Abstract
Diabetes Mellitus (DM) is a significant public health issue in Indonesia, shaped by underlying socioeconomic inequalities that affect its prevalence. This study aimed to comprehensively examine how socioeconomic status, technological accessibility, and environmental conditions influence DM prevalence, screening uptake, and medication adherence in Indonesia. Utilizing data from the Ministry of Health’s 2023 Indonesian Health Survey, a cross-sectional design was employed. The study sample consisted of 951 individuals with DM who met the inclusion criteria. Socioeconomic indicators based on the WHO’s framework for social determinants of health included access to healthcare, employment status, educational attainment, and economic standing. Analyses involved estimating DM prevalence and making comparisons across socioeconomic strata, as well as applying the Relative Inequality Index (RII) to quantify disparities in DM prevalence. Findings revealed a DM prevalence of 3.08% (n = 951), with prevalence increasing across higher wealth quintiles. Although only 3.03% of respondents utilized telemedicine, its use was associated with a significantly higher prevalence of DM. Conversely, individuals without telephone access had a lower recorded prevalence, while access to improve water sources and proximity to health centers correlated with reduced DM prevalence. Demographic factors such as lower educational attainment and unemployment were also associated with increased DM prevalence. Strengthening community health centers (Puskesmas) as hubs for early detection and chronic care, while expanding access to digital health tools and culturally sensitive education, is essential for reducing disparities and achieving more equitable diabetes management outcomes.
Plain Language Summary
Diabetes mellitus (DM) is a major health problem experienced by many people in various countries, including Indonesia, where the disease is not only influenced by biological factors but also by socioeconomic disparities. So far, research in Indonesia has only looked at socioeconomic disparities from several factors, whereas a broader analysis and understanding is needed, including access to health services, use of technology, housing and environmental conditions, and social aspects such as inclusion and prevention of discrimination. This study was conducted to understand how these factors affect the prevalence of DM in various community groups. The researchers analysed data from the 2023 Indonesian Health Survey managed by the Ministry of Health. The researchers investigated various socio-economic factors, such as the use of telemedicine, mobile phone ownership, education, occupation, economic status, and access to health services, to see their effect on diabetes. The researchers used the Relative Inequality Index (RII) to determine the difference in the number of people with diabetes in various social and economic groups. The researchers found that 951 people (3.08%) had diabetes. More patients came from higher economic groups. Few patients (3.03%) used telemedicine, which was linked to the high incidence of DM. DM rarely occurred in people who did not have a telephone. People with lower education and unemployment were at higher risk of developing DM. Community health centres were the places most frequently visited by patients. Social and economic differences affect the number of people with diabetes in Indonesia. Patients need to have easy access to health technology services, clean water, and services at health centres. It is important to create fair policies so that everyone can get better health services. Health education and the equitable distribution of health facilities must be improved to reduce differences in the treatment of diabetes.
Introduction
Diabetes mellitus (DM) is a major global public health concern, with Indonesia experiencing substantial challenges due to socioeconomic disparities that affect its prevalence (Kusumaningrum & Ricardo, 2022). Rapid socioeconomic development in the country has resulted in lifestyle transitions, particularly among urban populations. While urban residents generally have better access to healthcare services, they are also more exposed to risk factors such as unhealthy diets and sedentary behavior, both of which contribute to increased diabetes risk (Andriyani et al., 2024; Soewondo et al., 2013; Surendra et al., 2023). Studies suggest that urbanization and the adoption of Western dietary patterns have led to rising obesity rates, a key risk factor for type 2 diabetes mellitus (Soewondo et al., 2013; Tahapary et al., 2015). In contrast, rural communities often have limited access to healthcare and education, which hinders prevention and disease management efforts (Kamilah et al., 2021; Soewondo et al., 2013).
These disparities are also evident in screening participation and treatment adherence. DM screening rates remain low, particularly among individuals of lower socioeconomic status, contributing to delayed diagnosis (Mulyanto et al., 2019; Nopitasari & Ghozali, 2024). Treatment adherence likewise poses a challenge, with non-adherence often driven by low health literacy, inadequate family support, financial constraints, and limited access to healthcare (Handayani et al., 2019; Panduwiguna et al., 2023; Pertiwi et al., 2022).
Socioeconomic status (SES) plays a critical role in determining diabetes risk, as demonstrated by studies linking income, education, and employment to the prevalence of DM (Cleal et al., 2014; Kusumaningrum & Ricardo, 2022). This relationship extends beyond mere correlation and operates through complex behavioral and environmental pathways. The Social Determinants of Health (SDH) framework highlights how SES shapes an individual’s living conditions, which in turn influence access to healthy food, stress levels, and health literacy (Marmot & Allen, 2014). Lower education levels are often associated with reduced health literacy, which increases the likelihood of consuming inexpensive processed foods and neglecting preventive care (Shrestha et al., 2016) Paradoxically, higher income levels may contribute to unhealthy behaviors, including high-fat diets and physical inactivity a phenomenon referred to as the “disease of affluence” (Kusumaningrum & Ricardo, 2022; Mayawati et al., 2021). These mechanisms illustrate that SES influences diabetes risk through multidimensional, non-linear channels.
Environmental conditions also mediate the effects of SES on diabetes outcomes (Mutyambizi et al., 2019; Seifu et al., 2021). Economically disadvantaged populations often reside in “food deserts,” where access to affordable, nutritious foods is limited, leading to greater reliance on fast food, and increased risk of obesity. Meanwhile, wealthier groups may experience the affluence paradox, where high-sugar and high-fat diets combined with sedentary behavior result in rising diabetes prevalence (Zhang et al., 2020). Therefore, understanding DM requires an integrated perspective on these causal mechanisms, not merely a descriptive correlation.
In this study, SES is conceptualized as a key determinant of diabetes outcomes within the SDH framework. Socioeconomic factors including income, education, and occupation affect individual health behaviors (such as diet, physical activity, and medication adherence) and shape environmental exposures (such as healthcare access, food security, and neighborhood conditions). These processes collectively influence DM prevalence, screening participation, and treatment outcomes (Marmot & Allen, 2014), and guide the selection of variables in our analysis.
While previous research in Indonesia has identified some of these determinants, there is limited understanding of how socioeconomic and environmental factors interact in shaping disparities in diabetes screening and treatment adherence (Budiharsana, 2017; Kusumaningrum & Ricardo, 2022; Mihardja et al., 2014; Mulyanto et al., 2019). Much of the literature focuses narrowly on wealth, education, or age, without capturing the interplay of multiple factors across behavioral and ecological domains (Budiharsana, 2017; Mihardja et al., 2014). A more comprehensive analysis is needed to inform equitable policy interventions. By examining healthcare access, technology use, housing and environmental conditions, and social inclusion, we aim to provide a holistic understanding of the pathways linking SES and diabetes outcomes (Cheng et al., 2025; Johar et al., 2018).
This study aimed to comprehensively analyze how socioeconomic status, technological accessibility, and environmental factors affect the prevalence of diabetes mellitus, screening rates, and medication adherence in Indonesia. Using data from the 2023 Indonesian Health Survey, this study integrated household factors (such as access to water and Sosial Protection Card/Kartu Perlindungan Sosial [KPS] ownership), technology use (such as telemedicine), and access to healthcare facilities to understand how socioeconomic disparities influence DM risk, screening behaviors, and medication adherence. The findings of this study are expected to inform public policy for more equitable interventions, reduce healthcare costs, and improve the quality of life for DM patients in Indonesia.
Methods
Study Design and Population
This study employed a cross-sectional design to analyze secondary data from the 2023 Indonesian Health Survey (IHS), conducted by the Health Policy Agency of the Ministry of Health, Republic of Indonesia. The 2023 IHS is a nationally representative survey that integrates the Basic Health Research (Riskesdas) and the Indonesian Nutritional Status Survey (SSGI). It aims to assess the progress of health development outcomes over the past 5 years and monitor the nutritional status of children under 5 years of age between 2019 to 2024 (Kemenkes, 2023). All data used in this study were fully anonymized by the survey organizers prior to being made available to researchers, ensuring that no individual participant could be identified. Since this research involved no direct contact or intervention with participants, the study design posed minimal risk to individuals.
The design of this study inherently minimized participant risk due to two factors: (1) the absence of direct interaction or intervention between researchers and survey participants and (2) the use of data that had been processed and curated according to established data protection and confidentiality protocols. The potential benefits of this research are deemed to outweigh any minimal risks, as: (1) the findings contribute to the development of more targeted, evidence-based, and equitable health interventions for diabetes mellitus (DM) control in Indonesia and (2) the study enhances scientific understanding of socioeconomic disparities in DM prevalence, screening participation, and treatment adherence, thereby offering critical insights for public health policy formulation.
The original data collection process included written informed consent, with all participants voluntarily agreeing after being informed of the survey’s objectives, benefits, and potential risks. The data collection protocol was approved by the Health Research Ethics Committee of the Ministry of Health under the amended ethics protocol for the health survey (approval No. LB.02.01/I/KE/L/287/2023).
The survey employed face-to-face interviews using both individual and household questionnaires, applying stratified and multistage random sampling techniques. Data collection was conducted between May and July 2023, achieving a high response rate of 91.49%. The survey was implemented at both the national and sub-national levels, reaching down to the district and city levels (Kemenkes, 2023).
The 2023 SKI survey included a total of 315,646 ordinary households (rumah tangga biasa), representing diverse regions across Indonesia. From these households, 877,531 household members and 1,191,692 individual respondents were successfully interviewed. To ensure representativeness and address potential bias, a post-stratification data weighting process was applied by the survey team (Kemenkes, 2023).
For this study, the target population consisted of individuals diagnosed with Diabetes Mellitus (DM). A total of 30,278 individuals were identified with DM based on physician diagnosis. The analytic sample was determined using inclusion criteria, specifically: (1) a confirmed DM diagnosis by a medical professional and (2) complete information on socioeconomic variables. Based on these criteria, the final sample included 951 respondents, representing 3.08% of the DM-diagnosed population.
Measurement
The prevalence rate was defined as the proportion of individuals diagnosed with diabetes mellitus (DM). DM status was determined based on the physician’s diagnosis. This diagnosis was supported by fasting blood glucose measurements taken 2 hrs after glucose loading, using the Accu-Chek rapid detection device. Additionally, HbA1c levels were measured using an autoanalyzer with the immunoturbidimetry method. The screening rate was assessed using the question, “Are you routinely visiting doctors to control your diabetes?” Respondents could select from three options: (1) yes, routinely; (2) yes, occasionally; and (3) no. Responses (1) and (2) were categorized as indicating participation in screening, while response (3) was categorized as not participating. Medication adherence was evaluated through the question, “Are you taking diabetes medicine/injection as advised by the doctor?” with two responses options: (1) yes and (2) no. Respondents who answered “yes” were classified as adherent. Medication adherence was assessed only among individuals diagnosed with DM.
Socioeconomic variables were defined based on the World Health Organization’s (WHO) framework of social determinants of health. These included education, employment, economic status, access to affordable and quality healthcare services, technology use, housing conditions, basic infrastructure, environmental quality, social inclusion, and non-discrimination (WHO, 2025).
The education classification adapted the International Standard Classification of Education (ISCED) from UNESCO, with adjustments to the education system in Indonesia (UNESCO, 2011), which was classified into five categories: no/never to school, primary school, junior secondary school, senior secondary school, and above senior secondary school. Employment status was identified as (1) not working, (2) studying, (3) civil servant, (4) private employee, (5) self-employed, (6) farmer/farm laborers, (7) fishermen, (8) laborer/driver/household helper, and (9) others. Employment data were classified as not working, student, white collar, blue collar, and other.
Economic status refers to the position of an individual or household in relation to commonly accepted indicators of cultural integration, sufficient income, asset ownership, and participation in community activities (SKI, 2023). In this study, economic status was measured using an asset-based ownership index comprising twelve variables: type and power of household lighting, ownership of a washing machine, refrigerator, air conditioner (AC), computer or laptop, flat-screen television (minimum 30 inches), gold or jewelry (at least 10 g), personal vehicle, type of cooking fuel, as well as the type of house ceiling, floor, and wall materials. Each household received a composite score based on these indicators, which was then used to classify households into five economic status groups (welfare quintiles): bottom (the poorest 20% of the population), lower-middle (the next 20%, representing poor and vulnerable groups), middle (20% moderate expenditure), upper-middle (20% with upper higher expenditure), and top (the wealthiest 20%). This classification enabled stratified comparisons of diabetes-related outcomes across socioeconomic gradients (Johar et al., 2018).
Health services utilization is measured based on the health facilities most frequently used by households in the past year, such as community health centers (Puskesmas), private clinics, laboratories, or hospitals. This measurement is done by calculating how often households visit these facilities in a year. The results are then used to determine how often households utilize available health services by comparing the data with the number of households interviewed (Kemenkes, 2023).
Two technological factors were assessed in this study: mobile phone ownership (yes/no) and telemedicine utilization (yes/no). Telemedicine utilization was determined based on participants’ responses regarding their use of online health services within the past year. Housing, basic facilities, and environmental conditions were evaluated using indicators of access to drinking and non-drinking water, which were categorized as improved, unimproved, or surface categories. Social inclusion and non-discrimination were measured through household ownership of a Social Protection Card (Kartu Perlindungan Sosial, KPS), serving as a proxy for access to government support and inclusion in national welfare programs.
Data Analysis
We describe the prevalence of diabetes mellitus (DM) for the overall study population, as well as across subgroups defined by education level, occupation, wealth quintile, healthcare utilization, mobile telephone ownership, telemedicine use, sources of drinking and non-drinking water, and KPS (Social Protection Card) ownership (Kemenkes, 2023). DM prevalence was calculated as the number of diagnosed cases per 100 individuals and was age- and sex-standardized using the direct standardization method, with the survey population serving as the reference population (Kemenkes, 2023). Based on these standardized prevalence estimates, we computed both absolute and relative differences to compare DM burden across the aforementioned subgroups. The prevalence ratio was calculated by dividing the prevalence in the highest exposure group by that in the lowest group (Fleiss et al., 2003), thereby allowing for the assessment of socioeconomic gradients in diabetes prevalence.
The magnitude of socioeconomic inequality in disease prevalence was estimated using the relative inequality index (RII), which offers a comprehensive measure of disparities across the entire socioeconomic gradient. To account for potential confounding by demographic factors, the adjusted Relative Inequality Index (aRII), controlling for age and sex, was also calculated. RII was derived using a generalized linear model (GLM) with a log-binomial link function, treating socioeconomic position as a continuous variable based on a cumulative rank score ranging from 0 (lowest) to 1 (highest), following the method proposed by (Mackenbach & Kunst, 1997). In this study, both age and sex were included as covariates to ensure a more accurate estimation of inequality. In addition, healthcare utilization was analyzed using odds ratios (ORs), which quantify the strength of association between type of health facility use and health outcomes, thereby allowing assessment of how healthcare access influences disease prevalence, screening participation, and treatment adherence.
Statistical assumption checks were performed to ensure the validity of the models. These included assessments of linearity for the socioeconomic rank variable and evaluations of multicollinearity among covariates. To assess the robustness of the findings, sensitivity analyses were conducted using logistic regression as an alternative modeling approach, confirming the consistency of the RII estimates. Robust standard errors were applied to account for potential heteroskedasticity. All statistical analyses were conducted using Stata version 17.0, with a significance level set at p < .05, and 95% confidence intervals reported for all estimates.
Results
The prevalence of diabetes mellitus (DM) was calculated for all individuals diagnosed by a physician, while screening prevalence reflected the proportion of individuals with DM who reported undergoing regular check-ups. Medication adherence was defined as the proportion of individuals who reported taking diabetes medication or insulin injections as prescribed.
Out of 30,278 eligible respondents, 951 individuals were identified as having DM, yielding a prevalence of 3.08% (95% CI [2.79, 3.39]). The mean age of individuals with DM was 56.29 (SD = 10.61), and 61.77% were female (data not shown in Table 1). Among those diagnosed with DM, 733 individuals (77.07%) reported participating in regular screening, and 787 individuals (82.75%) reported adherence to prescribed treatment.
The Complex Relationship Between DM Prevalence and Various Social, Economic, and Health Factors.
Regarding technological factors, 87% of individuals with DM owned a mobile phone, while only 3.03% had used telemedicine in the past year. Screening and medication adherence rates were relatively similar between telemedicine users and non-users, although the overall utilization of telemedicine remained very low.
In terms of household factors, the number of individuals with DM increased with higher wealth quintiles, with the highest prevalence observed in the richest group (28.73%). Similarly, screening and medication adherence were common among individuals in higher economic groups. Notably, 78.75% of people with DM did not possess a Social Protection Card (KPS), and this group also showed higher rates of screening and adherence. A large majority of households had access to improved water sources, 96% for non-drinking water and 98.05% for drinking water.
With respect to sociodemographic characteristics, DM prevalence varied by education level. The highest proportions were observed among individuals with primary education (35.49%) and senior high school education (27.09%). Occupational status also showed variation, with the highest prevalence observed among unemployed individuals (41.51%) and white-collar workers (28.21%).
In terms of healthcare utilization, the most commonly visited facility was the community health center (Puskesmas), accounting for 43.59% of DM-related visits. Puskesmas was also the most frequently reported provider for screening (43.45%) and medication adherence (45.87%), followed by private clinics and hospitals. Detailed distributions for each subgroup are presented in Table 1.
Table 2 provides a comprehensive overview of socioeconomic inequality in the context of Diabetes Mellitus (DM) prevalence, screening, and medication adherence. The Relative Inequality Index (RII) and the adjusted Relative Inequality Index (aRII) serve as metrics for assessing inequality across various factors, such as technological, household, and demographic variables. This examination seeks to clarify the results shown in the table, concentrating on the consequences of these disparities for public health and policy.
Overview of Socioeconomic Disparities in the Context of Diabetes Mellitus (DM) Prevalence, Screening, and Medication Adherence.
Adjusted for age and sex.
p < .001. **p < .01. *p < .05.
The analysis of technological factors revealed different patterns of inequality. After adjusting for age and sex, telemedicine use was strongly associated with a higher prevalence of DM, indicating that individuals who reported using telemedicine were more often found among those with DM. In contrast, mobile phone ownership showed an opposite pattern, where individuals without a mobile phone had a higher prevalence of DM. No significant disparities were observed in screening and medication adherence.
Household inequality is evident in the distribution of wealth quintiles, the ownership of Social Protection Cards (Kartu Perlindungan Sosial/KPS), and the sources of household water. The prevalence of diabetes mellitus among individuals in the lowest wealth quintile was over four times greater than that observed in the wealthiest quintile. Conversely, households with a KPS had a lower prevalence of DM. Inequality was not found for screening and medication adherence. Similar findings were observed for household water sources. Households with unimproved or highly limited water sources paradoxically showed much lower DM prevalence. This result likely does not suggest a protective effect of poor-quality water; instead, it reflects restricted access to diagnostic facilities and a probable high prevalence of undetected diabetes cases in regions with inadequate basic infrastructure.
Demographic inequalities were evident in education and occupation. Lower educational attainment was associated with more than three times higher DM prevalence compared to higher education. However, this inequality did not persist for screening or medication adherence. For occupation, groups with lower socioeconomic status jobs showed a lower prevalence of DM. This pattern may be related to higher case detection among formal workers or higher-SES groups. However, the RII for medication adherence was not significant after controlling for age and sex, suggesting that occupation may not be a critical factor in adherence behavior.
Table 3 displays the odds ratios (OR) and adjusted odds ratios (aOR) for three outcomes: diabetes mellitus diagnosis, screening, and medication adherence, categorized by healthcare utilization. The healthcare utilization categories include “Never,” “Community Health Center (Puskesmas),” “Private Clinic,” and “Hospital.” The odds ratios (OR) reflect the likelihood of being diagnosed with DM, undergoing screening, and adhering to treatment compared to those who never accessed healthcare services. Individuals who accessed healthcare facilities were more likely to be diagnosed with diabetes than those who never visited. The strongest association was identified among patients who visited hospitals, with a more than threefold higher risk of being diagnosed with diabetes, followed by those visiting Puskesmas and private clinics. In terms of healthcare utilization, there was no variance in screening outcomes based on the type of healthcare facility accessed. In contrast, for medication adherence, patients who visited Puskesmas had the highest likelihood of following to their treatment regimen, followed by those who visited private clinics and hospitals.
Healthcare Utilization and Its Impact on Screening and Medication Adherence in Diabetes Mellitus Patients.
Adjusted for age and sex.
p < .001. **p < .01. *p < .05.
Discussion
Analyses of socioeconomic inequality in the context of Diabetes Mellitus (DM) prevalence, screening, and medication adherence have indicated significant disparities that warrant public health and policy attention. The data presented in Table 2 highlights the use of the Relative Inequality Index (RII) and adjusted Relative Inequality Index (aRII) to measure these disparities across a range of factors, including technological, household, and demographic variables. This comprehensive review underscores the multifaceted nature of the prevalence of diabetes and the important role of socioeconomic factors in shaping health outcomes.
Technological Factors
The study findings indicate that telemedicine utilization correlates with an increased prevalence of Diabetes Mellitus (DM), even when controlling for age and sex variables. This finding suggests that individuals utilizing telemedicine are more likely to have a prior diagnosis of diabetes mellitus (DM). In contrast, individuals lacking mobile phones exhibited a greater prevalence of diabetes mellitus; however, no significant correlation was identified between phone ownership and screening activity or adherence to medication.
The use of telemedicine likely reflects the consequences of an existing disease, where people with DM are more inclined to use this technology to monitor their condition. In this context, telemedicine use can be seen as an effort to overcome barriers to healthcare access, particularly among patients who have already been diagnosed with conditions such as DM. Telemedicine offers a convenient and efficient way to provide continuous care, especially when in-person visits are difficult, such as during the COVID-19 pandemic (Hari et al., 2024; Quinton et al., 2022; Rasekaba et al., 2018). On the other hand, lower mobile phone ownership may reflect more profound socioeconomic disparities. Not owning a mobile phone can indicate limited access to technology and health information more broadly, which affects their ability to prevent and manage type 2 diabetes mellitus (T2DM; AlOsaimi et al., 2025; Birhanu et al., 2024). Therefore, the inability to access a mobile phone may be a marker of socioeconomic vulnerability that increases the risk of DM, rather than just a consequence of an existing medical condition. The digital divide, characterized by unequal access to technology, is a significant social determinant of health. Individuals from lower socioeconomic backgrounds often face barriers to accessing and utilizing these technologies, which can lead to poorer health outcomes (Bosch-Frigola et al., 2023; Xu et al., 2025).
Although the use of telemedicine tends to be associated with a higher prevalence of DM, this is more likely to reflect the high need for monitoring among patients who have already been diagnosed with DM rather than being a cause of inequality. Policies that promote the expansion of access to and use of telemedicine must take into account the condition of patients with pre-existing diseases, and focus on education and information on how to use this technology effectively. Integrating telemedicine into diabetes care models can help address disparities in healthcare access, particularly for underserved populations. By providing remote consultations and continuous monitoring, telemedicine can help improve self-management behaviors and glycemic control in diabetes patients (Grauer et al., 2022; Marsh et al., 2022; Ward et al., 2023).
The findings on technological aspects also underscore the crucial importance of access to technology, such as mobile phones, in reducing socioeconomic disparities. Individuals without mobile phones may find problems in receiving information about DM prevention and control. Mobile phone ownership is vital for diabetes care as it provides access to telehealth services, appointment scheduling, and emergency contact (Fogel & Raymond, 2020; Momin et al., 2022). Therefore, one policy option worth exploring is the expansion of programs that enhance access to technology for individuals with low socioeconomic status. Collaborative efforts between government agencies and health institutions should aim not only to increase the availability of technological devices but also to provide training and education on their effective use for health management. Ensuring digital literacy and equitable access to health-related technologies may improve disease monitoring, patient engagement, and treatment adherence, particularly among underserved populations (Abramowski et al., 2006; Wijetunga, 2014).
Household Factors
The study findings show that household disparities are reflected in wealth distribution, ownership of Social Protection Card (Kartu Perlindungan Sosial/KPS), and household water sources. Individuals in the lowest wealth quintile have a much higher prevalence of Diabetes Mellitus (DM) compared to those in the highest wealth quintile, indicating that the burden of diabetes is heavier among groups with limited economic resources. In contrast, households with a KPS exhibit a reduced prevalence of DM, potentially indicating the protective benefits of health assistance programs. Nonetheless, under diagnosis may also contribute to disparities among economically disadvantaged populations. Comparable results are noted in households with insufficient or restricted water sources. Although DM prevalence appears lower in these groups, these results almost certainly do not reflect a protective effect of poor-quality water, but rather indicate limited access to diagnostic facilities and the likely high number of undetected DM cases in areas with poor basic infrastructure.
Restricted access to healthcare services, inadequate health literacy, and increased exposure to risk factors, including poor diet and physical inactivity, contribute to the elevated prevalence of diabetes mellitus (DM) among populations with low socioeconomic status (SES). Members of these groups frequently encounter obstacles in obtaining medical care and medications, resulting in undiagnosed or inadequately managed diabetes mellitus (DM), which contributes to increased prevalence and severity of the disease (Bosetti et al., 2021; De Silva et al., 2018; Meneghini et al., 2019; Santos et al., 2017). Furthermore, restricted income frequently results in unbalanced diets characterized by elevated intake of carbohydrate-dense foods that lack essential nutrients, alongside increased alcohol consumption, both of which exacerbate the onset and progression of diabetes (Assefa & Shifera, 2022; Bosetti et al., 2021). In areas with limited water sources or poor infrastructure, these challenges are compounded by a lack of health facilities and diagnostic technologies, resulting in under detection of DM even though its prevalence may be higher than recorded (Choukem et al., 2023). Unsupportive environments, characterized by poor dietary patterns and lack of physical activity, further aggravate these risk factors, but without proper medical detection, recorded prevalence tends to be lower (Joisten, 2023). The ownership of a KPS can help reduce DM prevalence in these groups by improving access to healthcare services, including early screening, disease management, and health education (Idris & Mursyid, 2023). Programs such as KPS make medical care more accessible, allowing individuals with low SES to be diagnosed and receive better treatment, although they remain more vulnerable to DM risk factors (Mihardja et al., 2014). Nevertheless, under detection may still occur in poorer areas with limited medical facilities, potentially leading to a lower recorded DM prevalence, even though similar challenges in disease management persist among households with KPS (Kusumaningrum & Ricardo, 2022).
Policies must focus on delivering comprehensive and affordable healthcare services, particularly diabetes screening and management, for populations with low socioeconomic status (Bosetti et al., 2021; Meneghini et al., 2019). Initiatives that facilitate access to nutritious food, provide nutrition education, and create opportunities for physical activity in economically disadvantaged communities may contribute to a decrease in diabetes prevalence (Assefa & Shifera, 2022). The KPS program should be expanded to include more low-income households, with a focus on improving medical care and chronic disease management, as well as education on healthy eating, stress management, and physical activity that supports the prevention and management of DM (Rachmawati et al., 2019). These programs should also include community-based education that emphasizes early detection, disease management, and healthy lifestyles (Khayyat Kholghi et al., 2018). To address structural inequalities, efforts need to focus on improving living conditions, such as providing decent housing, reducing poverty, and strengthening social support from families, communities, and social institutions (Walker et al., 2023). Interventions aimed at elderly populations and women in low socioeconomic status groups, including community-based health education, subsidized healthcare services, and initiatives to encourage healthy lifestyles, may mitigate the disparities identified after controlling for age and sex (Darvishi et al., 2024; Pichora et al., 2018).
Demographic Factors
Individuals with low educational attainment generally demonstrate limited health literacy and are frequently engaged in low-income employment, which heightens their vulnerability to risky behaviors such as high-sugar food consumption, insufficient physical activity, and smoking. They simultaneously face limited access to nutritious foods and environments that promote an active lifestyle (Carrillo-Alvarez et al., 2025; Gautam et al., 2023). A deficiency in comprehension often leads to an incapacity to recognize personal risk for the onset of diabetes mellitus (DM). Cultural customs, such as the habitual offering of high-carbohydrate or sugary foods as a sign of hospitality, intensify this problem (Nur Fadhillah, 2024). Nevertheless, these disparities do not appear to persist in the stages of screening or treatment adherence, indicating that universal public health programs, free or low-cost screening services, and culturally sensitive approaches have proven effective in improving participation across all social groups (Khatri et al., 2024). Furthermore, guidance from healthcare professionals acts as a significant motivator, as patients often adhere to medical advice despite having a limited comprehension. The motivation to pursue screening or treatment is heightened when individuals are made aware of their risk status or a diagnosis of diabetes mellitus (Hannawa et al., 2022; Krist et al., 2017).
The study’s findings suggest that individuals in low-socioeconomic status sectors exhibit a lower prevalence of diabetes mellitus, indicating a potential under diagnosis in these populations. The low prevalence of diabetes mellitus (DM) does not inherently suggest improved health outcomes; instead, it may reflect a reduced likelihood of diagnosis stemming from inadequate access to screening services or routine examinations (Hsu et al., 2012). Conversely, individuals in the formal sector and those from higher socioeconomic status groups obtain diagnoses more frequently, due to their access to corporate healthcare services, annual health assessments, and insurance that promotes early detection (Gabriel et al., 2025). The findings suggest that prevalence data for diabetes mellitus may be biased if variations in detection rates across social groups are not considered. Therefore, it is crucial to develop policies that expand screening initiatives for informal sector workers and populations with low socioeconomic status to enhance the epidemiological understanding of diabetes mellitus (Kruk et al., 2018).
The findings suggest the need to improve screening programs and develop more targeted and inclusive health education initiatives. It is essential for governments and healthcare providers to enhance community-based screening initiatives that are either free or low-cost, with particular emphasis on the informal sector and populations with limited educational attainment. Health education programs must be customized to fit cultural contexts and presented in an accessible manner, while enhancing the role of healthcare professionals as advisors and companions. Enhancing data recording systems for diabetes mellitus prevalence is crucial to reduce bias stemming from low detection rates in at-risk populations, thus providing more precise prevalence statistics to guide policy planning.
Healthcare Utilization
Healthcare utilization shows that individuals who access healthcare facilities are more likely to be diagnosed with diabetes, with the most substantial effect coming from hospital visits, followed by community health centers (Puskesmas) and private clinics. There was no variation in screening based on healthcare utilization, unlike medication adherence. Individuals with diabetes who regularly attended Puskesmas exhibited greater likelihoods of medication adherence than those who did not attend at all. A similar pattern was noted among individuals attending private clinics and hospitals, with a discernible decline in odds ratios from Puskesmas to private clinics and hospitals.
The relationship between healthcare utilization and diabetes diagnosis reflects detection bias, in which patients who access health services more frequently are more likely to be diagnosed due to receiving examinations (Deng et al., 2022). The predominance of hospitals in diabetes detection aligns with the concept of level of care, as secondary and tertiary services have more comprehensive diagnostic facilities (Tripathy et al., 2019). The lack of variation in screening suggests that it continues to be opportunistic, concentrating on symptomatic patients instead of employing a systematic methodology. Perceptions of screening among both patients and healthcare providers influence intervention strategies, which often focus on symptomatic individuals. Conversely, patients typically do not seek diabetes testing when they perceive themselves as healthy (Sanchis et al., 2025). The increased medication adherence linked to Puskesmas visits, in contrast to private clinics and hospitals, highlights the significance of continuity of care. Regular, community-based interactions allow healthcare workers to establish trust, deliver ongoing education, and consistently monitor patients’ medication usage (Religioni et al., 2025).
These findings indicate that the current healthcare system functions more in detecting diabetes after patients seek care, rather than in promoting prevention and early detection. The dominance of hospitals in detection shows that diagnostic capacity is concentrated at secondary and tertiary levels, while primary care services have not been fully optimized for early detection. The fact that screening remains opportunistic and focused on symptomatic patients highlights that proactive detection efforts are still weak, despite the much greater potential to prevent complications if screening were systematically conducted at the community level (Shi, 2012). The higher medication adherence observed at Puskesmas compared to private clinics and hospitals emphasizes the critical importance of continuity of care, where primary care services can build long-term relationships, foster trust, and ensure patient adherence. Health policies are needed to strengthen the role of Puskesmas as centers for chronic disease management while transforming primary care services into hubs for active screening (Arsyad et al., 2022). Puskesmas can implement follow-up programs for diabetes patients, regular education, and adherence monitoring. Private clinics and hospitals remain essential, but their functions need to be more optimized for detecting complications or providing intensive care, while daily support and therapy monitoring should remain anchored in primary care (Arini et al., 2022; Pamungkas et al., 2021).
We acknowledge the limitations of this study. The assessment of access based on the presence of health facilities in the district or neighboring districts may not fully describe the actual accessibility, such as travel time, cost, or quality of services. Information on access that relies on respondents’ perceptions (“don’t know”) may cause information bias. The low use of telemedicine (3.03%) may not be sufficient to provide a comprehensive picture of its impact on the prevalence of DM. Finally, this study did not consider environmental and genetic factors that may also contribute to DM prevalence, thus reducing the generalizability of the findings. Therefore, further studies with a longitudinal design and more comprehensive variables are required to understand the dynamics of DM prevalence in Indonesia.
Conclusion
This study demonstrates a substantial association between social, economic, and health-related factors and the prevalence of diabetes mellitus (DM) in Indonesia. The overall DM prevalence was 3.08%, with a significantly higher rates observed among individuals in the lowest wealth quintile. Despite limited utilization, telemedicine use was associated with a higher prevalence of DM, underscoring the need to expand equitable access to digital health technologies. Key risk factors such as low educational attainment, unemployment, economic disparity, inadequate access to nutritious food and active living environments were shown to exacerbate vulnerability to DM. Additionally, household-level variables, such as economic status, Social Protection Card (KPS) ownership, and access to clean water, influence diagnostic opportunities and suggest potential under diagnosis in socioeconomically disadvantaged settings.
The findings indicate that DM is often detected only when individuals actively seek care, with limited evidence of proactive or population-based screening. To address these disparities, policy efforts should focus on strengthening community health centers (Puskesmas) as hubs for chronic disease management, scaling up active case finding, enhancing access to health-related technologies, and providing culturally appropriate health education. Furthermore, improved health information systems are essential for monitoring disparities and targeting interventions effectively. A more community-oriented and inclusive public health approach is critical to ensuring equitable, effective, and sustainable prevention and control of diabetes in Indonesia.
Footnotes
Acknowledgements
This paper is a part of activities funded by the Indonesian Endowment Fund for Education (LPDP) on behalf of the Indonesian Ministry of Higher Education, Science and Technology and managed under the EQUITY Program, under Contract Number: 4300/B3/DT.03.08/2025 and 297/UN3/HK.07.00/2025.
ORCID iDs
Ethical Consideration
This study utilized secondary data from the 2023 Indonesian Health, which was conducted by the Health Policy Agency, Ministry of Health, Indonesia. All data were fully anonymized prior to being accessed by the researchers. This study involved no direct intervention or interaction with participants. All data collection procedures conducted by the survey organizer adhered to ethical standards and received approval from the Ministry of Health Research and Development health survey ethics protocol amendment (approval No. LB.02.01/I/KE/L/287/2023) on May 10, 2023.
Consent to Participate
Informed consent was obtained from all participants during the original data collection process.
Author Contributions
All authors contributed significantly to every aspect of the reported work, including concept design, study design, implementation, data collection, analysis, and interpretation. They also participated in the drafting, revision, and critical review of the article. Each author gave final approval to the version to be published and has approved the journal to which this article was submitted. Additionally, they agree to be responsible for all aspects of the work carried out.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.