Suicidal ideation in family caregivers of patients with dementia, cancer, stroke, and ischemic heart disease

Abstract

Background

Suicidal ideation among caregivers has drawn increasing attention. However, most prior studies have examined single diseases and provide limited comparative evidence. Furthermore, evidence focusing on family caregivers of community-dwelling adults with chronic diseases remains limited. Using nationally representative data of community-dwelling adults, we assessed the association between caregiving and suicidal ideation across four high-burden conditions: dementia, cancer, stroke, and ischemic heart disease (IHD).

Methods

We conducted a cross-sectional analysis using 2007–2021 data from the Korean National Health and Nutrition Examination Survey (KNHANES). Patients with the four target diseases were identified and linked to family members through a household identification number. The final sample included 7,025 caregivers and 66,872 non-caregivers. Suicidal ideation was assessed using a standardized questionnaire. Multivariable logistic regression and subgroup analyses were performed.

Results

Suicidal ideation was reported by 6.1% of non-caregivers and 6.9% of caregivers (adjusted odds ratio [aOR] = 1.18;95%CI = 1.05–1.32). Compared to non-caregivers, the risk of suicidal ideation was highest among caregivers of patients with IHD (aOR = 1.46;95%CI = 1.18–1.80) and patients or sibling caregivers (aOR = 2.21;95%CI = 1.34–3.62), caring for female patients (aOR = 1.18;95%CI = 1.03–1.36) and these caring for younger patients (< 60 years). In subgroup analysis, the association was notably stronger among older and female caregivers than male caregivers.

Conclusions

Mental health support should be tailored to disease and caregiving role. Interventions are especially needed for parents and spouses under high psychological burden. Policies must prioritize screening and access to mental health care for vulnerable caregivers.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12991-026-00670-5.

Introduction

Caregivers play a pivotal role in supporting patients with chronic and serious illnesses and often bear significant emotional, physical, and financial burdens [1]. Recently, the psychological consequences of caregiving, including an increased risk of depression and suicidal ideation, have been increasingly recognized [2]. The proportion of caregivers experiencing suicidal ideation ranged from 2.7% to 71% [3]. As the population ages and chronic conditions become more prevalent, family caregivers play an increasingly critical role in home-based care. In this context, understanding how mental health risks vary by disease type and caregiving role is essential for designing targeted interventions [4, 5].

Previous studies have shown that the psychological burden of caregiving varies considerably depending on both the type of illness and the caregiver’s role. Caregivers of cancer patients often face complex treatment schedules and uncertainty regarding prognosis [6], whereas those caring for individuals with dementia frequently face behavioral disturbances and the emotional exhaust of progressive cognitive decline [7]. In contrast, caregivers of patients with acute conditions that carry potential for recovery may experience different stressors, such as sudden caregiving demands and post-acute uncertainty, compared to those managing long-term stable conditions or degenerative illnesses [8]. Furthermore, the caregiver’s relationship to the patient significantly influences stress levels and coping mechanisms. Spousal caregivers may experience identity disruption and emotional isolation, while adult children often confront role reversal and competing familial or occupational responsibilities [9]. These observations underscore the importance of examining both disease context and caregiver role when assessing mental health risks.

Despite growing recognition of caregiver burden, most existing studies have focused on caregivers of patients living in institutional settings such as nursing homes or daycare centers, or those relying on formal social services [10]. In contrast, research systematically examining the burden on family caregivers of community-dwelling adults with chronic diseases remains limited. These caregivers operate not in acute or crisis situations but within the persistent demands of daily life, where continuous vigilance and emotional fatigue may gradually accumulate [11]. Furthermore, most prior studies have concentrated on a single disease or caregiving role, making it difficult to compare mental health risks across caregiving contexts or to identify subgroups based on characteristics such as sex, age, or familial relationship [10]. This limitation hinders the development of tailored interventions for diverse caregiver populations. To address these gaps, we examined disease-specific and role-specific patterns in caregiver suicidal ideation across four high-burden illnesses: dementia, cancer, stroke, and ischemic heart disease (IHD). Using data from the Korean National Health and Nutrition Examination Survey (KNHANES), a nationwide cross-sectional survey conducted by the Korea Disease Control and Prevention Agency. KNHANES uses a multistage, stratified, cluster sampling design to represent community-dwelling adults in South Korea. By combining disease-specific and role-specific comparisons using nationally representative data, our study provides a comprehensive assessment of the mental health risks faced by caregivers in different disease contexts.

Methods

Data sources and study participants

This cross-sectional study used data from the NHANES. The KNHANES uses a nationally representative cross-sectional design, focusing on non-institutionalized populations through a multistage cluster sampling approach [12]. In the KNHANES, participants completed the questionnaire once, ensuring one response per individual. These surveys are conducted annually as cross-sectional studies with each iteration involving a distinct sample population.

Between 2007 and 2021, a total of 79,946 individuals from 40,150 households were aged 19 years or older and lived in households with at least two members. We identified households in which at least one family member had dementia, cancer, stroke, or IHD. Using household family identification numbers, each patient was linked to their cohabitating family members. All eligible cohabitating family members, excluding the patients themselves (n = 6,134), were defined as caregivers. This process resulted in a total of 7,025 caregivers across 5,059 households. Disease-specific caregiver counts were 85 for dementia, 3,679 for cancer, 2,673 for stroke, and 2,959 for IHD. Participants who did not have any family members with these conditions were classified as the control group (66,872 individuals, 34,181 households) (Fig. 1). To address cases where a household included more than one patient, we assigned each caregiver to a single primary patient using a predefined disease-priority rule reflecting clinical burden (dementia > cancer > stroke > ischemic heart disease), and flagged with a separate indicator variable to capture multisystem disease burden.

Fig. 1.

Flow chart of the study

Survey participation followed the standardized national protocol administered by trained survey staff. Before any examination or interview, staff members verified participant eligibility, explained the purpose of the survey, described all components of data collection including health interviews, physical examinations, and laboratory tests, and informed participants of data confidentiality and voluntary participation. Participants provided written consent at the mobile examination center before the health interview and examination. For the nutrition survey conducted at home approximately one week later, written consent was confirmed again by trained nutrition interviewers. The KNHANES was approved by the Institutional Review Board (IRB) of the KDCA (2007-02CON-04-P, 2008-04EXP-01-C 2009-01CON-03–2 C 010-02CON-21-C 2011-02CON06-C 2012-01EXP-01–2 C 2013-07CON-03–4 C 2013-12EXP-03–5 C), whereas the review was waived in 2015 according to Article 1 (1) of the Bioethics and Safety Act and Article 2 (2)-1 of the Enforcement Decree of the Bioethics and Safety Act in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

Measurements

All measurements and definitions were based on standardized protocols from the KNHANES, which ensures comparability and reliability across survey years [12–15].

Caregiver identification and disease definition

We first identified individuals with cancer, stroke, or IHD using self-reported physician-diagnosed conditions. Cancer included cases of gastric, liver, colorectal, breast, cervical, lung, or thyroid cancer, as reported by a physician. Stroke and IHD were defined as physician-diagnosed stroke, or myocardial infarction/angina pectoris, respectively. In contrast, dementia was defined as a self-reported activity limitation attributed to dementia, based on responses to the activity limitation module in KNHANES. In the activity limitation module, participants are asked whether they experience limitations in daily or social activities and, if so, to identify the primary medical condition responsible for the limitation. Participants were classified as having dementia if they reported dementia as the main cause of their activity limitation. Using unique family identification numbers, we linked these patients to cohabiting family members within the same household to classify them as caregivers.

Suicidal ideation

The primary outcome was suicidal ideation, assessed using two standardized measures. Participants were considered to have suicidal ideation if they answered “yes” to the question, “Have you thought about wanting to die at any point in the past year?” or if they provided a positive response to item 9 on the Patient Health Questionnaire-9 (PHQ-9), which asks about thoughts of death or self-harm in the past two weeks [16].

Sociodemographic, behavioral covariates and comorbidities

We collected age, sex, education level, marital status, household income (per month), and employment status using structured self-report questionnaires. Health-related behaviors such as current smoking and alcohol consumption were also assessed via self-report. Comorbidities were identified using a combination of physical examinations and questionnaires. Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, a self-reported history of hypertension, or current use of antihypertensive medications. Dyslipidemia was defined as low-density lipoprotein cholesterol ≥ 130 mg/dL, high-density lipoprotein cholesterol ≤ 40 mg/dL, a self-reported history of dyslipidemia, or the current use of lipid-lowering medications [17]. Diabetes mellitus (DM) was defined as fasting serum glucose ≥ 126 mg/dL, a self-reported history of DM, or current use of glucose-lowering medications.

Statistical analyses

Since KNHANES uses a multistage, stratified cluster sampling design, weighted means and proportions were calculated and presented with standard errors. Because of the large sample size, baseline characteristics between participants with and without family members affected by dementia, cancer, stroke, or IHD were compared using standardized mean differences (SMDs) rather than P values. SMDs were calculated for all baseline variables, and values less than 0.1 were interpreted as indicating minimal imbalance.

We estimated the weighted prevalence of suicidal ideation among caregivers by disease type, using descriptive statistics with 95% confidence intervals. Multivariable logistic regression models were used to estimate adjusted odds ratios (aORs) and 95% confidence intervals (CIs) for suicidal ideation according to caregiving status and disease group. All models used participants without any family members affected by the four target diseases as the reference group (non-caregiver group). Covariates included age, sex, economic activity status, and comorbidities, selected based on theoretical relevance and prior literature. To account for the inclusion of multiple members from the same family, we adjusted for clustering at the family level. We also generated subcategories among caregiver group by role, multiple patients, patient’s sex, age and disease type, and them compare to the non-caregiver group.

We conducted subgroup analyses to evaluate whether the association between caregiving and suicidal ideation differed by caregiver age, sex, economic activity, and family income. Interaction terms were included in the logistic regression models, and P-values for interaction were obtained. In all subgroup analyses, the reference group remained individuals without any affected family members.

All tests were two-sided with a significance threshold of P < 0.05. Analyses were performed using R version 4.1.2.

Results

A total of 73,897 participants were included in the study. Among them, 66,872 had no family members with severe conditions (non-caregiver group), whereas 7,025 had at least one family member affected by dementia, cancer, stroke, or IHD (caregiver group) (Table 1). Compared with non-caregivers, caregivers were less likely to have a college or university education or higher (66.7% vs. 78.8%) and had a higher weighted percentage of current smoking (21.3% vs. 14.7%). Approximately 1.8% of all participants reported having two or more family members with chronic diseases. Across disease-specific caregiver groups, the majority were female, and the mean age was 48.3 years. Caregivers of individuals with dementia most frequently reported having more than two affected family members (Supplementary Table 1).

Table 1.

Characteristics of caregiver whose family have disease

	Non-caregiver (N = 66,872)	Caregiver (N = 7,025)	SMD
Caregiver
Sex, female % (SE)	50.2 (0.003)	48.4 (0.008)	0.037
Age, weighed mean (SE)	44.67 (0.1)	47.52 (0.28)	0.170
Education level % (SE)	12.1 (0.003)	17.5 (0.009)	0.171
Less than elementary school graduate	8 (0.003)	9.5 (0.01)
Middle school graduate	37 (0.003)	33.6 (0.01)
High school graduate	36.9 (0.003)	34.1 (0.011)
College/university graduate or above	78.8 (0.002)	66.7 (0.006)	0.275
Marital status, married % (SE)	3.02 (0.085)	3.3 (0.091)	0.001
Monthly income (1000 USD), weighed mean (SE)	61.8 (0.002)	59 (0.008)	0.057
Economic activity status, employed % (SE)	25.8 (0.004)	23.9 (0.011)	0.044
Smoking status, current, yes % (SE)	14.7 (0.003)	21.3 (0.009)	0.172
Alcohol status, current, yes % (SE)	75.8 (0.002)	71.4 (0.007)	0.109
Comorbidity % (SE)	10.2 (0.003)	14.6 (0.01)	0.135
Hypertension	5.7 (0.003)	8.4 (0.01)	0.102
Hyperlipidemia	50.2 (0.003)	48.4 (0.008)	0.037
Diabetes	44.67 (0.1)	47.52 (0.28)	0.170
More than 2 families with disease % (SE)		1.8 (0.011)	0.170
Role in the family % (SE)
Offspring	-	42 (0.011)
Spouse	-	51.2 (0.008)
Siblings	-	2.3 (0.012)
Parents	-	1.8 (0.01)
Patients
Sex, female % (SE)		52 (0.009)
Age, weighed mean (SE)		61.37 (0.22)
Type of disease % (SE)
Dementia	-	0.8 (0.012)	-
Cancer	-	54.9 (0.009)	-
Stroke	-	21.2 (0.011)	-
IHD	-	23.1 (0.011)	-

SE; Standard Error, IHD; Ischemic Heart Disease

Suicidal ideation was reported by 6.1% of non-caregivers and 6.9% of caregivers (adjusted odds ratio [aOR] = 1.18; 95% CI = 1.05–1.32) (Table 2). Compared to non-caregivers, the risk of suicidal ideation was highest among caregivers of patients with IHD (aOR = 1.46; 95% CI = 1.18–1.80), followed by caregivers of stroke patients (aOR = 1.26; 95% CI = 1.00–1.59). Regarding the caregiving role, parents to care for considerably younger patients (median age 43 years), whereas offspring and spouses predominantly cared for older adults (median ages 72 and 63 years, respectively) (Supplementary Table 2). Patients or sibling caregivers showed a significantly elevated risk of suicidal ideation compared to non-caregivers (aOR = 2.21; 95% CI = 1.34–3.62), followed by spouse caregiver (aOR = 1.23; 95% CI = 1.00–1.51). Caregiver who had more than two patients in family members had the highest prevalence of suicidal ideation (aOR = 1.41; 95% CI = 0.76–2.64), although confidence intervals were wide due to small sample size. In terms of patient characteristics, the odds of suicidal ideation were higher in caring for female patients (aOR = 1.18; 95% CI = 1.03–1.36) and these caring for younger patients (< 60 years) (aOR = 1.21; 95% CI = 1.04–1.41) (Table 2).

Table 2.

Association between caregiving status, disease type and suicide ideation

	Weighted proportion of suicide ideation	Adjusted* Odds Ratio (95% CI)
Caregiver
No	6.1 (0.1)	Reference
Yes	6.9 (0.4)	1.18 (1.05–1.32)
Disease type
Non-caregiver	6.1 (0.1)	Reference
Caregiver for dementia patients	7.8 (3)	1.23 (0.55–2.74)
Caregiver for cancer	6.1 (0.1)	1.01 (0.86–1.19)
Caregiver for stroke	8.2 (0.9)	1.26 (1.00-1.59)
Caregiver for IHD	8.9 (0.9)	1.46 (1.18–1.80)
Role of the caregiver
Non-caregiver	6.1 (0.1)	Reference
Offspring	6.2 (0.6)	1.11 (0.97–1.27)
Spouse	7.1 (0.4)	1.23 (1.00-1.51)
Parents or siblings	13.9 (2.9)	2.21 (1.34–3.62)
Caregiver with multiple patients
Non-caregiver	6.1 (0.1)	Reference
Only 1	6.9 (0.4)	1.17 (1.04–1.32)
More than 2	8.5 (2.5)	1.41 (0.76–2.64)
Patient Sex
Non-caregiver	6.1 (0.1)	Reference
Having male patients	6.2 (0.6)	1.13 (0.93–1.37)
Having female patients	6.9 (0.4)	1.18 (1.03–1.36)
Patient age
Non-caregiver	6.1 (0.1)	Reference
Having patients < 60 years old	7.1 (0.4)	1.21 (1.04–1.41)
Having patients ≥ 60 years old	6.1 (0.1)	1.02 (0.87–1.20)

SE; Standard Error, CI; confidence interval, IHD; Ischemic Heart Disease

* Adjusted for age, sex, economic activity, and comorbidities

Although caregiving was consistently associated with higher suicidal ideation across subgroups, the association was notably stronger among older age caregiver (aOR = 1.25;95% CI = 1.08–1.44) and female caregivers (aOR = 1.26; 95% CI = 1.08–1.47) than young and male caregivers, respectively (Fig. 2).

Fig. 2.

Subgroup analysis for association between caregiving status and suicide ideation. aOR; adjusted odds ratio, CI; confidence interval. * Adjusted for age, sex, economic activity, health behavior, and comorbidities

Discussion

This nationwide study found that caregiving was significantly associated with suicidal ideation. Notably, when compared to the non-caregiver reference group, the odds of suicidal ideation were highest among caregivers of patients with IHD and parents caring for affected children. In subgroup analyses, the association between caregiving and suicidal ideation was stronger in older and female caregivers.

Caregiving was associated with an increase in suicidal ideation risk compared to non-caregivers. As health systems shift toward community- and home-based care, the mental health of family caregivers must be considered a key public health priority. Although all caregiver roles showed higher odds of suicidal ideation compared with the non-caregiver group, caregivers for IHD or stroke also had higher the risk of suicide ideation compared with the non-caregiver group. Both conditions often involve sudden life-threatening events and prolonged recovery periods. Caregivers must manage patient rehabilitation while coping with emotional trauma, which can lead to fatigue and psychological strain [18–20]. Although cancer caregivers also showed an elevated risk, their levels were comparatively lower. This may reflect more structured care pathways, clearer prognostic expectations, and broader psychosocial support for oncology settings [21]. These findings suggest that the psychological burden of caregiving varies according to the disease type. Dementia caregivers may benefit from grief support and education such as understanding dementia as a biological disease to reduce social stigma, whereas stroke and IHD caregivers require early phase psychological support during rehabilitation. Cancer caregivers, particularly those in late-stage care, should also be included in mental health monitoring. The heterogeneity in mental health risk across disease groups suggests that caregiver support should not follow a uniform approach. Dementia caregivers may require long-term anticipatory grief counseling, while stroke and IHD caregivers may benefit from early crisis-phase psychological support. Tailored mental health interventions integrated into disease-specific care pathways could be more effective than general caregiver support services.

Notably, compared to the non-caregiver reference group, the odds of suicidal ideation were numerically highest among caregivers who were parents of the patients. Previous studies have reported elevated levels of depression, anxiety, and posttraumatic stress symptoms in caregivers of young children with chronic or life-threatening illnesses [22]. The population comparison consisted of 106,005 exposed parents and 1,060,050 matched unexposed parents; there was a 15% increased risk of parental suicide attempts during the first years after a child’s cancer diagnosis, particularly when the risk was 1.6 times when the child was diagnosed with a highly aggressive cancer, or when the child died due to cancer [23]. Parental caregivers may experience intense emotional distress related to the fear of losing their child, guilt over their perceived inability to protect them, and chronic anticipatory grief [24]. In fact, parents caring for younger patients were themselves more likely to be younger, and caregiving for younger patients often involved more intensive daily care, which in many households is disproportionately undertaken by women [25]. These factors may partly explain the slightly higher odds of suicidal ideation observed among caregivers of female patients and younger patients. However, our study extends these findings by demonstrating that this profound psychological burden persists even when caring for adult children. This suggests that the intense emotional distress related to the fear of loss, guilt, and grief is not limited to parents of young children but continues into adulthood. In addition, these parents may experience social isolation, financial strain, and the disruption of family dynamics or employment [26]. Unlike other types of caregivers, they may also suppress their own mental health needs because of their protective instincts toward the child [27]. Targeted mental health interventions could be essential for parent caregivers, particularly those caring for children with chronic or terminal illnesses, regardless of the child’s age.

Among older adults and females, the caregiver group showed higher suicidal ideation compared to the non-caregiver group. Age [28], gender [25], and socioeconomic disparities [29] in caregiver burden are well-documented. Older caregivers have also been shown to experience greater strain owing to declining physical health and limited resources [28]. Older individuals may lack physical resilience and experience social isolation. Female caregivers reported significantly higher rates of depressive symptoms and unmet needs than male caregivers [25]. Women are often expected to assume caregiving roles and may face role strain due to balancing employment, caregiving, and household responsibilities. The old age and female was associated with economic insecurity. The economic insecurity further compounds the stress, limiting access to mental health services and respite care [29]. These results suggest the importance of prioritizing screening and support for female and elderly caregivers, particularly those with limited financial resources or those who are unemployed. Policy measures such as caregiver allowances or tax credits could alleviate some of this burden.

This study has several limitations. First, the cross-sectional design of the KNHANES dataset limits the ability to infer causality between caregiving and suicidal ideation. Longitudinal data are necessary to determine temporal relationships and identify the trajectories of psychological burden. Second, suicidal ideation and disease status were measured based on self-reported responses to a single survey item, which may have been subject to recall bias or underreporting due to social desirability. More comprehensive assessments, including clinical interviews and validated scales, may yield more precise estimates. Third, although the analysis was adjusted for several sociodemographic variables, unmeasured confounders such as caregiving duration, severity of patient illness, availability of social support, and caregiver health status were not accounted for and may have influenced the observed associations. Fourth, although caregiver burden has been consistently reported in the context of mental illness [30], mental illness could not be included in our disease categories. Because KNHANES does not collect physician-diagnosed psychiatric conditions and self-reported identification of severe mental illness is also challenging in community surveys, further limiting the ability to capture these conditions reliably. Future research using data sources with validated psychiatric diagnoses is needed to examine caregiving burden across mental and physical illnesses. Fifth, the caregiving role was self-identified, without confirmation of whether the identified individual was the primary caregiver and not verified using clinical or administrative data. This may introduce heterogeneity in the definition of caregiving and limit the comparability across studies. Finally, KNHANES includes only community-dwelling adults and excludes institutionalized populations such as elderly individuals in nursing homes or long-term care facilities, where caregiving dynamics and mental health outcomes may differ. Consequently, patients with severe conditions, particularly advanced dementia requiring full-time institutional care, may be underrepresented in our sample compared to national disease prevalence statistics. However, this exclusion aligns with the primary objective of our study, which is to examine the psychological burden of family caregivers who co-reside with patients. By focusing on community-dwelling households, our study specifically captures the experiences of caregivers who provide direct, daily support within the home environment, distinct from those managing care for institutionalized family members. Furthermore, this study was conducted in a single national context (South Korea), which may limit its generalizability to other countries with different healthcare systems, cultural expectations of caregiving, or mental health stigma.

Despite these limitations, this study has several strengths. We used nationally representative data and applied comprehensive subgroup analyses, enhancing both the generalizability and clinical relevance of the findings. By analyzing a diverse range of conditions including neurodegenerative, cardiovascular, and oncologic diseases, this study identified disease-specific patterns in caregiver psychological burden that have been underexplored. These findings highlight the importance of developing tailored interventions that address the unique stressors faced by high-risk subgroups, particularly parents and spousal caregivers, to reduce caregiver burden and improve mental health outcomes.

Author contributions

Formal analysis and data visualization were performed by SO and JL. The first draft of the manuscript was written by SO and DK. Writing – review and editing were performed by HJK, and JC. Methodology and conceptualization were led by DK and JC, with validation support from JL and HJK. All authors drafted and reviewed the manuscript.

Funding

This work was supported by the Korea National Institute of Health (KNIH) [grant numbers 2025-ER1001-01] and by the Bio&Medical Technology Development Program of the National Research Foundation (NRF) funded by the Korean government (MSIT) (No.RS-2024-00440881).

Data availability

The datasets underlying this article were derived from sources in the public domain: [K‐NHANES, [https://knhanes.kdca.go.kr/knhanes/main.do](https:/knhanes.kdca.go.kr/knhanes/main.do) ].

Declarations

Human ethics

The KNHANES was approved by the Institutional Review Board (IRB) of the KDCA (2007-02CON-04-P, 2008-04EXP-01-C 2009-01CON-03–2 C 010-02CON-21-C 2011-02CON06-C 2012-01EXP-01–2 C 2013-07CON-03–4 C 2013-12EXP-03–5 C), whereas the review was waived in 2015 according to Article 1 (1) of the Bioethics and Safety Act and Article 2 (2)-1 of the Enforcement Decree of the Bioethics and Safety Act in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

Competing interests

The authors declare no competing interests.