Spatial health inequities among older adults in Thailand: a composite vulnerability index and relationship with local development

Abstract

Background

Thailand is undergoing a rapid demographic transition toward an aging society, with substantial challenges in ensuring health equity for older adults. While national-level health policies have expanded coverage and services, spatial disparities in health risks and local development remain underexplored, particularly through integrative spatial frameworks.

Objective

This study aimed to examine the spatial distribution of health vulnerability across Thai provinces and its spatial association with local development.

Methods

Data were obtained from the 2024 nationwide community-based screening program for older adults in Thailand, covering nine health risk indicators: cognitive impairment, mobility limitation, visual impairment, hearing impairment, urinary incontinence, activities of daily living limitation, oral health problems, malnutrition risk, depression risk. A composite health vulnerability index (CHVI) was constructed using principal component analysis, with the first component explaining 74.83% of total variance. Spatial clustering was assessed using Global Moran’s I and local indicators of spatial association (LISA), while bivariate LISA examined the spatial relationship between CHVI and local development, proxied by nighttime light (NTL) intensity derived from satellite imagery.

Results

The CHVI demonstrated strong positive loadings across all nine geriatric risk indicators. Global Moran’s I was 0.261 (p = 0.002), indicating significant spatial autocorrelation. LISA analysis revealed high–high clusters of CHVI predominantly in the Northeastern region, while NTL intensity was concentrated in the Central region. The negative bivariate Moran’s I (− 0.174, p < 0.05) indicated overall spatial dissimilarity between health vulnerability and local development. Notably, nine central provinces surrounding Bangkok exhibited high CHVI values spatially associated with lower neighboring NTL intensity.

Conclusion

The CHVI provides a spatially explicit measure for identifying provincial-level geriatric health risks. By examining spatial associations between health vulnerability and local development, this approach offers evidence to inform geographically targeted strategies aimed at promoting healthy aging and addressing regional inequities in Thailand.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-026-27051-7.

Introduction

Thailand is facing a rapidly aging population, with the proportion of older adults projected to exceed 30% by 2040, making it one of the fastest-aging societies in Southeast Asia [1, 2]. This demographic transition presents significant challenges for health systems, social protection, and regional development [3–5]. While aging is a global phenomenon, its impacts are deeply place-dependent. Spatial disparities in health outcomes and access to services among older adults may be associated with local development inequalities, particularly in low- and middle-income countries such as Thailand [6–8].

In Thailand, efforts to promote healthy aging have largely focused on individual behaviors and household-level care provision. However, growing evidence highlights that structural and spatial determinants including infrastructure, health system capacity, economic development, and urbanization are closely associated with variations in health risks among older adults [8–10]. Recent spatial epidemiological studies in Thailand have demonstrated that nighttime light (NTL) intensity serves as a robust proxy for urban growth and socioeconomic activity. Associations between NTL and both infectious and chronic disease patterns have been documented [11, 12], underscoring the structural role of development in shaping spatial health risks. This is particularly concerning in rural and remote provinces where access to geriatric services, transportation support, and nutritious food remains limited. Recent studies have emphasized the need for spatially explicit assessments of health vulnerabilities to inform equitable policy responses [13–15].

Despite the nationwide expansion of Thailand’s universal health coverage, spatial disparities in health outcomes and access to care persist across the country [7, 16]. Older adults living in resource-limited areas often face higher burdens of chronic and preventable conditions. These burdens are associated with structural variations in health infrastructure, economic capacity, service availability, physical accessibility, and institutional readiness—factors that collectively contribute to spatial inequities in health vulnerability across provinces. Recent decentralization reforms have created workforce challenges in local health systems [17], while digital access has emerged as a critical determinant of healthcare utilization among older adults [18]. These structural and technological disparities may amplify vulnerability in less developed provinces. Nevertheless, most existing literature has focused on individual- or household-level determinants, with limited application of spatial analytical approaches to examine place-based inequities in multidimensional health risk and development. In particular, the spatial relationship between local development and geriatric vulnerability has rarely been quantified at the provincial level [19, 20]. This represents a critical gap in evidence needed for spatially informed health and development policies in Thailand’s rapidly aging society.

This study addresses the need for spatially informed approaches to health equity among older adults by constructing a composite health vulnerability index (CHVI) through principal component analysis (PCA) of nine geriatric health risk indicators derived from Thailand’s national screening program. By integrating these indicators with satellite-based NTL intensity as a proxy for local development, the study applies spatial statistics to examine geographic clustering and misalignments between health vulnerabilities and local development levels. As an ecological study, the findings are interpreted as spatial associations rather than causal relationships and are intended to inform geographically targeted strategies aimed at promoting healthy aging and reducing regional health inequities across Thailand.

Materials and methods

Study design and setting

This ecological study employed a cross-sectional spatial analysis using provincial-level data from all 77 provinces of Thailand. The unit of analysis was the province, and the spatial relationships were analyzed using both global and local spatial statistics.

Data sources and variables

Health risk indicators

Data on nine geriatric health risk indicators were derived from the Annual Community-based Screening for Older Adults year 2024, conducted nationwide through a collaborative effort between the Ministry of Public Health and local health volunteers. Assessments were performed by village health volunteers (VHVs) using the digital application “Smart Oa So Mo” (Smart VHV), designed to standardize data collection across Thailand’s 77 provinces. Each indicator was operationalized as the proportion of affected older adults within each province and aggregated at the provincial level. The compiled screening data are maintained and managed under the Department of Health Service Support, Ministry of Public Health [21]. The nine indicators included in the analysis are listed in Table 1. These indicators reflect key domains of multidimensional geriatric vulnerability routinely assessed in Thailand’s national screening program. Rather than representing morbidity alone, they encompass cognitive function, mobility, sensory capacity, nutritional status, mental health, and functional independence. Collectively, they align with established frailty and health vulnerability frameworks, which conceptualize vulnerability as the accumulation of functional deficits across multiple physiological and psychosocial domains [22].

Table 1.

Operational definitions and screening methods of health risk indicators among older adults

No.	Health risk indicator	Screening tool / Question	Operational definitions of risk
1	Cognitive Impairment	Mini-Cog: 3-word registration, Clock Drawing, and 3-word recall	Total score ≤ 3 out of 5
2	Mobility Limitation	Timed Up and Go Test + Fall history	Time > 12 s or fall within past 6 months
3	Visual Impairment	Vision screening for daily activity (distance vision)	Inability to clearly see distant objects affecting daily function
4	Hearing Impairment	Finger Rub Test near each ear	Unable to hear sound in one or both ears
5	Urinary Incontinence	“Have you experienced involuntary urine leakage interfering with daily life?”	Answer = “Yes”
6	Activities of daily living (ADL) Limitation	“Has your ability to perform daily activities independently, without needing assistance, declined?”	Answer = “Yes”
7	Oral Health Problems	“Do you have difficulty chewing hard food?” and “Do you have oral pain?”	“Yes” to either = oral health problem
8	Malnutrition Risk	“Have you lost >3kg unintentionally in past 3 months?” and “Have you had reduced appetite?”	Answer = “Yes” to both
9	Depression Risk	2Q Depression Screening Tool	“Yes” to either question

Screening coverage varied across provinces, with the proportion of older adults screened ranging from 63.61% to 100%. Notably, data from Bangkok was incomplete due to reporting inconsistencies. Nevertheless, all risk proportions were calculated based on the actual number of older adults reported screened in each province. Eligible participants included: (1) older adults registered in the household registry and residing in the area for more than six months; and (2) older adults not registered locally but residing continuously in the area for at least six months. All prevalence estimates were calculated using the number of screened individuals as the denominator within each province. No imputation was performed for unscreened individuals. As such, provincial estimates reflect observed screening data rather than full population prevalence. For prevalence maps, fixed manual breakpoints (< 5%, 5–10%, 10–15%, 15–20%, and > 20%) were applied consistently across all indicators to enhance visual comparability.

This definition aligns with national screening protocols and ensures inclusion of both permanent and semi-permanent residents, particularly in rural or migrant-heavy areas. By accounting for varying levels of coverage and the locally grounded definition of “older adult residents,” the dataset offers a nuanced and practical basis for provincial-level analysis of health vulnerability.

Composite health vulnerability index

To integrate the nine geriatric health risk indicators into a single summary metric, principal component analysis (PCA) was employed. PCA transforms correlated variables into a set of orthogonal components, each representing a linear combination of the original variables, and ranks them according to the proportion of variance explained [23]. Prior to analysis, all indicators were standardized (z-scores) to ensure comparability across differing scales. sampling adequacy was acceptable, with an overall Kaiser-Meyer-Olkin (KMO) of 0.73, and the scree plot demonstrated a clear inflection after the first component (Supplementary Material S1). Only the first principal component (PC1) had an eigenvalue greater than 1 (6.73) and accounted for 74.83% of the total variance, supporting retention of a single dominant component. All nine indicators loaded strongly and positively on PC1 (loadings ranging from 0.77 to 0.92), indicating that the component represents a coherent multidimensional vulnerability construct rather than isolated morbidity measures. Uniqueness values were low (0.04–0.17), suggesting that most of the variance was captured by the dominant component.

PC1 was therefore interpreted as the Composite Health Vulnerability Index (CHVI). Higher CHVI values indicate provinces where multiple geriatric risk domains are concurrently elevated. Provinces with CHVI values greater than + 2 standard deviations were classified as very high vulnerability. This index provides a parsimonious yet robust representation of multidimensional health vulnerability at the provincial level [24, 25]. To assess the robustness of the PCA-based weighting scheme, an alternative equal-weight index was constructed using the same nine indicators. The correlation between the PCA-based CHVI and the equal-weight index was examined to evaluate sensitivity to weighting assumptions. The PCA-based CHVI was highly correlated with the equal-weight index (r = 0.989), indicating minimal sensitivity to weighting assumptions. For CHVI distribution map, class intervals were defined based on the empirical distribution of variable to enhance interpretability.

Local development indicator

NTL was employed as a proxy for local development, capturing spatial patterns of economic activity and infrastructure distribution [26]. NTL data for the year 2024 were obtained from the Visible Infrared Imaging Radiometer Suite satellite dataset, accessible via Google Earth Engine. This dataset provides high-resolution imagery with a spatial resolution of approximately 500 m, making it suitable for sub-national development analysis. To ensure comparability across provinces with varying land areas, raw NTL values were normalized by calculating the average NTL per square kilometer. These normalized values were then aggregated at the provincial level using zonal mean statistics. For the NTL distribution map, class intervals were defined based on the empirical distribution of variable to enhance interpretability.

Spatial analysis

Global spatial autocorrelation

To assess the overall spatial structure of health vulnerabilities, Global Moran’s I was calculated for each of the nine geriatric health risk indicators, the CHVI, and NTL intensity. The spatial weights matrix was constructed using Euclidean distance between provincial centroids with a fixed distance band. The threshold distance (1.002848 decimal degrees, approximately 111.3 km) corresponds to the minimum distance required to ensure that each province had at least one neighboring province, thereby preventing isolated spatial units (islands) and resulting in a fully connected spatial weights structure. This specification was consistently applied across all global and local spatial analyses. The significance of Global Moran’s I statistics was tested using 999 Monte Carlo random permutations. As a robustness check, spatial autocorrelation was re-estimated using a k-nearest neighbors (k = 4) weights matrix. Results were substantively consistent with the fixed-distance specification.

Local spatial autocorrelation

To identify local spatial patterns and statistically significant clusters of geriatric health risks and development disparities, Local Indicators of Spatial Association (LISA) were applied. The univariate LISA analysis was applied to detect spatial clustering patterns of individual health risk indicators and the CHVI across Thai provinces. Significant high–high (H–H) clusters indicate provinces with high levels of health risk surrounded by similarly high-risk neighbors, whereas low–low (L–L) clusters reflect low-risk provinces in low-risk neighborhoods. Spatial outliers—low–high (L–H) and high–low (H–L) clusters—capture provinces with divergent values relative to their neighbors and suggest localized disparities.

Bivariate LISA was conducted to assess the spatial relationship between CHVI and local development, proxied by NTL intensity. This analysis revealed H–L clusters where provinces with high health vulnerability are embedded within low-development contexts—indicating compounded disadvantage and potential priority zones for policy intervention. Conversely, L–H areas point to provinces with relatively low vulnerability surrounded by more developed neighbors, suggesting potential protective or spillover effects.

The spatial weight matrix for LISA was based on Euclidean distance with a fixed threshold of approximately 111.3 km (as defined in the Global Moran’s I analysis), ensuring consistent spatial neighbor definitions across all provinces. The significance of local clustering was tested using 999 Monte Carlo permutations, with a pseudo p-value < 0.05 considered statistically significant. Bivariate LISA maps were particularly useful in highlighting areas where provinces with high CHVI values were spatially associated with neighboring provinces characterized by low NTL intensity (i.e., high health vulnerability adjacent to low local development), offering insight into spatial health inequities that may require targeted intervention. Given the ecological design, the findings were interpreted as spatial associations rather than evidence of causal relationships.

All spatial and statistical analyses were conducted using GeoDa for spatial autocorrelation and LISA statistics, QGIS (3.30.2) for mapping and spatial visualization, and STATA version 15 (licensed to the Faculty of Public Health, Thammasat University, Thailand) for PCA and descriptive statistics.

Results

Descriptive statistics of health risk indicators

The descriptive analysis revealed marked heterogeneity in the distribution of health risks among older adults across Thai provinces (Table 2). Visual impairment was the most prevalent condition, with a median prevalence of 18.04% (IQR: 16.73%–20.59%), followed by oral health problems (median: 16.91%; IQR: 13.80%–19.34%) and mobility limitation (median: 12.83%; IQR: 11.12%–14.66%). In contrast, depression risk had the lowest median prevalence of 2.51% (IQR: 2.05%–3.17%). Other health risks including cognitive impairment, hearing impairment, urinary incontinence, ADL limitation, and malnutrition risk showed moderate median prevalence ranging from 5.73% to 7.82%, with relatively narrow IQR.

Table 2.

Descriptive statistics of health risk indicators among older adults in Thailand

Health risk indicators	Prevalence (%)
	Median	Min	Max	Q1	Q3
Cognitive impairment	6.71	2.78	14.29	5.90	8.04
Mobility limitation	12.83	5.89	19.32	11.12	14.66
Visual impairment	18.04	8.89	55.10	16.73	20.59
Hearing impairment	6.73	2.81	16.26	5.74	7.78
Urinary incontinence	5.73	3.16	11.80	4.85	7.13
ADL limitation	7.82	3.83	13.58	6.46	8.60
Oral health problems	16.91	7.24	34.70	13.80	19.34
Malnutrition risk	6.62	2.75	14.76	5.10	8.03
Depression risk	2.51	1.22	7.51	2.05	3.17

Spatial patterns of health risks among older adults

The spatial distribution maps reveal considerable heterogeneity in the prevalence of health risks among older adults across provinces in Thailand (Fig. 1). The results of the Global Moran’s I statistics indicate significant positive spatial autocorrelation for most health risk indicators, suggesting that provinces with similar risk levels tend to cluster geographically. The strongest spatial autocorrelation was found for oral health problems (Moran’s I = 0.547, p < 0.001), followed by urinary incontinence (I = 0.473, p < 0.001) and malnutrition risk (I = 0.466, p < 0.001) (Table 3).

Fig. 1.

Spatial distribution and LISA cluster maps of health risk domains among older adults across Thai provinces in 2024

Table 3.

Global Moran’s I for spatial autocorrelation of health risks domains among older adults across Thai provinces

Health risk domain	Moran’s I	p-value
Cognitive impairment	0.153	0.031
Mobility limitation	0.328	0.001
Visual impairment	0.111	0.085
Hearing impairment	0.296	0.003
Urinary incontinence	0.473	0.001
ADL limitation	0.269	0.001
Oral health problems	0.547	0.001
Malnutrition risk	0.466	0.001
Depression risk	0.340	0.001

The LISA cluster maps further highlight these spatial patterns. Notably, H–H clusters of oral health problems and urinary incontinence were concentrated in several northeastern provinces, indicating local hotspots of vulnerability. In contrast, visual impairment and cognitive impairment exhibited more scattered and non-significant patterns. L–L clusters were predominantly located in the northern and central regions, reflecting areas with relatively lower prevalence of these risks (Fig. 1).

Spatial patterns of the composite health vulnerability index

The CHVI, derived from PC1, demonstrated substantial spatial heterogeneity across Thai provinces (Fig. 2). The median standardized index was − 0.05 (IQR: -1.41 to 1.51), where higher values indicate greater overall health risk among older adults. Spatial analysis further revealed significant clustering of CHVI values across provinces. Global Moran’s I was 0.261 (p = 0.002), indicating statistically significant positive spatial autocorrelation.

Fig. 2.

Spatial distribution and clustering of the composite health vulnerability index among older adults in Thailand in 2024

The LISA cluster map revealed distinct spatial clusters of vulnerability among older adults. H–H clusters were primarily concentrated in northeastern provinces, including Nong Bua Lam Phu, Khon Kaen, and Buri Ram, indicating localized areas of elevated health risk surrounded by similarly vulnerable regions. In contrast, L–L clusters were observed in northern and central provinces such as Phrae, Sukhothai, Kamphaeng Phet, Ratchaburi, and Samut Sakhon, representing areas with relatively lower vulnerability. Additionally, H–L outliers were detected in Bangkok, Samut Songkhram, and Rayong.

Spatial patterns of local development

Figure 3 illustrates the spatial distribution of mean NTL intensity across Thai provinces, serving as a proxy for local development. Brighter areas on the left map indicate higher NTL values, with the most intense illumination concentrated in Bangkok and its surrounding provinces. The LISA cluster map (bottom-right) reveals significant spatial clustering. H–H clusters are in central Thailand, encompassing Bangkok, Samut Prakan, Nonthaburi, Pathum Thani, Phra Nakhon Si Ayutthaya, Ang Thong, Sing Buri, Saraburi, Chon Buri, Nakhon Pathom, Samut Sakhon, and Samut Songkhram. L–L clusters were observed in several northern provinces including Chiang Mai, Phrae, Sukhothai, Kamphaeng Phet. Additionally, L–H clusters emerged in Suphan Buri, Ratchaburi, Nakhon Nayok, and Chachoengsao, which are adjacent to high-development provinces. The Global Moran’s I statistics confirmed significant positive spatial autocorrelation (I = 0.414, p = 0.003), indicating regional dependence in development levels across the country.

Fig. 3.

Spatial distribution and clustering of local development measured by nighttime light intensity in 2024

Spatial inequity between health vulnerability and local development

The negative bivariate Moran’s I (− 0.174, p < 0.05) indicates an overall spatial dissimilarity between health vulnerability among older adults and local development across Thailand. In general, provinces with high CHVI values tend to be spatially associated with neighboring provinces characterized by lower levels of local development, and vice versa. This pattern reflects a pattern of spatial mismatch or spatial opposites rather than clustering of similar values at the national level.

The bivariate LISA cluster map (Fig. 4) provides further insight into localized spatial dynamics. A limited number of H–H spatial associations were identified, primarily in central Thailand, indicating provinces with high CHVI values spatially associated with neighboring provinces characterized by high levels of local development. Conversely, L–H clusters were observed in several central provinces surrounding Bangkok, reflecting provinces with relatively low health vulnerability that are spatially associated with neighboring provinces exhibiting higher development levels. These spatial patterns underscore regional disparities in the distribution, while also suggesting potential spillover or protective effects associated with proximity to highly developed regions.

Fig. 4.

Bivariate LISA cluster map showing spatial relationship between the composite health vulnerability index and NTL across provinces in Thailand in 2024

Discussion

This study revealed substantial spatial inequities in health risk vulnerability among older adults across Thailand, reflecting a critical intersection between population aging, structural disadvantage, and uneven development. By leveraging province-level spatial analytic techniques, including global and local Moran’s I statistics and bivariate spatial associations, this study identified spatially structured clusters of vulnerability that are not randomly distributed but systematically concentrated in specific regions, particularly in the Northeast.

Several health risk indicators, including oral health problems, urinary incontinence, and malnutrition, exhibited statistically significant positive spatial autocorrelation. These High–High clusters were primarily located in the Northeast, a region historically characterized by lower socioeconomic status, under-resourced health infrastructure, and limited access to geriatric care compared to more affluent central provinces [27, 28]. This pattern echoes findings from other low- and middle-income countries, where regional disadvantage and health system fragmentation contribute to a disproportionate burden of age-related conditions among rural and marginalized populations [19, 29, 30].

The composite index of health vulnerability derived from PCA further reinforced these spatial disparities. Vulnerability hotspots were overwhelmingly concentrated in northeastern provinces, reflecting not only the cumulative effects of individual-level health challenges but also deep-rooted systemic deficiencies in service accessibility, infrastructure, and equitable resource allocation. The Northeast of Thailand, long recognized for entrenched socioeconomic disadvantage, limited healthcare capacity, and restricted fiscal autonomy at the local government level, represents a classic case of structural vulnerability accumulated over the life course. Previous studies have demonstrated strong associations between poor health outcomes and residency in rural northeastern areas [31, 32]. Nearly half of the country’s population living in economically disadvantaged contexts is concentrated in this region, which continues to experience disproportionately high mortality from chronic illnesses such as liver cancer, diabetes, and renal disease [7, 33]. Regional economic indicators further highlight persistent developmental lags: the Northeast consistently records the lowest per capita gross regional product, alongside high household debt burdens and lower educational attainment [34, 35]. These structural disadvantages compound over time, reinforcing cycles of vulnerability among older adults and underscoring the need for spatially targeted, equity-oriented public health interventions. NTL intensity, used in this study as a proxy for local development, exhibited significant spatial autocorrelation, with pronounced clustering in central and metropolitan provinces. This pattern reflects Thailand’s long-standing centralization of infrastructure and economic activity [36].

The bivariate Moran’s I analysis revealed a statistically significant negative spatial relationship between health vulnerability among older adults and local development (I = − 0.174, p < 0.05). Overall, this finding indicates a broader spatial mismatch between development and geriatric vulnerability across Thailand. Such a pattern exemplifies a critical form of spatial inequity, whereby proximity to highly developed regions does not necessarily translate into improved health outcomes in adjacent provinces. This spatial separation between high-development and high-vulnerability areas is consistent with core–periphery dynamics observed in Thai urban epidemiology, where highly urbanized provinces concentrate infrastructure and disease clustering linked to elevated nighttime light intensity [11, 12].

Although the bivariate LISA map (Fig. 4) identified a limited number of H–H spatial associations—provinces with high vulnerability spatially associated with highly developed neighboring provinces—these clusters were relatively few and primarily located in central Thailand. Consistent with the negative global bivariate Moran’s I, such H–H associations represent localized exceptions rather than the dominant national pattern.

The L–H spatial associations observed in several central provinces surrounding Bangkok warrant nuanced interpretation. While these patterns reflect spatial asymmetry between health vulnerability and development, several structural mechanisms may help explain this pattern. First, health workforce distribution remains uneven, with physicians and specialized services concentrated in metropolitan areas. Peripheral provinces may lack geriatric-trained personnel and comprehensive chronic care services, limiting early detection and management of functional decline. Evidence from decentralized healthcare settings in Thailand indicates that insufficient workforce capacity and limited professional development opportunities contribute to service strain and burnout in local systems [17]. Second, transport infrastructure and physical accessibility influence service utilization, as provinces geographically proximate to developed urban centers may still face barriers related to mobility, referral systems, or service coverage. Third, lifecourse socioeconomic disadvantage may contribute to accumulated health risks in less developed regions, where older adults have experienced prolonged exposure to poverty, informal labor, or limited educational opportunity. Finally, urban bias in public investment may concentrate advanced health infrastructure and institutional capacity in central provinces, limiting the equitable diffusion of development benefits. Technological inequities may further compound vulnerability, as lower household internet access has been associated with reduced healthcare utilization among older adults in Thailand [18], potentially restricting access to health information and telemedicine in structurally disadvantaged provinces.

These dynamics align with the core–periphery model of spatial inequality, in which core urban centers such as Bangkok accumulate infrastructure, resources, and institutional capacity, while peripheral areas remain partially excluded from development externalities [37–40]. Policy fragmentation (e.g., limited interprovincial coordination) and infrastructural constraints (e.g., transport or service gaps) may further restrict equitable diffusion of development benefits [39, 41]. In light of these structural dynamics, spatially integrated policy responses may be necessary to address regional inequities. Cross-provincial infrastructure investment, improved health workforce capacity and redistribution, strengthened transport connectivity, and regionally coordinated geriatric health strategies could help ensure that development gains are more equitably shared across territorial boundaries.

Limitations

This study has several limitations. First, the use of province-level aggregated data may mask important intra-provincial heterogeneity, particularly in provinces with diverse urban–rural contexts. Future research should consider finer spatial units to better capture local variation. Second, although NTL intensity is a widely used proxy for local development, it primarily reflects physical infrastructure and economic activity, and does not capture dimensions such as service quality, social capital, or equity in resource distribution. Incorporating multidimensional development indicators would provide a more comprehensive assessment. Third, the health screening data were collected through routine administrative processes by village health volunteers nationwide. Variations in training, assessment practices, and protocol adherence may introduce measurement variability across provinces. Finally, screening coverage varied substantially, and data from Bangkok were incomplete. Because prevalence estimates were calculated using screened individuals as denominators, differential participation may introduce bias if coverage is systematically associated with underlying health risk. Accordingly, the CHVI reflects observed screening data rather than complete population-level prevalence.

Conclusion

The spatial inequities identified in this study raise important considerations for targeted public health planning and equity-oriented resource allocation. Provinces in northeastern Thailand, where high health risks intersect with low development, should be prioritized for place-based interventions, including expansion of mobile and community-based geriatric services, investment in rural health infrastructure, and cross-provincial collaboration to address inter-regional health disparities. Furthermore, the findings suggested that national development strategies should explicitly integrate health equity goals, especially in the context of a rapidly aging population. The failure to do so risks exacerbating existing health disparities and undermining efforts to achieve universal health coverage and sustainable development goals (SDGs), particularly SDG 3 (good health and well-being) and SDG 10 (reduced inequalities).

Authors’ contributions

C.T. and S.S. conceptualized the study. S.S. collected and analyzed the data. C.T. interpreted results. C.T. and S.S. drafted the initial manuscript. All authors reviewed and revised the manuscript for important intellectual content and approved the final version.

Funding

Not applicable.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study used aggregated, de-identified secondary administrative data, and no individual-level identifiable data were accessed; therefore, ethical approval was not required under national regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.