health, the importance of local data collection has been underscored in the context of devolving powers, responsibilities, and resources from the central government to subnational governments [4,5]. As the decentralization of healthcare governance gains political momentum in Korea [4], the significance of KCHS has increasingly come to the forefront.
The KCHS has been employed in various capacities, alongside societal efforts to reduce regional disparities in health outcomes. By legal delegation, the government has facilitated the implementation of community health promotion projects by local governments through health centers and health and medical clinics [2]. Serving as a governance entity for public health promotion in Korea [6–8], community health and medical institutions have used KCHS results to assess and manage public health interventions. Pilot projects adopting a small-area approach, which focuses on subdivided areas of districts under local governments, have utilized KCHS results as preliminary basic data [9,10]. Additionally, there has been an increase in academic research exploring the health status, health behaviors, and medical utilization of local residents, drawing on the findings from KCHS [11–15].
The current analytic frameworks for examining regional differences in health have several limitations. Traditional analyses based on urbanization levels do not capture the complexities of reality, which extend beyond the simple urban-rural dichotomy [16–19]. Studies conducted in 17 districts under the jurisdiction of regional governments do not consider variations within the districts themselves. Furthermore, analyzing geographical differences across 258 health centers and health and medical clinics presents challenges in generalizability, such as formulating policy recommendations, due to the overly specific nature of the data [20]. The necessity for differentiated evaluation has been previously discussed, taking into account regional variations in workforce, budget, environment, and scale [6,7].
In response to these challenges, this study proposes a new analytical framework that incorporates the diverse characteristics of local governments associated with health centers and health and medical clinics. This framework is designed to enhance our understanding of geographical variations in health services [16–19]. It categorizes local governments into 7 types, expanding the classification criteria to include sub-national levels, degrees of urbanization with an emphasis on an urban-rural multimodal category, and population size [6]. The study then explores whether there is significant variability in health status, behaviors, and healthcare utilization among residents based on the type of local government. The aim is to assess the effectiveness of this new analytical framework within the KCHS, providing a comprehensive rationale and identifying practical implications for addressing local health issues.
METHODS
Typology of Local Governments
Korea is a unitary state with a sub-national governmental system organized into 2 levels, as defined by the Local Autonomy Act [21,22]. As of 2023, there are 17 regional governments, including metropolitan cities and provinces, which constitute the higher tier [22,23]. Additionally, there are 226 local governments that make up the lower tier of sub-national governments [22,23]. The areas under the jurisdiction of local governments with autonomy are categorized by 1 of 3 designations. Si is the official term for districts within a city, gun refers to regions within a county, and gu denotes the autonomous districts (boroughs) in metropolitan cities [21]. Health centers and health and medical clinics, which are responsible for local healthcare and have specific legal obligations, have been established at the local government level [2]. Consequently, public healthcare governance in Korea is structured across 3 levels: the central government, the regional government, and the local government [6,7].
Local governments, which oversee community health and medical institutions, vary in terms of budgets, resources, infrastructure, and population structures [6,7]. Over the past few decades, as decentralization in Korea has evolved, it has led to asymmetrical decentralization in the political, administrative, and fiscal domains. Consequently, this has resulted in varying health impacts across communities [20,24]. A previous study [6] categorized local governments into 7 types, taking into account the healthcare governance system, industrial characteristics including urbanity, and the size of the population in each region.
Table 1 presents the classification of local governments. Types 1 and 2 represent autonomous districts in Seoul Metropolitan City and other metropolitan cities in Korea, respectively. This classification highlights the spatial concentration in Seoul Metropolitan City compared to districts outside Seoul, which contributes to regional inequality [25]. Types 3 and 4 categorize cities based on population size, acknowledging the right of cities with populations exceeding 500 000 to establish non-autonomous districts, similar to those in Seoul Metropolitan City and other metropolitan cities [21]. Type 5 pertains to urban-rural consolidated cities, a concept introduced in 1994 in response to the ongoing population decline in rural areas and the discrepancies between living and administrative jurisdictions [18,26]. Types 6 and 7 refer to rural areas, also segmented by population size as per the Local Autonomy Act [21]. Rural areas, known as gun, with populations over 50 000, may be upgraded to si (city) status under certain conditions [21]. This subdivision of rural areas also corresponds with the historical context of establishing health and medical clinics [27].
Classification of local governments into 7 types
Supplemental Material 1 presents the types of local governments along with the names of 258 health centers and health and medical clinics as of 2023, following the matching process. This matching was conducted using the administrative district information from the Ministry of the Interior and Safety and the 2023 population census data provided by Statistics Korea from the Korean Statistical Information Service [23,28].
Data Source and Study Population
The study utilized data from the 2023 KCHS, which investigated 145 items across 17 health-related domains through 258 community health centers [1,29]. The computer-assisted personal interviewing method was employed, involving researchers visiting households to conduct surveys with an electronic questionnaire on a tablet PC. This approach helped minimize errors due to omitted data and enhanced quality management [1,3,29]. Given that the KCHS sample was derived from a complex sampling design, rather than simple random sampling, it is essential to consider weights when estimating means and variances [8]. Consequently, all analyses in the study were conducted using values that accounted for individual weights, thereby accurately representing the community residents [8]. Of the total 231 752 participants in the 2023 KCHS, the final study population consisted of 228 399 participants after the exclusion of missing values, with a weighted total of 42 851 705 participants.
Measures
Dependent variables
The self-reported questionnaires from the KCHS focus on health outcomes, which include health status, health behaviors, and healthcare utilization. This study incorporated 3 specific variables as dependent variables: self-rated good health from the health status domain; current smoking status from the health behavior domain; and unmet medical needs from the healthcare utilization domain. The proportion of self-rated good health was defined as the percentage of participants who responded with “good” or “very good.” The proportion of current smoking status represented the percentage of participants who currently smoked any type of cigarette, including regular cigarettes, cigarette-type electronic cigarettes, or liquid-type electronic cigarettes. Lastly, the proportion of unmet medical needs indicated the percentage of respondents who needed healthcare services but were unable to visit healthcare institutions for various reasons.
General characteristics of study population
Sex, age group, marital status, monthly household income, educational status, and job were selected to examine the basic characteristics of the study population as a whole and of groups categorized by specific local government types. The age group was divided into 5 categories: 19–29, 30–39, 40–49, 50–59, and over 60. Marital status was initially categorized into 5 groups: married, divorced, widowed, separated, and single. However, for the purposes of this study, we simplified marital status into 2 categories: unmarried and married. Monthly household income was segmented into 5 brackets based on the classification criteria from Statistics Korea: <1, ≥1 and <3, ≥3 and <4, ≥4 and <6, and ≥6 million Korean won. Educational status was categorized into 3 levels: high school graduate or lower, college/university graduate or associate degree, and master’s degree or higher. Job categories were defined as follows: white-collar (including managers, professionals, and clerks), sales and service worker, general technician (including skilled agricultural, forestry, and fishery workers, craft and related trades workers, and plant and machinery operators and assemblers), elementary occupation, and other.
Statistical Analysis
Descriptive analyses were conducted to examine the characteristics of the study population when divided into 7 local government types. Analysis of variance (ANOVA) tests were conducted to compare differences among residents. To evaluate regional disparities in health status, health behavior, and healthcare utilization across different types of local governments, frequency analyses with weighted data were performed, and the results were displayed using box-and-whisker plots. The midline of a box plot indicates the median. The endpoints of a box plot represent the 25th percentiles and 75th percentiles, and the width of the box plot indicates the interquartile range (IQR). Whisker lines in each box plot extend to the most extreme values that are less than or equal to 1.5 times the IQR. Average values and outliers beyond the whiskers are shown with additional markers [30]. Statistical analyses were performed using SAS Enterprise Guide 7.4 (SAS Institute Inc., Cary, NC, USA). Statistical significance was set at p-value<0.05.
Ethics Statement
This study was approved by the Seoul National University College of Medicine/Seoul National University Hospital Institutional Review Board (IRB No. E-2404-003-1524). The requirement for informed consent was waived owing to the use of anonymized data.
RESULTS
General Characteristics of Study Population
Table 2 presents the characteristics of study subjects as a whole and divided into 7 groups based on the type of local government. The results of ANOVA tests on all basic characteristics of the study population indicate significant differences among residents of each local government type. According to these results, the male-female ratio exceeds 1 in local governments classified as counties in rural areas (types 6 and 7), which is not the case in other types of local governments. Notably, more than half of the residents in counties with populations under 50 000 (type 7) are presumed to be over 60 years old. The proportion of inhabitants in the lowest monthly household income bracket is smallest in cities with populations of 500 000 or more (type 3). Meanwhile, over 60% of residents in Seoul Metropolitan City (type 1) are believed to have attained an educational level beyond high school graduation. In contrast, more than 70% of individuals in counties with populations under 50 000 (type 7) are expected to have completed high school or less. The highest percentage of unmarried individuals is found among residents of Seoul Metropolitan City (type 1). Among the 7 local government types, Seoul Metropolitan City (type 1) reportedly has the highest proportion of people employed in white-collar jobs. Conversely, counties with populations under 50 000 (type 7) are expected to have the highest proportion of individuals in elementary occupations.
Characteristics of the study population across local government types in Korea
Comparison of Health Status, Health Behavior, and Healthcare Utilization
Table 3 presents summary statistics for the distribution of health status (self-rated good health), health behavior (current smoking status), and healthcare utilization (unmet medical needs), categorized by type of local government. Figure 1 displays the distribution of each variable, including outliers, using box-and-whisker plots. The mean and proportion of residents reporting self-rated good health show significant variation. Although the mean and average proportions for current smoking status remain under 20 percent across different types of local governments, the spread and skewness of these distributions vary. Among all dependent variables, the distribution of unmet medical needs is particularly noteworthy for its box width and the length of its whisker lines. Detailed information on each dependent variable for each health center and health and medical clinic is provided in Supplemental Material 2.
Summary statistics for the distribution of health status, health behavior, and healthcare utilization grouped by local government type
Distribution of health status, health behavior, and healthcare utilization categorized by local government type.
DISCUSSION
This study aims to discuss the necessity of analyzing geographical discrepancies in health-related statistics, using KCHS as a foundational data source for regional health projects. We introduce a new classification system that divides local governments into 7 distinct types for regional comparison. Each type of local government encompasses populations with statistically significant demographic differences. There are clear disparities in terms of health status, health behavior, and healthcare utilization among the comparative groups categorized by these local government types.
In this study, we took into account the characteristics of local governments when analyzing regional variations in health-related variables, a consideration that previous studies did not include [11–15]. Our new framework for regional comparison offers several distinct advantages. Firstly, it is designed with the administrative levels of the sub-national government system in Korea in mind. Additionally, it accounts for urban-rural differences and goes beyond simple dichotomies by introducing a classification known as “urban-rural consolidated city (type 5).” Moreover, incorporating a supplementary division based on population size reduces heterogeneity within the same type of local government.
Our analysis revealed significant discrepancies in population structure and health-related indicators across different types of local governments, confirming that these differences are not merely hypothetical. Variations in health outcomes can be attributed to individual-level attributes within local areas [20]. However, it is crucial not to overlook the contextual effects that contribute to these varying health outcomes [20,24]. Additionally, our findings highlight a considerable imbalance in health-related indicators within the same type of local government. This variation may stem from the persistence of asymmetric decentralization, despite efforts to achieve horizontal balance [4]. By identifying outliers within and between local government types, we can better determine the need for political, administrative, and fiscal support to address these spatial disparities in health.
This study presents a cross-sectional analysis based on the 2023 KCHS results, highlighting regional variations in health status, health behavior, and healthcare utilization. It should be noted that this analysis merely reflects the current situation and does not establish any causal relationships. Further research is necessary to understand the mechanisms behind spatial variations, potentially through qualitative analysis and interviews with stakeholders. Additionally, the study aims to identify the mere presence of regional variation across different types of local governments. Future research could focus on defining longitudinal trends in local health statistics within these government types. A limitation of this study is its narrow focus on only 3 dependent variables, which may hinder a comprehensive understanding of broader health-related regional issues. Therefore, the development of a single health-related indicator that encompasses comprehensive information is crucial. This need for a comprehensive indicator was highlighted in previous studies [3,4], emphasizing the importance of continuous monitoring to assess regional efforts effectively.
The significance of this study is underscored by its academic and policy implications. By introducing a new typology of local governments, the analysis deepens our understanding of the governance system’s capacity and condition. It presents a fresh perspective on reality, bridging the knowledge gap between national-level and small-area approaches. The understanding of local health-related issues could be enhanced by incorporating the KCHS with a multifaceted consideration of regional contexts. Consequently, policymakers and stakeholders could develop a more intuitive understanding, leading to more effective interventions and a reduction in regional health disparities.
KCHS is carried out by the Korean government under the Local Health Act to assess the health status and issues of its residents. This study proposes analyzing regional disparities using a new typology of local governments, an approach that is practically effective for incorporating into decision-making for community-centered health projects, particularly in community health planning. An analysis using KCHS should take into account regional characteristics comprehensively across 7 types of local governments.
Supplemental Materials
Supplemental materials are available at https://doi.org/10.3961/jpmph.24.388 .
Notes
Conflict of Interest
The authors have no conflicts of interest associated with the material presented in this paper.
Funding
None.
Author Contributions
Conceptualization: Jung H, Lee JY. Data curation: Bai H, Lee JR. Formal analysis: Bai H, Lee JR. Funding acquisition: None. Methodology: Park S, Bai H, Lee JR. Visualization: Park S. Writing – original draft: Park S, Bai H, Lee JR. Writing – review & editing: Kim S, Jung H, Lee JY.
Acknowledgements
None.