District-Level Inequalities in Hypertension among Adults in Indonesia: A Cross-Sectional Analysis by Sex and Age Group

Background: An estimated 1.28 billion adults 30–79 years old had hypertension globally in 2021, of which two-thirds lived in low- and middle-income countries (LMICs). Previous studies on geographic and socioeconomic inequalities in hypertension among adults have limitations: (a) most studies used individual-level data, while evidence from locality-level data is also crucial for policymaking; (b) studies from LMICs are limited. Thus, our study examines geographic and socioeconomic inequalities in hypertension among adults across districts in Indonesia. Methods: We combined geospatial and quantitative analyses to assess the inequalities in hypertension across 514 districts in Indonesia. Hypertension data were obtained from the Indonesian Basic Health Survey (Riskesdas) 2018. Socioeconomic data were obtained from the World Bank. Six dependent variables included hypertension prevalence among all adults (18+ years), male adults, female adults, young adults (18–24 years), adults (25–59 years), and older adults (60+ years). Results: We also found significant geographic and socioeconomic inequalities in hypertension among adults across 514 districts. All hypertension indicators were higher in the most developed region than in the least developed region. Districts in the Java region had up to 50% higher prevalence of hypertension among all adults, males, females, young adults, adults, and older adults. Notably, districts in the Kalimantan region had the highest prevalence of hypertension, even compared to those in Java. Moreover, income level was positively associated with hypertension; the wealthiest districts had higher hypertension than the poorest districts by up to 30%, but only among males and older adults were statistically significant. Conclusions: There were significant inequalities in hypertension among adults across 514 districts in the country. Policies to reduce such inequalities may need to prioritize more affluent urban areas and rural areas with a higher burden.


Background
The World Health Organization (WHO) estimated that 1.28 billion adults 30-79 years old had hypertension globally in 2021, of which two-thirds lived in low-and middle-income countries (LMICs) [1]. It is a serious medical condition of elevated blood pressure that increases the risks of diseases such as heart, brain, and kidney [1]. The latest Global Burden of Diseases study found that high blood pressure was the top leading risk of death and disability among adults in 2019 [2], which may contribute to ischemic heart disease, stroke, and chronic kidney diseases being among the top ten leading causes of deaths and disability in the same year [3]. Moreover, the economic burden is substantial. A recent study from Ethiopia showed total productivity loss due to premature mortality and morbidity was over USD 449,000, and the overall economic burden of hypertension was over USD 514,000 (or USD 106 per person per month) [4].
Indonesia is the fourth most populated country, with over 276 million people in 2021. Like many LMICs, Indonesia is a lower-middle-income country with an increasing burden of hypertension. The nationally representative surveys of the Indonesia Basic Health Survey (Riskesdas) found that hypertension among adults 18+ years old increased rapidly from 25.8% in 2013 to 34.1% in 2018 [5]. The latest national-level Global Burden of Study found that high blood pressure was the top risk factor attributable to deaths and disabilities in Indonesia, which may contribute to ischemic heart disease and cerebrovascular disease being the first and second leading cases of deaths and disabilities in the country [6].
The relationships between socioeconomic indicators and hypertension among adults have been well-studied, including in LMICs. Busingye et al. [7] conducted a meta-analysis and found that overall, there was a positive association between hypertension and income, while no association with educational status. However, they found that educational status was inversely associated with hypertension in East Asia but positively associated in South Asia. Mishra et al. assessed the socioeconomic inequalities using Nepal Demographic Health Survey data and found that adults from the highest education and income groups were 1.4 times and 1.7 times more likely to be hypertensive than those from the lowest education and income groups [8]. Previous studies have also shown some evidence of geographic inequalities in adult hypertension. Kershaw et al. [9] analyzed participants from six study sites in the United States and found that Blacks born in southern states were 1.11 times more likely to be hypertensive than non-southern states (findings were not significant for whites). Morenoff et al. [10] analyzed the Chicago Community Adult Health Study and found that hypertension was negatively associated with neighborhood affluence. Cho et al. [11] analyzed data from Korean National Health Insurance and found that neighborhood deprivation can exacerbate the influence of individual SES on all-cause mortality among patients with newly diagnosed hypertension.
Effective responses to reduce the inequalities in hypertension are crucial to achieving one of the global targets for non-communicable diseases-to reduce the prevalence of hypertension by 33% between 2010 and 2030 [1]. However, previous studies on geographic and socioeconomic inequalities in hypertension among adults have at least two limitations. First, the majority used individual-level data, including studies from Asia, Africa, and Latin America [7,8]. While such studies are essential, evidence from locality-level data (such as districts) is also crucial for policymaking, especially in a decentralized setting such as Indonesia, where some policies are transferred to the district level. Second, previous studies on geographic inequalities are mainly from high-income countries such as the United States and South Korea [9][10][11]. Studies from LMICs such as China and Thailand are limited to analysis using urban/rural or provincial levels [12][13][14]. Thus, our study aims to examine geographic and socioeconomic inequalities in hypertension among adults across 514 districts in Indonesia.

Study Design
Using a cross-sectional study, we analyzed geographic and socioeconomic disparities in hypertension among adults aged 18+ years in Indonesia. Geographic disparities were analyzed using geospatial analyses across 34 provinces and 514 districts. Socioeconomic disparities were assessed using multivariate regression analyses across 514 districts. Hypertension data as the primary dependent variable were obtained from the latest RISKESDAS 2018, a nationally representative health survey. The survey collected information on maternal and child health, nutrition status, communicable and non-communicable diseases and main risk factors, health behaviors, and disability among children and adults [5]. In total, the survey targeted 300,000 households using two-stage sampling. First, the team selected 30,000 census blocks in each urban and rural using probability proportional to size out of a total of 720,000 census blocks in the country. Second, ten households were systematically chosen using implicit stratification of the household head's education. For adults, the survey included 624,563 individuals aged 18+ years [5].

Independent Variables
The main independent variables included region, urban/rural, income, and education level at the district level, obtained from the World Bank database. For the region, we divided provinces and districts into five: Sumatera, Java (including Bali), Kalimantan, Sulawesi, and Papua (including Nusa Tenggara and Maluku). A reference to the provinces and regions is provided in Appendix A. In Indonesia, the western part is generally more developed (especially Java and Bali) than the eastern part (including Papua, Nusa Tenggara, and Maluku) [15][16][17]. In terms of urban and rural, we conducted the analyses using all districts, urban districts (i.e., cities) and rural districts (i.e., regencies). By income level, we grouped district-level poverty rates into five quintiles, with quintile one being the poorest (or highest poverty rates) and quintile five being the wealthiest (or lowest poverty rates). By education level, we grouped the net enrollment ratios of senior secondary into five quintiles, with quintile 1 being the least educated and quintile 5 being the most educated [15][16][17].

Dependent Variables
We used six indicators of hypertension as dependent variables: hypertension among all adults aged 18+ years, male adults, female adults, young adults aged 18-24 years, adults aged 25-59 years, and older adults aged 60+ years. Hypertension was defined as either systolic blood pressure 140+ mmHg, diastolic blood pressure 90+ mmHg, or both. A digital blood pressure monitor was used with respondents in a sitting position. Only two measurements were taken if the difference in blood pressure was less than 10 mmHg; otherwise, three were taken. For each participant, the average (mean) blood pressure was calculated from two measurements with the least difference. We assessed the prevalence by sex to observe variations for males and females. We evaluated the prevalence by age category to observe variations among young adults, adults, and older adults, which is crucial for better targeting NCD control and prevention efforts, including reforms toward effective health systems in Indonesia and other LMICs [18].

Data Analysis
For geospatial analyses, we divided the prevalence of hypertension among 34 provinces and 514 districts by quintile using ArcMap 10. For multivariate regression analysis, we performed Ordinary Least Square (OLS) models using STATA 15 to examine the associations between geographic indicators such as urban/rural and region and between socioeconomic indicators such as income and education level and each hypertension indicator: hypertension among all adults, male adults, female adults, young adults, adults, and older adults. We also calculated absolute and relative differences for the geographic and socioeconomic variations. We compared the differences between the most developed (the Java region) and the least developed region (the Papua region). We compared the differences between quintile 1 (poorest or least educated) and quintile 5 (wealthiest or most educated). All statistical significance was at the 5% level or lower. Figure 1 shows the prevalence of hypertension among adults by quintile at the province level. In panels a-f, hypertension among all adults ranged from 23.8% to 45.5%; that among male adults ranged from 23.9% to 42.4%; that among female adults ranged from 23.5% to 48.6%; that among young adults ranged from 8.3% to 21.9%; that among adults ranged from 24.0% to 46.0%; that among older adults ranged from 49.7% to 77.6%. Among all adults, hypertension was highest (quintiles 4-5) in all provinces in Kalimantan, most provinces in Java (except for Banten province), and some in Sulawesi (e.g., North Sulawesi and West Sulawesi). In Kalimantan, this patterning was similar in other indicators, including hypertension among males, females, young adults, adults, and older adults. In Java, the patterning was similar in all other indicators except among older adults, with only West Java having the highest prevalence. By sex, additional provinces with the highest prevalence (quintiles 4-5) include Bali for males and Lampung and South Sulawesi for females. By age group, additional provinces with the highest prevalence (quintiles 4-5) include Banten and Papua for young adults, Gorontalo for adults, and Riau Islands, Bangka Belitung, and Gorontalo for older adults.   Table 1 shows the prevalence of hypertension among adults by province. The top and bottom boxes show the ten wealthiest and poorest provinces, respectively. The grey-shaded cells show a prevalence higher than the national average for each column of the hypertension indicator. Five of the ten wealthiest provinces (including South Kalimantan, Central Kalimantan, North Kalimantan, East Kalimantan, and Jakarta) had consistently higher than average for at least five indicators. In contrast, none of the ten poorest provinces did.  Table 2 shows the descriptive statistics of districts in our analysis, including the prevalence of hypertension among adults. Of 514 districts, 97 (18.9%) were urban cities, and 417 (81.1%) were rural regencies. Urban cities were mainly in Java (36.1% of 97) and Sumatera (34.0%). Rural regencies were less concentrated, including 29.0% (of 417 regencies) in Java, 22.3% in Sumatera, 20.6% in Papua, 16.8% in Sulawesi, and 11.3% in Kalimantan). By the level of income, 79% of urban areas were wealthier (quintiles 4-5), while nearly half (47.2%) of rural areas were poorer (quintiles 1-2). By the level of education, 71.1% of urban cities had higher education (quintiles 4-5), while nearly half (46.8%) of rural regencies had lower education (quintiles 1-2). Regarding the dependent variables, the prevalence of hypertension was 33.3% among all adults, 30.4% and 36.0% among males and females, and 12.9%, 32.6%, and 63.2% among young adults, adults, and older adults, respectively. Compared to rural areas, hypertension among males, adults, and older adults was significantly higher in urban areas but significantly lower among females. Hypertension among males, adults, and older adults was 32.6%, 34.0, and 66.2% in urban areas and 29.9%, 32.3%, and 62.5% in rural areas. Hypertension among females was 34.6% and 36.4% in urban and rural areas.  Figure 2 shows the prevalence of hypertension by quintile at the district level, showing more granularity than at the provincial level. For instance, many districts in Aceh, North Sumatera, Riau, South Sumatera, Lampung, Bali, East Nusa Tenggara, West Nusa Tenggara, Central Sulawesi, Southeast Sulawesi, and Papua provinces had the highest prevalence of hypertension (quintiles 4-5) among all adults. In contrast, several districts in West Kalimantan and Central Kalimantan had a lower prevalence of hypertension (quintiles 1-2). This patterning is similar for hypertension among males, females, young adults, adults, and older adults.  In terms of socioeconomic disparities, Appendices C and D provide ten districts with the lowest and highest prevalence of hypertension among adults, respectively. For all adults, the prevalence of hypertension ranged from 9.7% in Nduga regency (Papua province) to 53.2% in Hulu Sungai Tengah (Papua). By sex, hypertension among males ranged from 11.0% in Nduga (Papua) to 51.1% in Kutai Barat (East Kalimantan); hypertension among females ranged from 8.0% in Nduga (Papua) to 57.2% in Ciamis (West Java). By age group, hypertension among young adults ranged from 1% in Buton Tengah (Southeast Sulawesi) and Mentawai Islands (West Sumatera) to 37.6% in Pegunungan Bintan Yalimo (Papua); that among adults ranged from 9.8% in Nduga (Papua) to 52.8% in Kutai Barat (East Kalimantan); that among older adults ranged from 0% in Yahukimo, Pegunungan Bintan, and Nduga (Papua) to 100% in Diyai (Papua). By urban/rural, all districts with the lowest prevalence of hypertension for all adults, by sex, and by age groups were rural. Similarly, most districts with the highest prevalence of hypertension for all adults by sex and age groups were rural. By income, the average poverty rates among the ten districts with the highest prevalence of hypertension were up to 14%, while the rates among the districts with the lowest prevalence were up to 35%. Table 3 shows the associations between geographic and socioeconomic indicators (i.e., region, income, and education) and hypertension. The absolute (relative) values indicate the difference (ratio) between the most (Java and Bali) vs. the least (Papua, Nusa Tenggara, and Maluku) developed regions, the wealthiest (quintile 5) and poorest (quintile 1) districts, and the most educated (quintile 5) and least educated (quintile 1) districts. By region, districts in the most developed region had a significantly higher prevalence of hypertension among all adults, males, females, young adults, adults, and older adults, compared to those in the least developed region. Districts in Java had 45%, 40%, 50%, 29%, 40%, and 27% higher prevalence of hypertension among all adults, males, females, young adults, adults, and older adults, respectively. However, districts in the Kalimantan region had the highest prevalence of hypertension among all adults, by sex, and by age group, compared to districts in all other regions, including Java. By income, the wealthiest districts had a higher prevalence of hypertension among all adults, by sex, and by age group than the poorest districts. However, only hypertension among males and older adults was statistically significant-the wealthiest districts had a 30% and 24% higher prevalence among males and older adults. By education, the associations were mixed but mostly not significant except for hypertension among young adults, which was significantly higher in the least educated districts compared to the most educated ones. The least educated districts had a 22.0% (i.e., 1/0.82 = 1.22) higher prevalence of hypertension among young adults. Results were similar in the urban and rural subgroup analyses. n/a n/a n/a n/a n/a n/a 32. Note: Q = Quintile; Java region includes Bali; Papua region includes Maluku and Nusa Tenggara. Income quintile used the district-level poverty rate (e.g., Q1 = 20% of districts with the highest poverty rate). Absolute (Relative)-Difference (Ratio) between Papua and Java as well as Q1 and Q5. For education, absolute (relative) was between Q1 and Q5 except among urban (Q2 and Q5). Boldface values show statistical significance at a 5l (see Appendix E for the regression outputs).

Discussion
We found a high prevalence of hypertension among adults 18+ years in Indonesia in 2018. The prevalence of hypertension was 33.3%, 30.4%, and 36.0% among all adults, males, and females, respectively. By age, the prevalence was 12.9%, 32.6%, and 63.2% among young adults (18-24 years), adults (25-59 years), and older adults (60 years and over), respectively. The findings are similar to the global the global estimates of agestandardized hypertension prevalence in adults 30-79 years of 32% in women and 34% in men in 2019 [19].
We also found a significant geographic and socioeconomic disparity in hypertension among adults across 514 districts in Indonesia. By urbanicity, while overall hypertension was generally higher in urban areas in Indonesia, we found mixed results by sex. Hypertension among males was significantly higher in urban areas (32.6% in urban vs. 29.9% in rural), but that among females was higher in rural areas (34.6% in urban vs. 36.4% in rural). This evidence aligns with a study in Turkey that found that women were more likely to be hypertensive in rural areas than in urban areas [20]. However, other studies from Nepal and Ghana found that hypertension among female adults was higher in urban areas [8,21]. Moreover, at the district level, while all districts with the lowest hypertension for all adults, by sex, and by age groups were rural, many districts with the highest prevalence were also rural. Thus, effective responses to reduce disparity in hypertension may need to prioritize not only urban areas but also rural areas with an already high burden of hypertension [22][23][24].
By region, all hypertension indicators were higher in the most developed region (i.e., the Java region, including Bali) than in the least developed region (e.g., the Papua region, including Maluku and Nusa Tenggara). Similarly, by income, the wealthiest districts had higher hypertension among all adults, by sex, and by age group than the poorest districts (although only among males and older adults was statistically significant). All this finding aligns with previous studies from LMICs. Studies on geographic variations across 31 provinces in China found that hypertension was higher in more developed areas such (e.g., Beijing and Shanghai) than in less developed areas such as (e.g., Hainan) [12,13]. In addition, a study across 76 provinces in Thailand found that hypertension was higher in Bangkok and metropolitan areas and lower in the northeast and southern provinces [14]. In contrast, studies from high-income countries such as the United States and South Korea found that hypertension was higher among less developed areas or neighborhoods [9][10][11].
For policy, hypertension is increasing among young adults and is already high among the adult population in the country, which is likely to produce a substantial economic burden from total productivity loss due to premature mortality and morbidity [4]. Also, the hypertension burden among older adults is very high. All this indicates the need for health systems reform towards improved prevention of non-communicable diseases and their main risk factors, especially hypertension. Reforms may include changes from the community to primary care and secondary care and integration with infectious disease platforms [25][26][27]. By region and socioeconomic status, effective responses to reduce inequalities in hypertension may need to prioritize more affluent urban areas and rural areas with higher hypertension burden and other risk factors for non-communicable diseases [28][29][30][31][32][33].
To the best of our knowledge, our study is the first in LMICs to examine geographic and socioeconomic inequalities in hypertension among all adults, males, females, young adults, adults, and older adults across many local units (over 500 districts). However, our study also has at least two limitations. First, we did not have information on ethnicity in our dataset, which limits our sub-group analysis by that variable [34,35]. Secondly, we used cross-sectional data and could not assess trends over time. Despite these limitations, our findings are highly relevant to health policies in Indonesia and other LMICs.

Conclusions
In Indonesia, hypertension prevalence was highest among females (36.0%) and older adults 60+ years (63.2%). We found significant geographic and socioeconomic inequalities in the prevalence of hypertension among adults across 514 districts. Hypertension was higher in the most developed region than in the least developed region. Districts in the Java region had up to 50% higher prevalence of hypertension among all adults, males, females, young adults, adults, and older adults. Notably, districts in the Kalimantan region had the highest prevalence of hypertension, even compared to those in Java. Moreover, income level was positively associated with hypertension; the wealthiest districts had higher hypertension than the poorest districts by up to 30%, but only among males, and older adults were statistically significant. Policies to reduce such inequalities may need to prioritize more affluent urban districts and rural areas with a higher burden.

Conclusions
In Indonesia, hypertension prevalence was highest among females (36.0%) and older adults 60+ years (63.2%). We found significant geographic and socioeconomic inequalities in the prevalence of hypertension among adults across 514 districts. Hypertension was higher in the most developed region than in the least developed region. Districts in the Java region had up to 50% higher prevalence of hypertension among all adults, males, females, young adults, adults, and older adults. Notably, districts in the Kalimantan region had the highest prevalence of hypertension, even compared to those in Java. Moreover, income level was positively associated with hypertension; the wealthiest districts had higher hypertension than the poorest districts by up to 30%, but only among males, and older adults were statistically significant. Policies to reduce such inequalities may need to prioritize more affluent urban districts and rural areas with a higher burden.     n/a n/a n/a n/a n/a n/a