SPATIAL ANALYSIS OF DENGUE HEMORRHAGIC FEVER BASED ON INFLUENCING FACTORS IN JOMBANG, 2014–2018

1 Faculty of Public Health, Universitas Airlangga, retno_trihastuti@yahoo.com 2 Department of Epidemiology, Faculty of Public Health, Universitas Airlangga, lucia-y-h@fkm.unair.ac.id Correspondence Author: Lucia Yovita Hendrati , lucia-y-h@fkm.unair.ac.id, Departement of Epidemiology Faculty of Public Health, Universitas Airlangga, Dr. Ir. H. Soekarno Street, Mulyorejo, Surabaya City, East Java, 60115, Indonesia


INTRODUCTION
Dengue Hemorrhagic Fever (DHF) is a disease that threatens human health. It is caused by the dengue virus, which is classified as an arthropod-borne virus of the genus Flavivirus and the family Flaviviridae. It is transmitted by mosquito bite (genus Aedes, especially Aedes aegepty). The genus has four serotypes: Den-1, Den-2, Den-3, and Den-4. In addition to DHF, Aedes mosquitoes can carry other diseases such as chikungunya, yellow fever, and the Zika infection (Ministry of Health RI, 2017).
DHF is widespread throughout the tropics and sub-tropics, with risk factors affected by rainfall, temperature, relative humidity, urbanization rate, and quality of vector control services. Before 1970, only nine countries were experiencing a DHF epidemic. Today, the disease is endemic to more than 100 countries in the African, American, Eastern Mediterranean, Southeast Asian, and Western Pacific regions. Symptoms arise 3-14 days after an infective bite, and DHF affects people of all ages, including infants, children, and adults. The symptoms are a mild to high fever accompanied by severe headaches, pain in the back of the eyes, muscle and joint aches, and rashes. No specific anti-virus medication for DHF is available (World Health Organization, 2017).
Based on the Jombang Distritcal Health Office (2018), there were 204,171 cases of DHF in 2016 at Jombang District, with the number of deaths totaling 1,598. In 2017, the number of cases decreased dramatically to 68,407 with a total of 493 deaths, and in 2018, there were 65,602 cases with 467 deaths. The decrease in the incidence rate (IR) of DHF in 2017 compared with 2016 was quite high, from 78.85 to 26.10 per 100,000 population, with a further decrease in 2018 to 24.75 per 100,000 population.
Jombang District is an endemic area for DHF, and the disease often appears as an outbreak with relatively high morbidity rates. The East Java province's IR target is 49 per 100,000 population in line with the national target. In 2016, Jombang District's IR reached 91.6 per 100.00 population. In 2017, it decreased to 28.10 per 100,000 population, but in 2018, it increased again to 41.95 per 100,000 population. Efforts to prevent DHF cases have been carried out, including mosquito eradication, support of three pillars for the implementation of PSN in various regions, traveling broadcasts to carry out mosquito eradication, counseling about handling DHF, and distributing circulars about early precaution systems to related institutions. These control efforts were carried out in 2017, but the DHF IR continued to increase in 2018 (Jombang Distritcal Health Office, 2018).
Increased cases occur due to climate change, which adds to the dengue vector habitat, thereby increasing the risk of transmission. In addition to endemic DHF, Jombang District is estimated to have other risk factors that support the occurrence of the disease, including population density, larvae free index, rainfall, coverage of healthy homes, and healthy lifestyle coverage (PHBS) (Jombang Distritcal Health Office, 2018).
The aim of this study was to spatially analyze various factors simultaneously (multivariate analysis) in relation to the IR of DHF in Jombang District during the period 2014-2018. One of the approaches to prevention of DHF utilizes spatial statistics with the aim of unveiling any factors that underlie the occurrence of geographical phenomena related to DHF cases. The difference between this study and previous research is that there additional variables were applied in the spatial test using GeoDa: the coverage of healthy home, and the PHBS coverage.

METHODS
This study was conducted as observational research, and the design of this research was an ecological study. The research used profile data from the provincial health Department of Jombang District. Sources of data regarding DHF IRs, larvae free index, coverage of healthy homes, and coverage of PHBS Behaviors for 2014-2018 were obtained from Profile Jombang District Health Office, while rainfall and population density data for the same period were retrieved from the Indonesia Statistics Agency.
The population sample consisted of 21 subdistricts in Jombang District during the period 2014-2018. The sample was taken by using the total population. The independent variables were the average larvae free index, the average rainfall, the average population density, and the coverage of PHBS for 2014-2018, as well as the coverage of healthy houses (average 2014-2015).
The dependent variable was the average IR in 2014-2018. The average IR was defined by identifying all patients who had been diagnosed with DHF by a doctor as recorded at the District Health Office Jombang in each year, dividing by the population of each sub-district (per 100,000 residents), and dividing over five years for 2014-2018. The average larvae free index was defined as the number of houses that were free of larvae per the number of houses or buildings inspected in each year, multiplied by 100%, and divided over five years. Average rainfall was defined as the average annual rainfall in the study site environment in Jombang district each year (2014-2018) and divided over five years. The average population density was defined by taking the number of residents living in the urban area per square kilometer in 2014-2018 and dividing over five years. Coverage of healthy houses was defined by residences that met health requirements, including building and structuring homes based on physical and biological requirements. The method of calculation was the number of healthy homes in an area within each year against the number of existing homes in the area, multiplied by 100% and divided for five years. In this variable, there was no categorization because the data used in accordance with the GeoDa application were ratio data.
Data were analyzed by a multivariate approach with classic regression. The variables analyzed were various factors related to the IR of DHF cases based on spatial analysis in Jombang District. The study was conducted in August-September 2019. The research location was the working area of Jombang District Health Office. The research was carried out using spatial analysis through GeoDa. This research received a permission that granted by the ethics committee of the Faculty of Dental Health (No. 509/HRECCFODM/VII/2019). Table 1, DHF cases in Jombang District in 2014-2018 tended to fluctuate. The highest number of cases occurred in 2016 (1,142 cases), while 2017 saw a decline of 69%, with the number of cases dipping to 351. However, in 2018, cases of DHF increased again by 50% from the previous year. Table 2 shows that, for average IRs across the five years, the highest value was in Diwek subdistrict, at 92 per 100,000 inhabitants, and the lowest was in Ngusikan sub-district, at 14 per 100,000 inhabitants. The highest average value for larvae free index across the five years was in the Perak sub-district, at 89%, while the lowest was in the Diwek sub-district, at 50%.

According to
The highest average population density across the five years was in the Jombang sub-district, with 3,983 people/km², and the lowest density was in the Wonosalam sub district, with 261 people/km². The highest average rainfall value across the five years was in Wonosalam subdistrict, at 207 mm/year, and the lowest rainfall level was found in Jogoroto District, at 113 mm/year (Table 2). Table 3 shows the results of GeoDa analysis per variable. A p-value of 0.05 was calculated for the larvae-free and PHBS variables, which shows the relationship of these variables to DHF. For the variables of population density, rainfall, and healthy home coverage, there was a p-value of 0.05, which indicates that there was no correlation between these variables and DHF. The results of the multivariate analysis returned a value of p < 0.05, indicating a connection between all the variables and DHF.
For the variable of average healthy home coverage across the five years, the highest value was in the Sumobito sub-district, at 90%, and the lowest was in the Plandaan sub-district, at 27%.
For the average PHBS coverage across the five years, the highest value was in the Jogoroto subdistrict, at 82%, and the lowest was in the Wonosalam sub-district, at 15%.

Incident Rate
Jombang District is a region with a relatively number of DHF cases. This district's IR target of 49 per 100,000 population is in line with the East Java Province incident rate. There were eight subdistricts with an average case rate of DHF higher than 49 per 100,000 population during 2014-2018: Diwek, Jogoroto, Ngoro, Peterongan, Sumobito, Jombang, Bareng, and Mojowarno. The lowest incidence rates were recorded in Ploso, Wonosalam, and Ngusikan.

Population Density
Population density was defined as the number of residents per square kilometer of area. Population density is the ratio between the number of inhabitants and the area (units = km²) (Subekti & Islamiyah, 2017). Based on the analysis of this study, there was no relationship between the IR and the population density. This is in line with another study that was conducted in Yogyakarta, which found that population density was not related to the incidence of DHF (Setiawan, Supardi, & Bani, 2017). Other studies indicating no correlation between population density and DHF cases include research in Malayang subdistrict, Manado City (Paomey, Nelwan, & Kaunang, 2019) and in Banjarmasin (Kasman & Ishak, 2018).
Population density in downtown areas tends to be higher than in suburban areas (Handayani, Fannya, Roza, & Angelia, 2015). The denser the population, the easier it is for DHF to be transmitted. This is because the mosquito's flight distance is 50 meters; therefore, transmission can spread more quickly (Fatati, Wijayanto, & Sholeh, 2017). Another factor that explains why population density has no impact on the incidence of DHF is population mobility, which was not measured in this study. Both Murwanto, Trigunarso, & Purwono (2019) and Taryono, Ispriyanti, & Prahutama (2018) report that population mobility is a factor that influences the incidence of DHF.

Rainfall
Rainfall refers to the amount of rainwater that falls in an area in a certain time period. High rainfall will indirectly lead to high levels of mosquito breeding as it leads to more mosquito breeding locations being formed (Azhari, Darundiati, & Dewanti, 2017). Current results show no relationship between rainfall level and the DHF IRs, as supported by a study in Tanah Datar, which also found that rainfall did not show a significant relationship with DHF cases (Masrizal & Sari, 2016). Other research showing that rainfall does not affect DHF cases has been conducted in Kenali Besar of Jambi City (Hardianti, 2017) and Ternate (Tomia, Hadi, Soviani, & Retnani, 2016).
In the rainy season, the population of Aedes aegypti increases because eggs that have not yet hatched begin to do so when the breeding habitat starts to fill with rainwater. This condition will also increase the mosquito population and thus cause an increase in the transmission of DHF (Ministry of Health RI, 2017). Research in Bandung confirms that high rainfall provides ideal locations for breeding larvae (Fitriana & Yudhastuti, 2018), however in Nisaa, Hartono, & Sugiharto (2016) study stated that rainfall can sweep mosquito breeding places; thus, rainwater and flooding can cause the disappearance of mosquito breeding sites (Ayumi, Iravati, & Umniyati, 2016). This process possibly indicates why there is no relationship between rainfall and the IR of DHF.

Coverage of Healthy Houses
A healthy house is defined as a house that meets health requirements (Jombang Distritcal Health Office, 2018). The sanitation condition of the home environment is related to an increase in regional dengue cases. This is because of the low awareness of environmental sanitation, which leads to an increase in the presence of Aedes aegypti mosquito breeding places in the home environment (Sitepu, Nasution, Supriyadi, & Depari, 2018). Indicators of a healthy home include assessments of ventilation, lighting, and air quality. Ventilation is closely related to air humidity-good ventilation helps to maintain air temperature and humidity. The movement of mosquito populations in some parts of their habitat is regulated by many factors, including temperature, humidity, the appeal of the house, and places to settle (Sari, Wahyuningsih, & Murwani, 2017).
This study's findings indicate that there is no relationship between the coverage of healthy homes and the IR. This is not in line with another study, in Purwokerto Subdistrict, which found a relationship between the types of houses and dengue cases (Mardiah, Satoto, & Pramono, 2015). In the inspection of healthy homes in Jombang, there are three components investigated based on the form of the Ministry of Health: building and structuring homes based on physical and biological requirements, sanitation facilities, and occupant behavior. The investigations are carried out regularly and accumulated across a year. The test is less than optimal because it does not use lighting and humidity measurement tools. This process indicates a possible reason why no relationship is indicated between healthy home coverage and the incidence of DHF.

Healthy Lifestyle Coverage
Clean living behavior is defined as a set of actions that is practiced based on awareness to be able to play an active role in public health programs (Jombang Distritcal Health Office, 2018). This study's findings show that there is a relationship between PHBS coverage and IR. Previous research in Lowokwaru Sub District, Malang City has reported that there is a relationship between PHBS and the prevalence of DHF (Madeira, Yudiernawati, & Maemunah, 2019). Research in Yogyakarta has also indicated a similar relationship between (Wahyu & Widayani, 2018).

Various Factors Associated with DHF IR
The results of our analysis indicate a relationship of various factors with DHF based on spatial analysis in Jombang. Multiple factors which were rainfall, population density, and PHBS in the spatial analysis affect the incidence of DHF in Jombang. The spatial report per variable also indicates that rainfall and population density did influence the spatial analysis. Different results were obtained from the bivariate and multivariate analyses. This was because the multivariate analysis involved analyzing all factors simultaneously in relation to the incidence of dengue, and in the bivariate analysis, a partial review of DHF occurrence was conducted. Other factors that are assumed to have an influence include mobility, which was not measured in this study. The findings are in line with research conducted in Blitar Sub-district, with Moran's I test results indicating a p-value of 0.53 (p > 0.05), which means that there was not a relationship between the various factors and the incidence of DHF (Maryanto, 2019).
The strength of this study is that the data used were valid data from the health department in Jombang for 2014-2018. We applied six variables that were suspected to be factors in the spread of DHF. The results suggest that the variables of larvae free index and PHBS coverage do affect DHF rates. This research is intended be used as input for relevant agencies to help reduce and break the chain of transmission of dengue disease. Contributions to this research could be used by the Office of Health or related agencies concerned with handling of the problem of dengue cases intensively by plotting the chain of transmission of disease incidence. The recommended activities to be carried out are eradicating mosquito nests by draining and scrubbing water reservoirs, closing water reservoirs, utilizing or recycling used items that can collect rainwater (3M) thoroughly, effectively, simultaneously, routinely, and continuously, and larvae control via administering larvicide routinely in areas that are difficult to water or to drain.

Research Limitations
In this study, the main limitation was the rainfall data recorded in Jombang District. Three sub-districts (Bandar Kedung Mulyo, Ngusikan, and Plandaan) did not have rainfall registers, so their data values were entered as 0.

CONCLUSION
Most of the average IRs of DHF cases in Jombang during the period 2014-2018 exceeded the East Java Province IR target. Based on the results of the bivariate analysis, there are connections between IR and the variables of larvae free index and PHBS. The variables of population density, rainfall, and coverage of healthy homes show no correlation with IR. Various spatial factors are simultaneously related to IR, even though there are only two in variables related to IR according to the bivariate analysis.

CONFLICT OF INTEREST
In this study, there were no conflicts of interest declared by any parties involved.

AUTHOR CONTRIBUTIONS
RT and LYH conceived the study. RT led on statistical analysis and drafting the manuscript. LYH led on interpretation of the findings. All authors read and approved the final version of the manuscript. All authors have agreed to authorship and order of authorship for this manuscript.