Identification of Spatial Patterns of Community Health Centers and Health Disasters: Learning from the Covid-19 Pandemic in Magelang District, Central Java, Indonesia

Regional development in an area will have consequences for the health status of the surrounding community. This paper discusses the adequacy of community health center (CHC) facilities in taking an important role in managing health disasters such as the Covid-19 pandemic. Covid-19 is one of the most infectious environmental-based diseases. The research objective was to spatially identify the availability of CHC and their relationship to Covid-19 cases during the pandemic. Ecological studies are used with a spatial approach. The population in this study were all sub-districts in Magelang District with a total sampling. Spatial analysis makes use of the QGIS and Geoda applications. The results showed that there was grouped spatial autocorrelation (Moran’s I = 0.089, Io = 0.05) between CHC in Magelang. Covid-19 in Magelang residents has a positive autocorrelation with CHC (Moran’s I = 0.248, Io = 0.05) and forms a cluster pattern. The spatial lag regression further clarifies that there is a spatial autocorrelation between the two variables (Coef: -0.175; p value= 0.569), and the Covid-19 variable has a significant influence on the CHC (p=0.0022). Analysis using the Local Indicator Spatial Association (LISA) method found that Ngluwar Sub-district is in the High-High quadrant, while Mungkid Sub-district is in the Low-High quadrant and the other sub-districts are not significant. Spatial pattern heterogeneity is formed in CHC and there is a spatial autocorrelation relationship between Covid-19 cases and CHC. Systematic planning is needed to overcome the adequacy of the CHC to help improve the quality of public health.


Introduction
The role of the government in its capacity as a coach for regional development may play an important role and facilitation efforts in the direction of regional development management procedures that occur in regions [1].The four components of regional development impact are settlement, production, protection, and infrastructure.Infrastructure is a critical pillar supporting the fundamental 1264 (2023) 012040 IOP Publishing doi:10.1088/1755-1315/1264/1/012040 2 goal of promoting enhanced care standards and well-being for the whole population, along with a good experience in the health care system [2].Health sector development is a development strategy that is an alternative to development and becomes an important aspect of regional development, contributing to the general improvement of people's living conditions [3,1].
Community health centers in Indonesia offer primary care and serve as gatekeepers for all health care services [4].Each Community health centers (CHC) has responsibility for community health efforts, which include public health activities for the population in their working area [5,6].In 2020, there were roughly 10,203 Puskesmas were dispersed across the country, with around 30,000 inhabitants on average in each catchment region and fewer people in rural areas.Furthermore, Puskesmas offers both curative and public health services, with an emphasis on critical service areas such as disease control and prevention; health promotion; community nutrition; family planning; child and maternal health; and environmental health (including sanitation and water), making Puskesmas the cornerstone for providing health care services [7,8].
The function of Puskesmas during a pandemic is crucial and plays a frontline role in prevention, control, and mitigation efforts [9].Puskesmas is a part of the healthcare system in Indonesia and has a strategic role in providing healthcare services to the community, especially at the most basic or primary level.The Puskesmas plays a very important role because it is in the middle of a residential area, and related to the population, there are many studies that explain the determinants of Covid-19 related to the population, such as socio-economic conditions, density, geographical factors, and sociocultural [10,11,12].Some functions of Puskesmas during a pandemic are case detection and search.They conduct early identification of symptoms and signs of diseases, collect case data, and report to local or higher health authorities.Puskesmas provides public health services including vaccination, health examinations, health education, and dissemination of important health information to the community.During a pandemic, this includes providing up-to-date information on preventive measures and health protocols that must be followed [13].Isolation and Referral: Puskesmas is responsible for providing isolation services for patients suspected of being infected with pandemic viruses.They may also refer more severe cases to hospitals or other healthcare facilities that are better equipped to handle serious cases.Contact Tracing: Puskesmas assists in tracing close contacts of confirmed COVID-19 patients or other infectious diseases.This is done to prevent further spread and identify potential new cases.Education and Socialization: Puskesmas plays a role in providing education and socialization to the community about the importance of adhering to health protocols, such as handwashing, wearing masks, maintaining physical distance, and avoiding crowds.This education is essential to raise public awareness in facing the pandemic.
When viewed from a regional perspective, Indonesia, which is an archipelagic country with the third largest population in Asia, makes the Puskesmas have a strategic role and is the most important basic health service facility during a pandemic because the puskesmas is able to reach the smallest communities and can help the main hospital was not overwhelmed [14,15].Nationally, the ratio of puskesmas to sub-districts is 1.39, meaning that an average of 1 sub-district has 1.39 puskesmas.There are 3,623 inpatient health centers and 6,370 non-inpatient health centers.Puskesmas is the coordinator and person in charge of health services in its working area which includes private health facilities [16].Covid-19 has the character of its rapid spread because it is related to the presence of the population [17,10].The effectiveness of the performance of the Puskesmas in serving the population will be closely related to the population or the proportion of the number of Puskesmas to the population [18].
It is interesting to observe the proportion of CHC with Covid-19 cases and the concept of spatial autocorrelation is suitable for exploring the causal correlation between the proportion of CHC and the spread of cases during the pandemic [19].The concept of autocorrelation comes from a geographical perspective close to the concept of community health [20].The spread of infectious diseases is always close to the concept of regional neighborliness considering that between regions are connected because location will be closely related to all the activities of the population in the area, and health is one of the factors that drives it [21].For this reason, the existence of basic health services is one of the factors that needs to be observed in the spread of disease in an area.
The existence of CHC which have different proportions in each region to the total population raises the question of how far the benefits of CHC are for the distribution of the spread of Covid-19 during the pandemic?This paper aims to explore the relationship between the proportion of Puskesmas as basic health services and the incidence of Covid-19 by using a spatial autocorrelation approach between regions.The results of this study lead to more local, regional knowledge about how the Puskesmas made an impact in Magelang District, Central Java Province during the Covid-19 pandemic.

Data & Methods
This study was conducted in Magelang District, Central Java Province, Indonesia (Figure 1).Official data is used to support research sourced from official agencies such as the Office of Communication and Information, the Office of Health and the Central Agency on Statistics of Magelang District.Data in aggregate form on the population in 21 districts.The variability of Covid-19 sufferers during the 2021 pandemic and data on the proportion of puskesmas to the population were quantified using the QGIS Geographic Information System (GIS) software version 3.10 and the Geoda application to see the distribution of Covid-19 and balance it with the proportion of puskesmas in Magelang District, Central Java.Geostatistical analysis with global spatial autocorrelation technique was performed and produced the Morans Index.If the Morans I value is above the Io value, it means that there is a spatial autocorrelation between variables.Moran's I values can range from -1 to +1, With positive values indicating positive spatial autocorrelation indicating spatial grouping or the presence of similar values in the nearest or neighboring sub-district locations [22].If the opposite value obtained is negative, it can be interpreted that there is an indication of spatial distribution, but the value obtained is different at the location of the surrounding sub-district or neighbors.The test was continued with the Morans I autocorrelation bivariate test to see the form of the relationship between the two variables involved.If the results indicate that there is autocorrelation between regions, the test can be continued at the spatial regression stage using spatial lag regression analysis to see the significance level of the relationship between these variables.Local indicator of spatial association (LISA) analysis was used in this study to see the relationship between regions in terms of the strong tendency of neighboring relations [23]

Result and Analysis
Covid-19 is spread evenly throughout the region with an average case of 1021.38 cases per subdistrict in the period November 2020 to August 2021.The map of the distribution of Covid-19 in Magelang District shows Mertoyudan sub-district is the area with the highest cases with 3198 cases (15%), then Secang sub-district 1984 cases (9%), and Mungkid sub-district 1755 cases (8%).The map shows that the area south of Magelang District is the area with the highest cases, this area is directly adjacent to the Special Province of Yogyakarta The visualization on the map shows that there are 4 sub-districts with the highest case category with over 1430 cases and 5 sub-districts with under 446 cases.Magelang District is the Regency with the highest cases during 2020 with 4116 cases under Semarang City with 11038 cases in Central Java Province [24].To find out whether there is a tendency for spatial correlation between regions between the variable proportion of CHC and the Covid-19 cases that have occurred, it is necessary to carry out a spatial autocorrelation test so that the morans index is known.If the Morans index (I) is greater than the basic index (Io), it means that there is spatial autocorrelation between regions.The value of Io in the sample is calculated using the equation :  () = − 1 −1 with nilan n are 21 sub-districts in Magelang District.In the calculation, the value of Io is 0.05 (Figures 2 and 3).The existence of CHC throughout Magelang District shows a statistically significant spatial pattern, this is evidenced by the value on the spatial autocorrelation test with an indication of the Morans index of 0.089 or above the Io value of 0.05 (I > Io) (Figure .2).The positive value formed from this test indicates that the existence of the Puskesmas tends to be clustered spatially and can be interpreted that there is a similar pattern or the same pattern between neighbouring sub-districts in Magelang District.This condition is possible because the existence of CHC in each sub-district area tends to be close to and is in the center of the sub-district economy and close to the sub-district office, markets, schools and residential centers.  .3).The results of the morrans I global bivariate autocorrelation test show a value of 0.248, which means that there is an indication of a spatial autocorrelation between the proportion of the number of CHC and the Covid-19 Incidence in Magelang District.The I value resulting from the test is far above the Io value (I > Io).The results of the autocorrelation analysis show a tendency for spatial autocorrelation between variables, to see how much influence the autocorrelation has, the analysis can be continued at the stage of the spatial regression test between variables.  1 is the final stage in the spatial regression analysis model that uses analytical techniques with the spatial lag regression analysis method.The results of the analysis show that there is influence (W-y variable) between sub-districts in Magelang District with a coefficient value of -0.175 and a p value of 0.569.A negative value on the coefficient as an indicator showing an increase in Covid-19 cases in a sub-district area tends not to contribute to the surrounding area.On the other hand, the p-value of Covid-19 events (0.0022) or below the p value (p <0.005) indicates that the incidence of Covid-19 has a significant effect on the proportion of Community Health Centers.Further analysis was carried out using local indicator of spatial association (LISA) analysis techniques.LISA analysis is used to see the types of relationships between regions locally (Figure 4).The LISA Cluster map identifies 21 sub-districts in Magelang District where in the visualization of the cluster map that is formed, it can be seen that 1 sub-district has a high-high relationship type (Ngluwar) and 1 sub-district has a low-high relationship type (Mungkid).Meanwhile, the LISA significant map depicts that there are two sub-districts with a high level of significance (p <0.05), namely Ngluwar and Mungkid sub-districts.The remaining 19 sub-districts do not have a significant level between regions.
Covid-19 is spread evenly in 21 sub-districts with an average of 1021 cases per sub-district, the highest cases are in the area bordering Temanggung district on the north side and Special Province of Yogyakarta on the south side.The distribution of cases shows that areas close to transportation routes between districts or between provinces are areas with high cases of the spread of Covid-19.These findings are in accordance with research in Finland where in this study areas with high patterns of interaction and bordering other regions are areas that have high socio-economic and transportation interactions both in terms of population movements or shifts due to economic reasons [25] [26].In general, Covid-19 cases in Magelang District have decreased from year to year, initially in 2020 it became the second highest in Central Java Province after Semarang City, then began to decline in 2021.However, the condition of Magelang District which is among the big cities in Indonesia, namely Semarang and Yogyakarta, have made Magelang District an area that deserves attention because of its vulnerability to disease pandemics.For this, a strong basic health system is needed with the Puskesmas as the determining factor in it [27].
Spatial univariate and bivariate correlation tests Morrans index indicates a positive autocorrelation where in the univariate test the proportion of puskesmas variables (Morrans I = 0.089, with Io = 0.05) in the Magelang District area.This finding is reinforced by the results of the spatial lag regression analysis which shows that there is a spatial autocorrelation between the proportion of CHC and the incidence of Covid-19 (Coefficient= -0.175; p=0.059) and the variable proportion of CHC has a significant effect on the incidence of Covid-19 (p=0.0002) .These results show a link between the proportion of the number of CHC and Covid-19 cases.In other words, the results of the analysis explain that the number of Covid-19 cases that occur in each sub-district is related to the proportion of the number of Community Health Centers.A similar condition occurred in study A in the Thailand region which showed an autocorrelation between basic health services in that country and the incidence of the Covid-19 infectious disease [28].The condition of the autocorrelation that is formed and shows that there is a correlation can occur considering that during the pandemic the basic health service system that could reach patients was the Puskesmas, the uncertainty and confusion of patients due to the pandemic situation made the Puskesmas the official health system service provided by the government making the Puskesmas the most trusted health service in society.
Further analysis on a local area basis was carried out using the LISA method.The results of the LISA cluster map analysis show that the Ngluwar sub-district has a high-high correlation type and the Mungkid sub-district has a low-high correlation type, 19 other districts in the analysis were found to be insignificant.These results indicate that the two sub-districts have a significant correlation for the variable proportion of CHC with the incidence of Covid-19 in their areas.This type of high-high relationship can be interpreted that the Ngluwar sub-district area is in a situation of high significance and is surrounded by areas with the same indications.Meanwhile, the low-high relationship type means that the Mungkid District area is an area with low significant relationship indications but is surrounded by areas with a high level of significance.

Conclusion
This research indicates spatially geographical heterogeneity and autocorrelation between regions on the variable proportion of CHC with the incidence of Covid-19 in Magelang District, Central Java Province.The proportion of the number of CHC has a spatial correlation with the incidence of Covid-19 in the research object area, however, there are only 2 sub-districts that are indicated to be significant, namely Ngluwar sub-district and Mungkid sub-district.This level of significance is explained by the existence of a significant relationship between the proportion of CHC and the incidence of Covid-19.
Ngluwar sub-district is a sub-district with a high-high regional local correlation type and Mungkid subdistrict is a sub-district with a low-high regional local correlation type.
Puskesmas as the closest basic health service to the community has a spatial autocorrelation with the incidence of Covid-19.For this reason, attention is needed to systematic and measurable program planning and evaluation to optimize the health system in Magelang District.The balance of basic health services and population can be the basis for improving the quality of health in an area and can minimize disparities between regions so that the development of basic health service facilities can be more targeted

Figure 2 .
Figure 2. Global Spatial Univariate Autocorrelation of Health Facility Services

Figure 3 .
Figure 3. Global Spatial Bivariate Autocorrelation of the Proportion of Community Health Centers and the Incidence of Covid-19 in Magelang District

Figure 4 .
Figure 4. LISA Cluster Map and LISA Significant Map Furthermore, a bivariate autocorrelation test was carried out between the 2 variables involved in the study, namely the variable proportion of the number of CHC with Covid-19 Incidence in Magelang District (Figure 6

Table 1 .
The Results of the Spatial Autoregression Analysis of the Spatial Lag Model*