Spatial analysis of tuberculosis based on geographic information systems in Sleman district, Special Region Yogyakarta

International Journal of Public Health Science (IJPHS) Vol. 14, No. 4, December 2025, pp. 1876~1885 ISSN: 2252-8806, DOI: 10.11591/ijphs.v14i4.26826  1876 Journal homepage: http://ijphs.iaescore.com Spatial analysis of tuberculosis based on geographic information systems in Sleman district, Special Region Yogyakarta Makhrum Irmaningsih, Angga Eko Pramono Department of Health Services and Information, Vocational School, Universitas Gadjah Mada, Indonesia Article Info ABSTRACT Article history: Received Jul 29, 2025 Revised Sep 25, 2025 Accepted Nov 3, 2025 The number of tuberculosis cases continues to rise annually, with Sleman Regency recording 2,372 cases in 2024, making it one of the highest in the Special Region of Yogyakarta Province. This study aims to analyze spatial autocorrelation and spatial relationships of tuberculosis cases in Sleman Regency in 2024 using geographic information systems (GIS) and spatial analysis. A quantitative cross-sectional design was applied to 1,406 tuberculosis cases across 86 villages. Bivariate local indicators of spatial association (LISA) analysis were performed using GeoDa software, while geographically weighted regression (GWR) in R Studio examined local environmental influences. Bivariate LISA results showed no significant spatial autocorrelation for population density, air temperature, air humidity, precipitation, and altitude (p-values: 0.173, 0.265, 0.138, 0.312, and 0.401, respectively). GWR revealed negative correlations between these variables and tuberculosis cases. Findings highlight spatial patterns and inform targeted interventions, recommending enhanced tuberculosis awareness and treatment access in low-density, high-incidence areas, along with public education on ventilation and preventive measures during colder seasons, and strengthened prevention in high-risk lowland villages. Keywords: Bivariate LISA GIS GWR Spatial analysis Tuberculosis This is an open access article under the CC BY-SA license. Corresponding Author: Makhrum Irmaningsih Department of Health Services and Information, Vocational School, Universitas Gadjah Mada Caturtunggal, Depok, Sleman, Yogyakarta 55281, Indonesia Email: makhrum.irman[email protected].ac.id 1. INTRODUCTION In 2015, 193 countries declared and agreed on the Sustainable Development Goals (SDGs), which contain the dimensions of the Millennium Development Goals (MDGs), which emphasize the eradication of world poverty by 2030. The SDGs are based on 3 pillars, based on 17 points, which are broken down into 169 targets and 241 indicators that influence each other [1]. In point 3, 9 core targets are all related to health, and one of the 9 targets (point 3.3) is related to efforts to end tuberculosis. This target also includes the commitment of the World Health Organization (WHO) and United Nations SDGs, which include reductions for the tuberculosis incidence rate, the absolute number of deaths caused by tuberculosis, and the cost faced by people with tuberculosis and their households [2]. Tuberculosis is an infectious disease caused by infection with Mycobacterium tuberculosis complex, which is transmitted through the air [3]. Other studies have shown that risk factors for tuberculosis transmission result from suboptimal anti-tuberculosis treatment, socioeconomic factors, and the environment [4]. Globally, in 2023, the number of tuberculosis cases has reached 8.2 million people, indicating an increase from previous years. In that year, Indonesia became the second country contributing to tuberculosis cases (10%) after India Int J Public Health Sci ISSN: 2252-8806  Spatial analysis of tuberculosis based on geographic information systems … (Makhrum Irmaningsih) 1877 (26%) [2]. Based on data from the World Health Organization (WHO), in 2023, there were 281 million people in Indonesia, 1.09 million of whom had been exposed to tuberculosis. The Special Region of Yogyakarta Province continues to face significant challenges in controlling tuberculosis. In 2024, the province ranked 26th nationally in estimated tuberculosis cases, with Sleman Regency reporting the highest burden within the province. A preliminary study conducted in January 2025 by the Sleman Regency Public Health Office revealed an increasing trend in tuberculosis cases: 933 in 2020, 978 in 2021, 1,955 in 2022, 2,509 in 2023, and 2,372 in 2024. Although the health office has utilized geographic information systems (GIS) to map tuberculosis cases on the SmartDinkes platform, the data are limited to aggregated figures at the village level. This study addresses this gap by providing a more granular spatial analysis of tuberculosis distribution in Sleman Regency, applying advanced spatial statistical methods such as local indicators of spatial association (LISA) and geographically weighted regression (GWR). By integrating these techniques, the research offers novel insights into the localized spatial patterns and environmental factors influencing tuberculosis incidence, which have not been previously explored in this region. GIS has become an essential tool in public health, enabling the visualization, analysis, and interpretation of spatial data to improve understanding of disease patterns and guide evidence-based interventions [5]. Specifically, GIS applications enhance tuberculosis surveillance, prevention, and mapping of accessibility to health facilities [6]. Mapping the spatial distribution of tuberculosis cases allows identification of high-risk areas and supports the development of targeted control strategies, thereby facilitating informed decision-making. 2. METHOD This study employed a quantitative method emphasizing objectivity and accuracy in measuring variables and drawing conclusions from the population sample [7]. A cross-sectional design was used, with data collected from the Sleman Regency Health Office in 2024. The total sampling included 2,372 tuberculosis cases undergoing treatment in Sleman Regency. After applying the inclusion criteria—patients residing in Sleman Regency with clearly identified residential villages—a final sample of 1,406 cases was selected. The data analyzed originated from 86 villages across the Sleman Regency area. The study analyzed several independent variables, including the population density, the air temperature, the air humidity, the precipitation, and the altitude. Population density and altitude data were obtained from the Central Statistics Agency of Sleman Regency, while air temperature, air humidity, and precipitation data were sourced from NASA satellite imagery. Spatial autocorrelation analysis is an analysis method to determine the pattern of relationships or correlations between observation locations [8]. It includes global autocorrelation analysis and local autocorrelation analysis (LISA) [9]. This study used GeoDa software to conduct spatial autocorrelation analysis. With its user-friendly interface, GeoDa enables researchers to perform LISA bivariate analysis without requiring advanced programming skills, thereby making the interpretation of results more accessible. Index Moran's I had been used in global autocorrelation analysis with a range of [-1, 1] [10]. When the I index value approaches 1, it indicates a clustered pattern. Then, if the I index value approaches -1, it indicates a spread pattern, and the I index = 0 indicates a random pattern. This analysis used a significance level of 5% [11]. If the p-value <0.05, then the initial hypothesis is rejected, so that there is spatial autocorrelation [12]. Local autocorrelation analysis (LISA) is used to determine the degree of correlation between adjacent areas [13]. This analysis produces 4 quadrants consisting of quadrant 1 (high-high), quadrant 2 (low-high), quadrant 3 (low-low), and quadrant 4 (high-low). The study also employed R Studio software for spatial correlation analysis using geographically weighted regression (GWR). Unlike classical regression models that produce a single global parameter estimate for the entire study area, GWR generates unique parameter estimates at each location, thus providing localized results [14]. GWR is particularly suited for spatial data because traditional regression assumptions of independent and homogenous residuals are often violated in spatial contexts; using classical regression in such cases may yield biased estimates and incorrect conclusions [15]. 3. RESULTS AND DISCUSSION 3.1. Result 3.1.1. Local indicators of spatial autocorrelation (LISA) analysis Based on the results of the bivariate LISA analysis, the Moran's I index between the number of tuberculosis cases and population density is 0.0464, with an expected value E(I) of -0.0118 as shown in Table 1. Since Moran's I value is greater than E(I), this indicates positive spatial autocorrelation or clustering in the data (Figure 1(a)). However, the significance test shows a p-value of 0.173, which is greater than the 0.05 threshold, indicating that the spatial autocorrelation between tuberculosis cases and population density is not statistically significant. Figure 1(b) illustrates two villages in quadrant I (high-high), representing locations where high numbers of tuberculosis cases are surrounded by areas of high population density. These areas are Kalitirto and Tamanmartani. Then, there are seven villages in quadrant II (low-high), representing locations where a low  ISSN: 2252-8806 Int J Public Health Sci, Vol. 14, No. 4, December 2025: 1876-1885 1878 number of cases are surrounded by areas with low population density. These areas are Sendangtirto, Tegaltirto, Jogotirto, Sambirejo, Umbulmartani, Hargobinangun, and Umbulharjo. Furthermore, in quadrant III (low-low), there is one village, namely Sidomulyo. Quadrant III indicates an area with a low number of cases surrounded by areas with low population density. Meanwhile, there are no areas included in quadrant IV (high-low). Based on the results of the bivariate LISA analysis, the Moran's I index between the number of tuberculosis cases and air temperature is -0.0343, with an expected value E(I) of -0.0118 (Table 1). Since Moran's I value is less than E(I), this indicates a negative spatial autocorrelation or a dispersed pattern in the data (Figure 1(c)). However, based on the significance test, the p-value is 0.265, which is greater than 0.05, indicating no statistically significant spatial autocorrelation between the number of cases and air temperature. Figure 1(d) shows two villages in quadrant I (high-high), where areas with a high number of cases are surrounded by areas with high air temperatures. These areas include Argomulyo and Umbulmartani. Then, there are two villages in quadrant II (low-high): Argomulyo and Umbulmartani. Quadrant II indicates that areas with low tuberculosis cases are surrounded by areas with high air temperatures. Furthermore, in quadrant III (Low-Low), there is one area that meets the criteria, namely Sendangarum. Quadrant III indicates an area with a low number of tuberculosis cases surrounded by areas with low air temperatures. Meanwhile, in quadrant IV (high-low), there are three villages. There are Sendangagung, Sendangrejo, and Sidorejo. Quadrant IV indicates that areas with a high number of tuberculosis cases are surrounded by areas with low air temperatures. Based on the results of the bivariate LISA analysis, the Moran's I index between the number of tuberculosis cases and air humidity is 0.0593, with an expected value E(I) of -0.0118 (Table 1). Since Moran's I value is greater than E(I), this indicates positive spatial autocorrelation or clustering in the data (Figure 1(e)). However, the significance test shows a p-value of 0.138, which is greater than 0.05, indicating no statistically significant spatial autocorrelation between tuberculosis cases and air humidity. Figure 1(f) illustrates two villages in quadrant I (highhigh), where areas with a high number of cases are surrounded by areas with high air humidity. These areas include Lumbungrejo and Tlogoadi. Meanwhile, in quadrant II (low-high), there is one village, namely Sinduadi. Quadrant II indicates that areas with a low number of tuberculosis cases are surrounded by areas with high air humidity. Furthermore, in quadrant III (low-low), there are no areas that meet this quadrant. Then, in quadrant IV (high-low), there are four villages. There are Margokaton, Kalitirto, Sumberharjo, and Gayamharjo. Quadrant IV indicates areas with a high number of tuberculosis cases surrounded by areas with low air humidity. Based on the results of the bivariate LISA analysis, the Moran's I index between the number of tuberculosis cases and precipitation is 0.0248, with an expected value E(I) of -0.0118 (Table 1). Since Moran's I index is greater than E(I), this indicates positive spatial autocorrelation or clustering in the data (Figure 1(g)). However, the significance test shows a p-value of 0.312, which is greater than 0.05, indicating no statistically significant spatial autocorrelation between tuberculosis cases and precipitation. Figure 1(h) shows two villages in quadrant I (High-High), where areas with a high number of cases are surrounded by areas with high precipitation. These areas include Tlogoadi and Wedomartani. Meanwhile, in quadrant II (Low-High), there are five villages. There are Tamanmartani, Sendangadi, Tridadi, Triharjo, and Pandowoharjo. Quadrant II indicates that areas with a low number of tuberculosis cases are surrounded by areas with high precipitation. Furthermore, there are no areas included in quadrant III (low-low). Meanwhile, in quadrant IV (high-low), there are four villages. There are Margokaton, Kalitirto, Sumberharjo, and Gayamharjo. Quadrant IV indicates that areas with a high number of tuberculosis cases are surrounded by areas with low precipitation. Based on the results of the bivariate LISA analysis, the Moran's I index between the number of tuberculosis cases and altitude is -0.0122, with an expected value E(I) of -0.0118 (Table 1). Since Moran's I value is slightly less than E(I), it indicates negative spatial autocorrelation or a dispersed spatial pattern (Figure 1(i)). The significance test yielded a p-value of 0.401, which is greater than 0.05, indicating no statistically significant spatial autocorrelation between tuberculosis cases and altitude. Figure 1(j) shows that no areas fall within quadrant I (high-high). Meanwhile, in quadrant II (low-high), there are 7 villages. There are Balecatur, Sidomulyo, Sidokarto, Sidoarum, Banyuraden, Tirtoadi, and Margoluwih. Quadrant II indicates that areas with a low number of tuberculosis cases are surrounded by high altitude. Furthermore, in quadrant III (low-low), there are 5 villages. There are Triharjo, Sendangadi, Umbulmartani, Selomartani, and Sambirejo. Quadrant III indicates that areas with a low number of tuberculosis cases are surrounded by low altitude. Meanwhile, there are no areas included in quadrant IV (high-low). Table 1. The result of the bivariate LISA analysis Variables Moran’s I E(I) Z-score P-value Classification Population density 0.0464 -0.0118 0.9562 0.17300 Not Significant Air temperature -0.0343 -0.0118 -0.6061 0.26500 Not Significant Air humidity 0.0593 -0.0118 1.1089 0.13800 Not Significant Precipitation 0.0248 -0.0118 0.4702 0.31200 Not Significant Altitude -0.0122 -0.0118 -0.2880 0.40100 Not Significant Int J Public Health Sci ISSN: 2252-8806  Spatial analysis of tuberculosis based on geographic information systems … (Makhrum Irmaningsih) 1879 (a) (b) (c) (d) (e) (f) (g) (h) (i) (j) Figure 1. Scatterplots and clustering maps of population density, air temperature, air humidity, precipitation, and altitude: (a) population density scatterplot, (b) population density clustering map, (c) air temperature scatterplot, (d) air temperature clustering map, (e) air humidity scatterplot, (f) air humidity clustering map, (g) precipitation scatterplot, (h) precipitation clustering map, (i) altitude scatterplot, and (j) altitude clustering map 3.1.2. Geographically weighted regression (GWR) analysis Spatial relationship analysis was performed using geographically weighted regression (GWR) in R Studio, starting with an ordinary least squares (OLS) test. The OLS model yielded a multiple R-squared of 0.3758 and an AICc of 676.45, while the GWR model showed a higher R-squared of 0.7482 and an AICc of  ISSN: 2252-8806 Int J Public Health Sci, Vol. 14, No. 4, December 2025: 1876-1885 1880 649.08. These results indicate that GWR provides a better fit for the spatial data than OLS. Local GWR coefficients between tuberculosis cases and environmental factors are presented in Table 2. Table 2. The local GWR coefficient between tuberculosis cases and environmental factors in Sleman Regency in 2024 Variable Min. 1st Qu. Median 3rd Qu. Max Intercept -3,501.0 -1,618.1 -769.95 -3.768 244.6836 Population density -0.0013101 0.0008785 - 0.0024384 0.0046412 0.0151 Air temperature -0.57189 6.702 12.384 24.406 49.0803 Air humidity -2.2463 2.7164 6.5674 13.42 29.0272 Precipitation -44.332 -24.844 -11.51 -6.4871 0.0160 Altitude -0.077513 -0.013763 0.0013857 0.033011 0.3032 Table 2 shows that all environmental factors have a negative correlation with tuberculosis cases. Based on the GWR analysis, spatial relationships between tuberculosis cases and environmental factors are determined by p-values (p < 0.05). The results indicate that 27 villages show significant spatial relationships between tuberculosis cases and population density (Figure 2(a)), including Banyuraden, Candibinangun, Caturtunggal, Condongcatur, Donoharjo, Donokerto, Girikerto, Harjobinangun, Maguwoharjo, Minomartani, Nogotirto, Pakembinangun, Pandowoharjo, Purwobinangun, Sardonoharjo, Sariharjo, Sendangadi, Sendangtirto, Sinduadi, Sinduharjo, Sukoharjo, Tegaltirto, Trimulyo, Umbulmartani, Wedomartani, Widodomartani, and Wonokerto. For the air temperature variable, 13 villages were found to have a spatial relationship with tuberculosis cases (Figure 2(b)). There are Bokoharjo, Jogotirto, Kalitirto, Madurejo, Maguwoharjo, Purwomartani, Selomartani, Sendangtirto, Sumberharjo, Tamanmartani, Tegaltirto, Tirtomartani, and Wedomartani. Regarding air humidity, 32 villages exhibited a spatial relationship with tuberculosis incidence (Figure 2(c)), including Bangunkerto, Banyuraden, Candibinangun, Caturtunggal, Condongcatur, Donoharjo, Donokerto, Girikerto, Hargobinangun, Harjobinangun, Maguwoharjo, Minomartani, Nogotirto, Pakembinangun, Pandowoharjo, Purwobinangun, Sardonoharjo, Sariharjo, Sendangadi, Sendangtirto, Sinduadi, Sinduharjo, Sukoharjo, Tegaltirto, Tridadi, Trihanggo, Trimulyo, Umbulmartani, Wedomartani, Widodomartani, Wonokerto, and Wukirsari. Based on precipitation, 26 villages also showed a spatial relationship with tuberculosis cases (Figure 2(d)), namely Candibinangun, Caturtunggal, Condongcatur, Donoharjo, Donokerto, Girikerto, Harjobinangun, Maguwoharjo, Minomartani, Nogotirto, Pakembinangun, Pandowoharjo, Purwobinangun, Sardonoharjo, Sariharjo, Sendangadi, Sendangtirto, Sinduadi, Sinduharjo, Sukoharjo, Trihanggo, Trimulyo, Umbulmartani, Wedomartani, Widodomartani, and Wonokerto. A total of 41 villages were found to have a significant spatial association between tuberculosis cases and altitude (Figure 2(e)), including Argomulyo, Bangunkerto, Banyuraden, Bimomartani, Candibinangun, Caturtunggal, Condongcatur, Donoharjo, Donokerto, Girikerto, Hargobinangun, Harjobinangun, Maguwoharjo, Minomartani, Nogotirto, Pakembinangun, Pandowoharjo, Purwobinangun, Sardonoharjo, Sariharjo, Selomartani, Sendangadi, Sendangtirto, Sidoarum, Sidomoyo, Sinduadi, Sinduharjo, Sukoharjo, Sumberadi, Tirtoadi, Tlogoadi, Tridadi, Trihanggo, Triharjo, Trimulyo, Umbulharjo, Umbulmartani, Wedomartani, Widodomartani, Wonokerto, and Wukirsari. 3.2. Discussion 3.2.1. Spatial analysis of the tuberculosis cases and population density Based on the LISA analysis results, there was no spatial autocorrelation between the number of tuberculosis cases and population density in Sleman Regency. This may occur because some areas have a high number of cases but low population density, and vice versa. These findings are consistent with research conducted in Central Java Province in 2022, which reported a p-value of 0.449 [16]. A bivariate analysis in Kupang from 2022 to 2023 also reported p-values of 0.318 and 0.353 between tuberculosis cases and population density [17]. However, these results contrast with studies in Surakarta, northern China, Bekasi, and West Java [18]-[21]. Based on the local coefficient in GWR analysis, a negative correlation was shown, with 27 villages exhibiting a spatial relationship with the number of tuberculosis cases and population density. This means that in certain local areas, higher population density is associated with fewer reported tuberculosis cases. This may be due to the relatively adequate availability of healthcare facilities in high population density areas, along with the presence of good personal hygiene practices, which together contribute to reducing the incidence of tuberculosis [22]. Meanwhile, this result contrasts with other studies conducted in Makasar City in 2022 and Gombak, Malaysia [23], [24]. Other research has shown that population density is a factor associated with increased tuberculosis cases [25]. High population density can accelerate the spread of tuberculosis bacteria, Int J Public Health Sci ISSN: 2252-8806  Spatial analysis of tuberculosis based on geographic information systems … (Makhrum Irmaningsih) 1881 leading to a higher number of cases [12]. Moreover, high population density can result in the formation of slum areas, where 22 studies reported that a person's chance of contracting tuberculosis is almost three times higher than the national average [26]. From this research, it is hoped that health workers will be more active in socializing tuberculosis disease and increasing access to and coordination of treatment services, especially in areas with low population density but still high tuberculosis cases, so that prevention and control can be more effective. (a) (b) (c) (d) (e) Figure 2. Geographically weighted regression map between tuberculosis cases with population density, air temperature, air humidity, precipitation, and altitude. Geographically weighted regression map of (a) population density, (b) air temperature, (c) air humidity, (d) precipitation, and (e) altitude 3.2.2. Spatial analysis of the tuberculosis cases and air temperature Based on the LISA analysis results, no spatial autocorrelation was found between the number of tuberculosis cases and air temperature in Sleman Regency. However, these results differ from a study in Nepal during 2020-2023, where Moran's I index showed a positive spatial autocorrelation between land surface  ISSN: 2252-8806 Int J Public Health Sci, Vol. 14, No. 4, December 2025: 1876-1885 1882 temperature and the tuberculosis prevalence [27]. Moreover, the GWR analysis revealed a negative correlation, with 13 villages exhibiting a spatial relationship with the number of tuberculosis cases and air temperature, indicating that the lower temperature increases the risks of tuberculosis cases [28]-[30]. Another study in Mozambique showed higher tuberculosis incidence in areas with colder temperatures compared to other areas [31]. Tuberculosis transmission may increase due to more indoor activities and reduced ventilation during colder weather, which facilitates transmission [32], [33]. Additionally, cold air temperatures can reduce the body's immunity, thereby increasing individual susceptibility to infection and enhancing the survival of tuberculosis bacteria in airborne droplets [34]. Further studies have reported strong associations between environmental factors, including air temperature, and tuberculosis cases [31], [35]-[37]. Environmental factors, allergic reactions in the respiratory tract, and decreased temperatures may also cause bronchoconstriction and airway narrowing, potentially damaging respiratory epithelial cells and worsening tuberculosis conditions [38]. Given the negative correlation between air temperature and tuberculosis incidence, governments can seek to increase awareness and preventative interventions, particularly during cold periods, by improving indoor ventilation and encouraging protective behaviors to reduce the risk of transmission in vulnerable populations. 3.2.3. Spatial analysis of the tuberculosis cases and air humidity Based on the results of the bivariate analysis using LISA, there was no spatial autocorrelation between air humidity and the number of tuberculosis cases in Sleman Regency. However, GWR analysis revealed a negative correlation, with 32 villages showing significant spatial relationships. Similar findings have been reported in the southeast and northwest regions of China during 2010-2017 [39]. A study conducted in Beijing, China, also found a negative correlation between the number of tuberculosis cases and air humidity during 2004-2016 [40]. Other studies examining air humidity have reported associations between tuberculosis incidence and humidity levels [41], [42]. Low air humidity is known to increase resistance in the respiratory tract to tuberculosis bacteria, due to reduced protective mucus on the respiratory surface [30], [40]. From this condition, it is hoped that the government will enhance public education on the importance of optimal ventilation and air humidity as part of intensified interventions to prevent tuberculosis transmission. 3.2.4. Spatial analysis of the tuberculosis cases and precipitation Based on the LISA analysis results, there was no spatial autocorrelation between the number of tuberculosis cases and precipitation in Sleman Regency. In contrast, previous studies reported positive spatial autocorrelation between tuberculosis prevalence and precipitation in Nepal (2020-2023) and China (20052015) [10], [27]. Meanwhile, the GWR analysis showed a negative correlation, identifying 26 villages with a spatial relationship to tuberculosis cases. This suggests that the lower precipitation in Sleman Regency is inversely related to higher tuberculosis incidence. Supporting this, a study in Khuzestan Province, Southern Iran, and China (2010-2017) found that increased average precipitation is associated with decreased tuberculosis cases [39], [41]. Additionally, areas with low annual precipitation and dry climates have a higher risk of tuberculosis spread [41]. Conversely, other research indicates that high precipitation increases air humidity, which can support tuberculosis bacteria growth [43]. During the rainy season, individuals have a 3.33 times higher risk of contracting tuberculosis due to increased indoor physical activity [44]. Thus, precipitation can create ideal conditions for the growth of tuberculosis-causing bacteria [45]. 3.2.5. Spatial analysis of the tuberculosis cases and altitude Based on the results of the bivariate LISA analysis, there was no spatial autocorrelation between the number of tuberculosis cases and altitude in Sleman Regency. However, the GWR analysis revealed a negative correlation between tuberculosis cases and altitude, with 41 villages showing a spatial relationship with tuberculosis case numbers. The altitude in Sleman Regency ranges up to less than 2500 meters above sea level, increasing toward the peak of Mount Merapi on the northern side, while the southern side has lower altitudes. Data presentation shows that moderate to high tuberculosis cases are mostly found in areas below 500 meters above sea level. Therefore, the lower the altitude in Sleman Regency, the higher the number of tuberculosis cases, indicating an inverse relationship. Lowland areas have a greater potential for tuberculosis transmission compared to highlands, as indicated by the decrease in tuberculosis cases with increasing altitude [46]. It is recommended that the government and stakeholders focus on strengthening tuberculosis prevention and control efforts in lowland areas that have a higher risk, by improving access to health services, health education on risk factors and transmission, and optimizing early detection and complete treatment programs. Additionally, specific interventions related to environmental management and enhancing vitamin D levels through adequate sunlight exposure in low-altitude areas should be considered to support the population’s immune response. Furthermore, highland areas receive higher exposure to UV-B rays, which can increase vitamin D levels, enhancing immune response and reducing the risk of tuberculosis reactivation. Previous epidemiological studies in Kenya, Peru, Mexico, and Vietnam similarly revealed decreased tuberculosis death and notification Int J Public Health Sci ISSN: 2252-8806  Spatial analysis of tuberculosis based on geographic information systems … (Makhrum Irmaningsih) 1883 rates in highland areas [47]. A comparable pattern is observed in China, where tuberculosis prevalence is higher in low-altitude and coastal regions, while mountainous and hilly areas have lower prevalence [48]. Results from the bivariate LISA analysis indicated no significant spatial autocorrelation between tuberculosis cases and the environmental and demographic variables studied. This lack of significance may suggest weak or spatially inconsistent relationships at local scales or the influence of unmeasured confounding factors. Consequently, GWR was employed to capture spatial heterogeneity and local variations in the associations between tuberculosis cases and predictors such as population density, air temperature, air humidity, precipitation, and altitude. The GWR model revealed significant negative spatial correlation across all examined variables, implying that higher values of these environmental factors were locally associated with reduced tuberculosis case counts. However, this study is limited by its reliance on secondary tuberculosis case data, which may lack granularity and timely updates. Furthermore, the environmental variables of humidity, rainfall, and temperature were aggregated on an annual basis, potentially overlooking seasonal fluctuations. To address these limitations and gain a more comprehensive understanding of the spatiotemporal dynamics influencing tuberculosis, it is recommended that future analyses incorporate data at quarterly intervals to reflect seasonality more accurately. In addition, supplementary analyses that include socioeconomic factors such as poverty levels, sanitation conditions, and health education are necessary, as these determinants may have a more substantial role in tuberculosis transmission dynamics within the study area. 4. CONCLUSION This study aimed to investigate the spatial relationship between environmental factors and tuberculosis cases. This study found no significant spatial autocorrelation between population density, air temperature, air humidity, precipitation, and altitude with tuberculosis cases. However, all these variables showed a negative correlation with tuberculosis cases, indicating that higher values are linked to fewer cases. Advanced spatial and temporal analyses are necessary to identify hidden risk factors and local dynamics, which will clarify temporal trends and spatial variability in tuberculosis distribution. The study also highlights the need for health workers to actively promote tuberculosis awareness and enhance treatment access, especially in low-density areas with high tuberculosis cases. Given the negative correlation between air temperature and tuberculosis cases, governments should increase public awareness during colder periods by improving indoor ventilation and encouraging protective behaviors. Intensifying education on optimal ventilation and humidity is essential. Governments are advised to strengthen tuberculosis prevention in high-risk lowland villages through improved healthcare access, targeted health education, early detection, and comprehensive treatment programs. FUNDING INFORMATION This study was carried out independently without any financial assistance or sponsorship from external organizations. All funding was provided by the researcher personally. AUTHOR CONTRIBUTIONS STATEMENT This journal uses the Contributor Roles Taxonomy (CRediT) to recognize individual author contributions, reduce authorship disputes, and facilitate collaboration. Name of Author C M So Va Fo I R D O E Vi Su P Fu Makhrum Irmaningsih ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Angga Eko Pramono ✓ ✓ ✓ ✓ ✓ C : Conceptualization M : Methodology So : Software Va : Validation Fo : Formal analysis I : Investigation R : Resources D : Data Curation O : Writing - Original Draft E : Writing - Review & Editing Vi : Visualization Su : Supervision P : Project administration Fu : Funding acquisition CONFLICT OF INTEREST STATEMENT The authors declare that there is no conflict of interest regarding the publication of this paper. ETHICAL APPROVAL This study was conducted after obtaining a research eligibility letter (ethical clearance) from the Health Research Ethics Committee of the Yogyakarta Ministry of Health Polytechnic, numbered No.DP.04.03/e-KEPF.1/409/2025. All regulations related to research ethics and data protection have been  ISSN: 2252-8806 Int J Public Health Sci, Vol. 14, No. 4, December 2025: 1876-1885 1884 described in the research eligibility letter. Through this letter, the confidentiality of patient data has been guaranteed and its use is limited to research purposes. DATA AVAILABILITY The data that support the findings of this study are available from the corresponding author, [AEP], upon reasonable request. REFERENCES [1] A. R. Nunes, K. Lee, and T. O’Riordan, “The importance of an integrating framework for achieving the Sustainable Development Goals: the example of health and well-being,” BMJ Global Health, vol. 1, no. 3, 2016, doi: 10.1136/bmjgh-2016-000068. [2] World Health Organization (WHO), “Global tuberculosis report 2024,” Geneva, Switzerland, 2024. [Online]. Available: https://www.who.int/publications/i/item/9789240098073%7D. [3] M. Pai et al., “Tuberculosis,” Nature Reviews Disease Primers, vol. 2, no. 1, p. 16076, Oct. 2016, doi: 10.1038/nrdp.2016.76. [4] K. F. Ortblad, J. A. Salomon, T. Bärnighausen, and R. Atun, “Stopping tuberculosis: a biosocial model for sustainable development,” The Lancet, vol. 386, no. 10010, pp. 2354–2362, 2015, doi: 10.1016/S0140-6736(15)00324-4. [5] P. W. Biu, C. N. Nwasike, O. A. Tula, C. A. Ezeigweneme, and J. O. Gidiagba, “A review of GIS applications in public health surveillance,” World Journal of Advanced Research and Reviews, vol. 21, no. 1, pp. 030–039, Jan. 2024, doi: 10.30574/wjarr.2024.21.1.2684. [6] S. Mishra, P. K. Sahu, A. K. Sarkar, B. Mehran, and S. Sharma, “Geo-spatial site suitability analysis for development of health care units in rural India: effects on habitation accessibility, facility utilization and zonal equity in facility distribution,” Journal of Transport Geography, vol. 78, pp. 135–149, Jun. 2019, doi: 10.1016/j.jtrangeo.2019.05.017. [7] A. Queirós, D. Faria, and F. Almeida, “Strengths and limitations of qualitative and quantitative research methods,” European Journal of Education Studies, vol. 3, no. 9, pp. 369–387, 2017. [8] G. Grekousis, Spatial analysis methods and practice. Cambridge University Press, 2020, doi: 10.1017/9781108614528. [9] Y. Yu, B. Wu, C. Wu, Q. Wang, D. Hu, and W. Chen, “Spatial-temporal analysis of tuberculosis in Chongqing, China 2011-2018,” BMC Infectious Diseases, vol. 20, no. 1, p. 531, Dec. 2020, doi: 10.1186/s12879-020-05249-3. [10] Y. Zhang et al., “Spatial distribution of tuberculosis and its association with meteorological factors in mainland China,” BMC Infectious Diseases, vol. 19, no. 1, 2019, doi: 10.1186/s12879-019-4008-1. [11] M. da Saúde and S. Brasil, “Sistemas de informações geográficas e análise espacial na saúde pública,” (Série B. Textos Básicos de Saúde) (Série Capacitação e Atualização em Geoprocessamento em Saúde; 2), vol. 2, p. 148, 2007. [12] X. Chen, M. Emam, L. Zhang, R. Rifhat, L. Zhang, and Y. Zheng, “Analysis of spatial characteristics and geographic weighted regression of tuberculosis prevalence in Kashgar, China,” Preventive Medicine Reports, vol. 35, p. 102362, Oct. 2023, doi: 10.1016/j.pmedr.2023.102362. [13] Y. Wang, Y. Huang, Y. Guo, L. Wang, Z. Cao, and M. Wu, “Two-mode fluorescent detection of cyanide by a simple AIE-based chemosensor with red emission,” Zeitschrift fur Anorganische und Allgemeine Chemie, vol. 646, no. 11–12, pp. 526–531, 2020, doi: 10.1002/zaac.202000159. [14] O. Raza, M. A. Mansournia, A. Rahimi Foroushani, and K. Holakouie-Naieni, “Geographically weighted regression analysis: a statistical method to account for spatial heterogeneity,” Archives of Iranian Medicine, vol. 22, no. 3, pp. 155–160, 2019. [15] R. Hidayat N, B. W. Otok, Z. Mahsyari, S. H. Sa’diyah, and D. A. Fadhila, “Geographically weighted regression for prediction of underdeveloped regions in East Java Province based on poverty indicators,” in Proceedings of the 2nd International Conference Postgraduate School, SCITEPRESS - Science and Technology Publications, 2018, pp. 898–907, doi: 10.5220/0007553708980907. [16] S. Puspita, “Spatial autocorrelation analysis of pulmonary tuberculosis cases in Central Java Province,” Jurnal Biometrika dan Kependudukan, vol. 13, no. 1, pp. 90–99, Jul. 2024, doi: 10.20473/jbk.v13i1.2024.90-99. [17] M. Lengari, P. Weraman, Y. K. Syamruth, L. P. Ruliati, and A. A. Adu, “Spatial autocorrelation of population density, HIV/AIDS, and diabetes mellitus with pulmonary tuberculosis in Kupang, East Nusa Tenggara, Indonesia,” Indonesian Journal of Medicine, vol. 10, no. 2, pp. 93–104, 2025, doi: 10.26911/theijmed.2025.10.2.825. [18] D. Nabila Silva Diba, B. Murti, and N. A. Setiyadi, “Spatial analysis of pulmonary tuberculosis risk in Surakarta, Central Java, Indonesia,” Journal of Epidemiology and Public Health, vol. 9, no. 3, pp. 386–406, 2024, doi: 10.26911/jepublichealth.2024.09.03.12. [19] X. Wang et al., “Spatiotemporal epidemiology of, and factors associated with, the tuberculosis prevalence in northern China, 20102014,” BMC Infectious Diseases, vol. 19, no. 1, 2019, doi: 10.1186/s12879-019-3910-x. [20] Y. R. Inggarputri, I. Trihandini, P. D. Novitasari, and M. R. Makful, “Spatial analysis of tuberculosis cases diffusion based on population density in Bekasi Regency in 2017-2021,” BKM Public Health and Community Medicine, p. e6462, 2023, doi: 10.22146/bkm.v39i01.6462. [21] E. P. Madao, E. Hermawati, N. A. A. Putri, and M. R. Makful, “Evaluating spatial analysis of tuberculosis prevalence to identify priority districts or municipalities that need policy attention in West Java,” BKM Public Health and Community Medicine, p. e12160, 2024, doi: 10.22146/bkm.v40i04.12160. [22] A. Efendi and T. I. Darwis, “Spatial analysis of the influence of residential density on the spread of tuberculosis cases in Pasar Rebo General Hospital Service Area,” Cities and Urban Development Journal, vol. 1, no. 1, Jun. 2023, doi: 10.7454/cudj.v1i1.1000. [23] S. M. Meliyana, A. S. Ahmar, and Siti Nurazizah Auliah, “Geographically weighted poisson regression (GWPR) model with fixed Gaussian Kernel and fixed bi-square Kernel weights,” ARRUS Journal of Social Sciences and Humanities, vol. 5, no. 2, pp. 896–909, 2025, doi: 10.35877/soshum3812. [24] N. A. Mohidem, M. Osman, Z. Hashim, F. M. Muharam, S. M. Elias, and R. Shaharudin, “Association of sociodemographic and environmental factors with spatial distribution of tuberculosis cases in Gombak, Selangor, Malaysia,” PLoS ONE, vol. 16, no. 6 June 2021, 2021, doi: 10.1371/journal.pone.0252146. [25] M. De Abreu E Silva, C. Di Lorenzo Oliveira, R. G. Teixeira Neto, and P. A. Camargos, “Spatial distribution of tuberculosis from 2002 to 2012 in a midsize city in Brazil,” BMC Public Health, vol. 16, no. 1, 2016, doi: 10.1186/s12889-016-3575-y. [26] E. Noykhovich, S. Mookherji, and A. Roess, “The risk of tuberculosis among populations living in slum settings: a systematic review and meta-analysis,” Journal of Urban Health, vol. 96, no. 2, pp. 262–275, 2019, doi: 10.1007/s11524-018-0319-6. [27] R. K. Mahato et al., “Spatial autocorrelation with environmental factors related to tuberculosis prevalence in Nepal, 2020–2023,” Infectious Diseases of Poverty, vol. 14, no. 1, 2025, doi: 10.1186/s40249-025-01283-y.