Mental Health Service Utilization among Older Adults in Indonesia: Nationwide Retrospective Cohort Study Using the National Health Insurance Claims Data, 2015–2023

Article information

Ann Geriatr Med Res. 2025;29(4):525-533

Publication date (electronic) : 2025 October 17

doi : https://doi.org/10.4235/agmr.25.0108

Yeni Mahwati¹, Syarif Rahman Hasibuan^,2

¹Sekolah Tinggi Ilmu Kesehatan Dharma Husada, Bandung City, Indonesia

²Faculty of Medicine, Universitas Pembangunan Nasional Veteran Jakarta, Jakarta, Indonesia

Corresponding Author: Syarif rahman Hasibuan, SKM, MKM Faculty of Medicine, Universitas Pembangunan Nasional Veteran Jakarta, 12450, Jakarta, Indonesia E-mail: syarifrhasibuan@upnvj.ac.id

Received 2025 July 14; Revised 2025 September 6; Accepted 2025 October 12.

Abstract

INTRODUCTION

Mental health disorders among older adults are a growing global concern. As populations age, the burden of conditions such as depression, anxiety, dementia, and schizophrenia has increased substantially, contributing to disability and excess mortality. Depression is the most prevalent disorder, with global estimates ranging from 13.3% to over 30%, and even higher in developing countries, where prevalence may exceed 40%.1-3) These disorders are frequently underdiagnosed and undertreated, especially in low-resource settings, exacerbating their impact on individuals and public health systems.4,5)

Older adults in low- and middle-income countries face disproportionate risk due to limited access to geriatric mental health services, inadequate detection, stigma, and social vulnerabilities. Risk factors for mental illness in late life are well-documented, including the female gender, chronic illness, low socioeconomic status, lack of social support, and loneliness were particularly among those who are widowed or live alone.6,7) In this context, the burden of depression, anxiety, and dementia is expected to rise significantly.

In Indonesia, mental health problems among the elderly are prevalent and concerning. Several local studies report depressive symptoms affecting between 16% and 60% of older adults, depending on the population studied and measurement tools used.8-10) Emotional disorders such as anxiety and stress are also common, particularly among individuals with comorbid physical conditions, sensory impairments, and functional limitations.11-13) These mental health issues are closely linked to reduced quality of life and increased risk of suicidal ideation.8,14) Untreated mental disorders in older adults may progress to severe disability and, in some cases, suicide. Although the World Health Organization reports a low age-standardized suicide rate in Indonesia (2.6 per 100,000), these figures are classified as low quality and are likely subject to severe underreporting due to stigma, weak registry systems, and cultural taboos surrounding suicide.15) This lack of reliable data underscores the challenge of quantifying the true burden of psychiatric morbidity in Indonesia and reinforces the importance of leveraging administrative claims data to examine mental health service utilization among older adults.

The urgency is further amplified by Indonesia’s rapid demographic transition. The proportion of older adults has increased from 8.4% in 2015 to 12.0% in 2024, accompanied by a rise in life expectancy from 70.8 to 72.4 years.16) This aging trend is expected to generate greater demand for age-appropriate mental health services. Moreover, the number of older adults enrolled in Indonesia’s National Health Insurance or Jaminan Kesehatan Nasional (JKN) program administered by the Social Health Insurance Administration Body (Badan Penyelenggara Jaminan Sosial (BPJS) Kasehatan) has exceeded 15 million in 2021.17) This large coverage offers a unique opportunity to examine how mental health services are actually utilized by older adults across the country. The use of service utilization data could provide evidence as a crucial first step in quantifying the magnitude of the problem and identifying gaps in care.

The JKN program launched in 2014 is a mandatory, publicly financed health insurance program with different funding sources depending on employment status available to all Indonesian citizens.18) The JKN program guarantees a broad range of health services spanning preventive, promotive, curative, and rehabilitative care, across primary, secondary, and tertiary levels.19) This includes comprehensive coverage for mental and behavioral disorders, with services for assessment, diagnosis, pharmacological and non-pharmacological treatment, hospitalization, and follow-up care. Mental health conditions classified under the International Classification of Diseases 10th revision (ICD-10) codes F00–F99 are included in the benefits package, and claims for these services are submitted through the INA-CBG (Indonesia Case-Based Groups) system. As a result, BPJS Kesehatan administrative data provide a robust source of information for understanding mental health service utilization among insured individuals in Indonesia.

Beyond the JKN system, social care for older adults in Indonesia remains largely family-based, complemented by community initiatives such as Posyandu that provide health monitoring and social activities.20,21) Institutional care, including nursing homes managed by local governments, is limited due to cultural preferences for family support and resource constraints.22) These characteristics highlight that while JKN ensures financial access to health services, broader social support for older adults is uneven and heavily reliant on family and community structures.

Despite growing awareness, there remains a limited understanding of how older adults in Indonesia utilize mental health services. Most existing studies rely on localized surveys or small-scale observational data, which do not reflect broader patterns of service access and diagnosis across the national health system. This lack of system-level insight hinders efforts to evaluate the reach and effectiveness of mental health care for the elderly, and limits the ability to plan and allocate resources based on actual needs.

To address this gap, this study aims to examine the utilization of mental health services among older adults in Indonesia using administrative claims data from BPJS Kesehatan between 2015 and 2023. By identifying the number and characteristics of elderly individuals who received care for mental health conditions, as well as the types of diagnoses they received, this study provides foundational evidence for understanding mental health care patterns within a large, nationally insured population.

MATERIALS AND METHODS

This study employed a retrospective cohort design using secondary data from the 2024 release of the BPJS Kesehatan Mental Health Contextual Sample. The dataset was constructed by BPJS Kesehatan using a stratified random sampling method based on individual-level data. Stratification was performed using three key variables: province of residence, mental health diagnosis category (including anxiety disorders, depressive disorders, schizophrenia, autism spectrum disorder, and others), and insurance membership category (government-subsidized groups by national budget [PBI-APBN], government-subsidized groups by local budget [PBI-APBD], informal workers [PBPU], formal workers [PPU], and non-workers).23) Only individuals who utilized mental health services in 2023 were eligible for sampling into the dataset. The resulting dataset includes linked tables for membership data, primary care utilization (capitation and non-capitation), referral care utilization (FKRTL), and secondary diagnoses. This dataset captures only individuals who accessed mental health services and received a recorded diagnosis, excluding older adults with unmet needs who did not seek care. The sample selection process by BPJS Kesehatan can be seen in Fig. 1.

Fig. 1.

Sampling framework based on stratified random sampling of mental health service users. Source: adapted and translated from the original visual published by BPJS Kesehatan.

For this study, we included individuals aged ≥60 years, and were confirmed to be alive as of December 31, 2023. From this group, we identified those who had at least one recorded diagnosis of a mental disorder based on ICD-10 codes F00–F99, whether as a primary or secondary diagnosis. The sample cleaning process can be seen in Fig. 2.

Fig. 2.

Sample cleaning process.

From a total of 54,820 individuals included in the Mental Health Contextual dataset released by BPJS Kesehatan in 2024, a subset of participants was selected based on this study’s criteria. Final sample of this study consisted of 5,966 participants.

The primary outcome of this research was mental health service utilization, operationalized as the total number of healthcare visits during which any mental disorder diagnosis (F00–F99) was recorded. Independent variables included: age, sex, insurance membership category, and the insurance ward class. For descriptive analysis, unweighted data were used to report the real counts of individuals and visits.

We conducted descriptive analyses using actual unweighted visit counts. For inferential statistics, we applied negative binomial regression to examine the association between sociodemographic factors and total mental health service utilization, accounting for overdispersion in the count data.24) Although the dataset was constructed using stratified random sampling to enhance representativeness, it only captures individuals who accessed mental health services under the BPJS Kesehatan scheme. Therefore, older adults who did not seek care are not included, which may introduce selection bias and limit the generalizability of the findings to the broader elderly population. All analyses were performed using SPSS (IBM, Armonk, NY, USA). No missing data were identified in the study sample.

RESULTS

A total of 5,966 older adults were included in the analysis, selected from an initial pool of 54,820 individuals. The sociodemographic characteristics of the sample, including age distribution, sex, marital status, registered hospital ward class, and health insurance segmentation, are presented in Table 1.

Table 1.

Sociodemographic background of the participants

The mean age of participants was 66 years, ranging from 60 to 96 years, with a standard deviation of 6 years. The majority of participants were female (55.7%), while 44.3% were male. Most participants were married (78.0%), with 16.0% divorced or separated, and 6.1% never married. Regarding hospital ward class registration, nearly all participants (95.6%) were enrolled in Class III, with a small proportion in Class I (3.1%) and Class II (1.3%). In the JKN system, hospital care is stratified into three ward classes (Class I, II, and III), which correspond to the level of inpatient accommodation and the size of insurance premiums paid. Class I beneficiaries contribute the highest premium per month.

In terms of participant segmentation under the National Health Insurance (JKN), the largest group was PBI-APBD at 44.3%, followed by PBPU at 55.7%. Other segments included PPU at 6.1%, PBI-APBN at 13.2%, and non-workers at 21.7%.

As presented in Table 2, from 2015 to 2023, a total of 93,557 mental health service utilizations were recorded among the older adult participants. On average, each participant utilized mental health services 4.94 times per year, with a standard deviation of 5.18. The minimum number of utilizations per person was 1, while the highest reached 49 visits per year. These figures suggest a skewed distribution of service use, where a small group of individuals accounted for a disproportionately high number of visits.

Table 2.

Mental health service utilization among older adults (2015–2023)

Our trend analysis of service utilization between 2015 and 2023 is presented in Supplementary Fig. S1. There was a steady increase in utilization from 2015 to 2019, followed by a significant increase during the COVID-19 pandemic in 2020–2021, and a consistent increase in 2022–2023. These patterns indicate that the pandemic might affect older adults' mental health.

Geographic variation in utilization is presented in Supplementary Table S1. Service utilization was higher in cities compared to districts. Similarly, provinces outside Java and Bali accounted for a greater share of total utilization than those within Java and Bali. These descriptive patterns suggest potential inequalities in access across provinces and between cities and district settings.

Fig. 3 shows the distribution of mental disorder diagnoses among older adults based on ICD-10 categories. The most frequently diagnosed condition was anxiety and neurotic disorders (F40–F48), affecting 35.95% of the sample. This was followed by psychotic disorders (F20–F29) with 28.29%, and mood (affective) disorders (F30–F39) at 18.97%.

Fig. 3.

Proportion of mental disorder diagnoses among older adults by the International Classification of Diseases 10th revision (ICD-10) category (n=93,557). Data source: BPJS Kesehatan, processed by authors.

Other notable categories included dementia and other organic mental disorders (F00–F09) at 11.43% and behavioral syndromes associated with physiological factors (F50–F59) at 3.58%. Diagnoses from other groups—such as childhood behavioral disorders, developmental disorders, personality disorders, substance use disorders, and unspecified mental disorders—affected less than 2% of the population.

These findings highlight that late-life anxiety, psychosis, and depression are the most prevalent mental health diagnoses among older adults who utilize services in Indonesia’s national health insurance system. A small proportion of cases were coded under ICD-10 categories F70–F89 and F90–F98, which are conventionally defined as childhood-onset disorders. These likely represent lifelong conditions persisting into old age rather than new late-onset diagnoses, as intellectual and developmental disabilities may continue throughout the life course.25) We retained these categories to present the data as released by BPJS Kesehatan, but this also highlights the need for stronger verification of ICD-10 coding in administrative claims submitted by hospitals.

Table 3 presents the results of the negative binomial regression analysis examining factors associated with the total number of healthcare utilization events among older JKN participants. Several sociodemographic and insurance-related characteristics were significantly associated with healthcare utilization.

Table 3.

Significant factors associated with total healthcare utilization (negative binomial regression)

Compared to PPU group, other participant groups had significantly lower utilization rates. The PBI-APBN group (B=-0.40; 95% confidence interval [CI] -0.42, -0.38; p<0.01) and local budget (APBD) (B=-0.25; 95% CI -0.28, -0.22; p<0.01) showed the strongest negative associations. Informal workers (PBPU) (B=-0.08, p<0.01) and non-workers (B=-0.05, p<0.01) also had significantly lower utilization.

Ward class showed strong associations. Participants in Class I (B=-1.14, p<0.01) and Class II (B=-0.68, p<0.01) wards had significantly lower utilization compared to those in Class III.

Marital status and age were also significant predictors. Individuals who were never married had lower utilization (B=-0.15, p<0.01) than those who were married. Increasing age was associated with a small but significant decrease in utilization (B=-0.06 per year, p<0.01). Sex and divorce status were not significantly associated with healthcare utilization.

DISCUSSION

Our findings reveal that mental health service use among older JKN participants is concentrated on neurotic, psychotic, and mood disorders. Despite the comprehensive coverage of the JKN scheme, annual service utilization remained modest and declined over time. Striking differences were observed across insurance segments and ward classes, suggesting persistent inequalities in access or treatment-seeking behavior, even within a universal health coverage system. These patterns underscore the complexity of mental health care utilization in later life and raise questions about service adequacy and equity across participant groups.

To our knowledge, there are no previous national-level studies in Indonesia examining mental health service utilization among older adults enrolled in JKN. This study is among the first to explore this issue using data from BPJS Kesehatan. This is despite the fact that several researchers in Indonesia explored the prevalence of emotional disorders in late life. These studies, conducted in areas such as Kendal, Bandung, and Bali, consistently found high rates of depressive symptoms, with moderate-to-severe depression ranging from 41.6% to 71%.26–28) These figures reflect mental health needs in the community, while our findings provide a complementary perspective by capturing actual service utilization. The discrepancy between high prevalence and modest service use may reflect barriers in detection, referral, or treatment uptake among older adults.

The analysis revealed substantial disparities across insurance categories. Older adults enrolled as PPU utilized mental health services significantly more than those in the PBI-APBD, PBI-APBN and PBPU group. This likely reflects differences in socioeconomic status, awareness, and access to care. Individuals in the PBI group are economically disadvantaged and may face barriers such as transportation costs, stigma, or lower perceived need. These findings align with previous research highlighting the role of income and education in shaping service use.29)

In our study, the highest mental health service utilization was observed among elderly participants with PPU status, predominantly in the lowest premium class. This may not reflect socioeconomic advantage, but rather administrative patterns in the JKN system that the retirees often retain PPU status but are entitled only to Class III ward based on former office policy, or they are enrolled as dependents of their working children. These conditions may also relate to better awareness of health entitlements and easier navigation of services through family support. Previous studies have shown that perceived need, social support, and insurance coverage are critical enablers of healthcare access among the elderly.30,31)

The higher utilization among Class III beneficiaries may reflect their dominant representation in the sample (95.6%) and the financial protection from government subsidies. Supplementary Table S2 confirms this pattern, with the average number of utilizations per year substantially higher in Class III (5 vs. 2 in Class I/II). Notably, this contrasts with international evidence where lower socioeconomic status is usually linked to lower utilization.32,33) Indonesia’s JKN scheme appears to mitigate these barriers, enabling disadvantaged groups to use services more frequently, offering novel insights into how universal coverage can reshape socioeconomic patterns of mental health care.

Our findings also indicate that unmarried older adults had significantly lower levels of mental health service utilization. This may be due to a lack of social support, especially from spouses or close family members, which plays a key role in recognizing mental health symptoms, accessing services, and ensuring follow-up.32,34) The absence of a spouse may also heighten feelings of isolation or stigma, further deterring care-seeking behavior.35,36) This is consistent with our findings, where unmarried JKN members showed significantly lower mental health service utilization, possibly due to these structural and emotional barriers. Given the lack of national evidence in Indonesia, global findings are used here to contextualize the results.

Gender was not significantly associated with mental health service utilization in this study. This contrasts with prior evidence suggesting higher use among older women, attributed to greater health awareness and help-seeking behavior.37,38) In the Indonesian context, gender differences may be attenuated by structural and cultural factors, including stigma and access limitations, which constrain utilization across both sexes.39)

Similarly, being divorced was not significantly associated with mental health service utilization among older adults. Although divorce can lead to psychological distress, recent studies suggest its influence on care-seeking behavior may vary depending on individual coping capacity, time elapsed since separation, and the availability of alternative social support systems. Some older individuals may adapt over time or rely on informal support networks rather than formal services.40) Additionally, the lack of significance in our analysis may reflect limited statistical power within this subgroup or confounding effects from related variables such as living arrangement or economic security.

Our analysis utilizes administrative data from BPJS Kesehatan, the agency administering Indonesia’s National Health Insurance (JKN) scheme. This dataset includes service utilization records from all contracted referral care providers (both public and private), that submitted mental health claims for older adults. Given the near-universal coverage of JKN, these data offer a unique opportunity to assess mental health service use on a national scale, capturing population groups and facilities that may be underrepresented in other surveillance systems. In particular, this study focuses on services for mental and behavioral disorders, a high-priority area for aging populations yet underexplored in Indonesian health research. By linking diagnostic and service-level data, the dataset allows stratified analysis by participant segment, care class, and diagnostic category. These insights can inform future policy to improve access, equity, and continuity of care for older individuals with mental disorders. This approach mirrors efforts in countries such as Japan and South Korea, where large-scale population-based databases (e.g., JAGES in Japan and the Korean Urban Rural Elderly study) have been successfully used to identify aging-related health patterns and inform policy development.41,42)

Our findings highlight persistent disparities in mental health service utilization among older adults, particularly by insurance segment, ward class, marital status, and age. To address these gaps, policies should strengthen community-based and outreach services, enhance detection and management of common disorders at the primary care level, and ensure sustained financial protection through JKN subsidies for the most vulnerable groups. Leveraging BPJS Kesehatan big data for routine surveillance, akin to international initiatives like JAGES in Japan, could further support evidence-based planning and equitable mental health care for Indonesia’s aging population.

Mental health in later life is strongly influenced by chronic disease, disability, poverty, education, and social support.43-45) While JKN reduces financial barriers, these broader determinants remain important in shaping service use, as also reflected in our findings by marital status and insurance segment. Taken together, prior studies consistently show that social and health determinants drive disparities in mental health outcomes, suggesting that insurance coverage alone is insufficient without attention to these underlying factors.

This study has several limitations. The dataset only includes claims from referral care facilities, excluding mental health services delivered in primary care under the capitation payment system. As a result, early-stage consultations or cases managed exclusively in primary care are not captured. Clinical detail is limited; the dataset lacks information on symptom severity, treatment adherence, outcomes, or comorbid conditions, restricting the ability to assess quality or effectiveness of care. Another limitation of our study is not capture comorbidities or disability status, which are known predictors of psychiatric morbidity. Mental disorder diagnoses were identified through ICD-10 codes recorded in the JKN case-mix system. These codes are assigned by hospital case-mix managers, who are generally physicians, based on clinical documentation. However, as with all secondary administrative data, we could not validate these diagnoses against biomarkers, symptom checklists, or independent clinical assessments. This limitation raises the possibility of under- or misclassification, which should be considered when interpreting the findings. Although JKN covers a large share of the population, it does not include individuals outside the national insurance system, such as those uninsured or seeking care through alternative pathways, who may have different utilization patterns. Additionally, this dataset reflects only those who accessed services, excluding older adults with unmet mental health needs who never engaged with the healthcare system.

In conclusion, this study addresses the existing evidence gap by examining the utilization of mental health services among older adults in Indonesia using national claims data from BPJS Kesehatan between 2015 and 2023. We found that despite comprehensive coverage under the JKN scheme, the overall use of mental health services in later life remains underutilized, with notable disparities across participant types, ward classes, marital status, and age groups. Utilization is likely influenced by a combination of structural factors, health literacy, caregiving support, and institutional capacity. Our findings underscore the need for targeted policy interventions to strengthen detection, referral, and care pathways for older adults, particularly among socially isolated individuals or those from disadvantaged segments. Strengthening community-based detection and primary care management, while maintaining financial protection through JKN for the most vulnerable elderly, is critical to reducing disparities and ensure equitable access to mental health services in Indonesia.

SUPPLEMENTARY MATERIALS

Article information Continued