INTRODUCTION
Korea has experienced rapid population aging, with older adults representing a significant proportion of its population. Jeju Province is no exception, with older adults accounting for 19.0% of the population at the end of 2024 [1]. Jeju Province is on the verge of becoming a super-aged society, as this proportion is expected to exceed 20.0% soon. As the elderly population increases, understanding their health challenges is an increasingly important public health concern.
Older women are more vulnerable than their male counterparts because of their longer life expectancy, lower socioeconomic status, and disparities in access to financial and healthcare resources. Many women outlive their spouses, experience economic hardship, and have limited social support, elevating their risks of poor physical and mental health outcomes [2-5]. Of these, depression is particularly concerning. Women tend to experience depressive symptoms at nearly twice the rate of men throughout their lifespan [6-8]. In rural areas, such as Jeju, geographic and economic factors may further exacerbate these challenges [9-11].
Despite these well-documented vulnerabilities, sex differences in aging and mental health remain insufficiently explored [12,13], particularly in population-based studies. Existing studies often treat older adults as a homogeneous group, failing to capture the distinct differences between men and women, or between different age subgroups.
This study examined the health status and mental well-being of older adults in Jeju, Korea, focusing on sex- and age-related differences, with a particular focus on depressive symptoms. By identifying sex differences in the factors associated with depressive symptoms in older adults in Jeju, this study aimed to provide insights to inform more tailored interventions.
METHODS
1. Study population and data source
The Community Health Survey (CHS) has provided regional statistics on the health of adults aged 19 years and over since 2008. The survey population focuses on adults living in residential housing (stratification variable: apartment, general housing) at each sampling point in Tong/Ban/Ri at the time of the survey.
Raw data from the 2017 CHS were used in this study [14]. Among the 5,043 adults surveyed in Jeju Province, individuals aged under 65 years were excluded, resulting in 1,558 older adults. Subsequently, cases with incomplete information (responding ‘don’t know’ or ‘refused to answer’ for key variables, such as education, marital status, economic status, economic activity, dietary condition, and mental health) were excluded. Education is an important variable that is frequently used to measure socioeconomic status in epidemiological studies or health equity studies. Moreover, a known relationship exists between educational level and academic background, and health [15]. The final analysis included 1,471 older adults (606 men and 865 women). Participants were classified into early elderly (65-74 years) and late elderly (≥75 years) groups (Fig. 1).
This study was exempt from Institutional Review Board review (2025-03-016) as it analyzed secondary data from the CHS. The dataset did not include private information, such as patient name, social security number, address, or phone number. Only non-identifiable results were released and made openly available for public research purposes.
2. Assessment
All assessment tools used in this study were derived from the 2017 CHS. The tools are as follows [14].
1) Socioeconomic status
The variables representing socioeconomic status in the Jeju regional CHS included residence, marital status, education level, economic activity, economically vulnerable status, and dietary conditions. Residence was classified into two categories: urban area (Dong area) and rural area (Eup or Myoun area). Marital status was categorized as living with a spouse or not having a spouse (bereaved, divorced, or single individuals). Educational level was classified as no formal education, village school, elementary school, middle school, high school, university, or higher. Economic activity was grouped into employed and unemployed. Economic vulnerability was determined based on whether the individual was a recipient of basic livelihood support. Dietary conditions were assessed based on the household food security status. Household food security was evaluated using the following four response options: 1) My family members could eat a sufficient quantity and variety of foods as desired; 2) My family members could eat a sufficient quantity of food but not a variety of foods; 3) Due to financial difficulties, we sometimes did not have enough to eat; and 4) Due to financial difficulties, we often did not have enough to eat.
2) Health status
Disease prevalence was assessed by asking whether the respondent had experienced any illness or discomfort due to chronic or acute diseases, accidents, or poisoning within the past 2 weeks. Participants were also asked if they had ever been diagnosed by a physician with chronic conditions, such as hypertension, diabetes, dyslipidemia (including hyperlipidemia), or osteoarthritis, and whether they were currently receiving treatment for these conditions. Additionally, participants were asked about the following health-related experiences: chewing discomfort (difficulty or discomfort when chewing food due to dental-, denture-, or gum-related issues), fall accidents (incidents of slipping, tripping, collapsing, or falling within the past year), and bedridden conditions (a state requiring the participant to remain in bed for most of the day due to illness or injury within the past month).
3) Mental health status
Sleep duration was measured by asking the participants to report their average daily sleep time. Subjective stress levels were assessed using a single question regarding the extent to which the participants generally felt stressed. Responses were ‘feel very much’, ‘feel a lot’, ’feel a little’, and ‘hardly feel it.’
Nine depressive mood items from the Patient Health Questionnaire-9 (PHQ-9) were assessed by asking, over the past 2 weeks, how often have you been bothered by the following nine symptoms? Responses to each item were categorized as follows: presence of the symptom (several days, more than a week, or almost every day), absence of the symptom (never), and clinically significant depression was defined according to the CHS raw data guidelines, using a PHQ-9 total score of 10 or higher as the cutoff [14]. Based on the responses to the nine items, a binary variable, having depressive symptoms (depressive symptom prevalence), was created. Participants who reported experiencing at least one of the nine symptoms were classified as having depressive symptoms, while those who reported none of the symptoms were classified as not having depressive symptoms. The nine depressive symptoms assessed were as follows. 1) Little interest or pleasure in doing things (loss of interest in activities). 2) Feeling down, depressed, or hopeless (depressed mood, hopelessness). 3) Trouble falling asleep, staying asleep, or sleeping too much (sleep disturbance). 4) Feeling tired or having little energy (fatigue). 5) Poor appetite or overeating (eating disturbance). 6) Feeling bad about oneself, feeling like a failure, or feeling that one has let down oneself or their family (feelings of worthlessness). 7) Trouble concentrating on tasks, such as reading the newspaper or watching television (difficulty concentrating). 8) Moving or speaking so slowly that others notice, or, conversely, feeling so restless that it is noticeable to others (psychomotor changes). 9) Thoughts of being better off dead, or thoughts of self-harm in any way (suicidal ideation). Suicidal ideation was assessed by asking whether the participants had experienced thoughts of death or self-harm.
3. Statistical analyses
A complex sample design (stratification, clustering, and weighting) was applied to the analysis to ensure that the findings could be generalized to a population of 93,858 individuals aged 65 years and older in Jeju Province. Analyses were conducted within the framework of comparing socioeconomic status, physical health, and mental health characteristics according to age and sex. Rao-Scott chi-square tests and t-tests were used for univariate analyses. Logistic regression analysis was conducted to examine factors associated with depressive symptoms. The model demonstrated statistical significance, and the global null hypothesis (BETA=0) was rejected (P<0.001). All statistical analyses were conducted using SAS software (version 9.4; SAS Institute, Cary, NC, USA). Statistical significance was set at P<0.05.
RESULTS
The average age of older adults in Jeju was 69.3±0.19 years for men and 69.9±0.19 years for women in the early elderly group (65-74 years), and 79.9±0.48 years for men and 80.6±0.27 years for women in the late elderly group (≥75 years).
1. Socioeconomic challenges among elderly women in Jeju
In the early elderly group, 60.7% of men and 69.3% of women lived in urban areas. However, in the late elderly group, this pattern was reversed (63.6% for men and 46.8% for women), indicating that the proportion of elderly male residents surpassed that of elderly female residents.
The proportion of elderly women living with their spouses was significantly lower than that of elderly men. Among men, the spousal cohabitation rate was 92.1% in the early elderly group and 87.9% in the late-elderly group. In contrast, among women, the rate was markedly lower, at 68.3% in the early elderly group and 37.1% in the late-elderly group.
The proportion of elderly women with no formal education who attended traditional village schools (Seodang) or classical Chinese education was significantly higher than that of elderly men. Among women, this proportion was 17.1% in the early elderly group, increasing to 47.3% in the late elderly group. Among men, this was 2.6% and 4.9% in the early and late elderly groups, respectively.
The proportion of individuals engaged in economic activity (employed) was significantly higher in men (67.6%) compared to women (48.5%) in the early elderly group. In the late elderly group, the proportion decreased in both men (33.5%) and women (32.1%) and the difference between the two groups was no longer statistically significant.
Moreover, the proportion of basic livelihood recipients (classified by whether they were economically vulnerable or not) in the early elderly group did not differ significantly between men and women. However, in the late-elderly group, women were significantly more vulnerable. Household food security status followed the same pattern as the variable, basic livelihood recipients (Table 1).
2. Physical health challenges in elderly women in Jeju
The proportion of individuals who experienced illness in the past 2 weeks was significantly higher in men (23.7%) than in women (14.0%) in the late elderly group. The proportion of men diagnosed with diabetes was significantly higher than that of women. In the early elderly group, the prevalence was 23.3% in men and 15.6% in women, while in the late elderly group, it was 29.5% in men and 16.3% in women. Similarly, the proportion of individuals currently receiving treatment for diabetes was also significantly higher in men, with 22.5% of men and 14.6% of women in the early elderly group, and 18.5% of men and 15.9% of women in the late elderly group.
In contrast, the prevalence of dyslipidemia (including hyperlipidemia) was significantly higher in women than in men, but only in the early elderly group (26.4% in men and 39.9% in women). Similarly, the proportion of those currently receiving treatment was 20.6% in men and 31.5% in women in the late elderly group. Likewise, the prevalence of arthritis was significantly higher among women in all comparisons. In the early age group, arthritis affected 20.2% of men and 48.0% of women, whereas in the late-age group, 23.7% of men and 55.0% of women were affected. The proportion of individuals on current treatment was also significantly higher in women (10.8% of men and 34.4% of women in the early elderly group, and 20.7% of men and 38.6% of women in the late elderly group).
Experiences of chewing discomfort and fall accident did not differ significantly between the two sexes in the early and late elderly groups. However, approximately one in three individuals in the early elderly group and one in two individuals in the late-elderly group experienced chewing discomfort, indicating that its prevalence increased with age. The prevalence of fall accidents and being bedridden was higher in women than in men in both age groups. Notably, the prevalence of most health-related variables was higher in the late-elderly group than in the early elderly group (Table 2).
3. Mental health stability among late elderly women in Jeju
The average sleep duration was significantly shorter in women than in men (Table 3). The average daily sleep duration was 6.5 hours (standard error [SE], 0.10) for men and 6.1 hours (SE, 0.11) for women in the early elderly group, and 6.5 hours (SE, 0.12) for men and 6.2 hours (SE, 0.09) for women in the late elderly group.
The response distribution for subjective stress levels showed that elderly women experienced higher levels of stress than elderly men in the early elderly group. However, no significant difference was observed between sexes in either age group.
Although the results of the comparison by sex did not reach statistical significance in the early elderly group, the prevalence of depressive mood was 4.2% in men and 7.9% in women, whereas the prevalence of clinically significant depression was 1.4% in men and 2.8% in women. Similarly, in the late-elderly group, depressive mood remained more prevalent in women (4.5% in men vs. 6.0% in women). Notably, the prevalence of clinically significant depression was reversed between sexes, with higher rates observed in men (4.6%) than in women (3.3%).
Depressive symptoms were most commonly observed in the following order: fatigue, sleep disturbances, depressed mood/hopelessness, eating disturbances, and loss of interest in activities. In contrast, feelings of unhappiness, self-deprecating thoughts, anxious behavior, and difficulty in concentrating were relatively less common. In the early elderly group, symptoms, such as loss of interest in activities, depressed mood/hopelessness, sleep disturbance, and fatigue, were significantly more common in women than in men. Additionally, suicidal ideation was more common in women than men in both age groups, although with no significant differences (Table 3).
When comparing the early and late elderly groups among elderly men, almost all depressive symptoms were more prevalent in the late elderly group than in the early elderly group, with two symptoms (depressed mood/hopelessness and eating disturbance) showing statistically significant increases. Conversely, among women, six of the nine depressive symptoms improved, and three of these improvements (loss of interest in activities, depressed mood/hopelessness, and sleep disturbance) were statistically significant (Table 4).
4. Sex differences in factors associated with having depressive symptoms among older adults in Jeju
Multiple logistic regression analysis was performed to identify the factors associated with depressive symptoms in elderly women and men. Adjustments were made for age (late elderly group), low educational level (no formal education), marital status (widowed, divorced, single), low economic status (basic livelihood security recipient), place of residence (urban), and physical health conditions (diagnosis of hypertension, diabetes, dyslipidemia, hyperlipidemia, arthritis, and chewing discomfort).
Among women, those belonging to the late elderly group showed a significant protective association against depressive symptoms (adjusted odds ratio [aOR], 0.66; 95% confidence interval [CI], 0.45-0.98; P=0.041). Conversely, chewing discomfort (aOR, 1.67; 95% CI, 1.10-2.52, P=0.016) was identified as a significant risk factor for depressive symptoms. In contrast, no significant association was found among elderly men (Table 5).
DISCUSSION
This study investigated the socioeconomic status, health, and mental well-being of older adults in Jeju, Korea, by analyzing sex- and age-related differences, with a particular focus on depressive symptoms. The findings confirmed that older women were more vulnerable than men across multiple dimensions, including socioeconomic status (e.g., living in urban rather than rural areas, living without a spouse, lower educational attainment, lower labor force participation, and greater food insecurity), physical health (e.g., dyslipidemia and arthritis), and mental health (e.g., shorter sleep duration and depressive symptoms; see Tables 1-3 for descriptive differences by sex and age group). These examples reflect patterns observed in the descriptive statistics and are not intended to represent statistically significant associations in a multivariable analysis. Contrary to expectations, late-old women did not exhibit the highest levels of depressive symptoms. Instead, their symptoms were comparable to those of early aged women and seemed to be even better than those of late elderly men, who showed significantly higher depressive symptoms. These results suggest that resilience may play a role in mitigating mental health decline among late elderly women, despite their greater vulnerability.
Our findings are consistent with the results of previous research, indicating that older women generally face greater socioeconomic and physical health vulnerabilities than men, including lower educational attainment, higher economic insecurity, and a higher prevalence of chronic conditions, such as arthritis and chewing discomfort [16-18].
Many studies have consistently reported that women across all ages are at a higher risk for depressive symptoms than men [6,7]. A key novel aspect of this study is the unexpected pattern observed among late elderly women. In particular, among late elderly women, depressive symptoms were not significantly higher than those of early elderly women and were lower than those of late elderly men. After adjusting for marital status, education, place of residence, economic status, hypertension, diabetes mellitus, dyslipidemia, arthritis, and chewing difficulties, being in the late elderly age group was identified as a significant protective factor against depressive symptoms (aOR, 0.66; 95% CI, 0.45-0.98; P=0.041) only among women. Women had the only risk factor of shewing discomfort (aOR, 1.62; 95% CI, 1.10-2.52; P=0.016).
Previous studies have reported mixed findings regarding the relationship between aging, sex, and mental health. Some studies have suggested that the accumulation of social and physical disadvantages over time leads to worsened mental health in older women [8,9,19]. However, other studies have found that older adults may demonstrate psychological resilience, contributing to the maintenance of better mental well-being despite increasing vulnerabilities [20-23]. A nationwide register-based cohort study in Sweden compared the standardized incidence rate of clinically diagnosed psychiatric disorders between men and women across the lifespan. While depressive disorders were higher in women across all age groups, this sex difference disappeared after 80 years of age. Similarly, anxiety disorders, which are known to have a higher incidence in women across all age groups, were similar in both men and women after mid-50s, and stress-related disorders after 80 years did not demonstrate a significant sex difference [24].
Our study provides empirical evidence supporting the latter perspective. In particular, among late elderly women, our results suggest that resilience mechanisms may buffer against the observed higher levels of depressive symptoms in men despite the vulnerability of late elderly women.
Another important result of this study is that the factors associated with depressive symptoms among older adults in Jeju, Korea differed between different sex and age groups. Given that different population subgroups encounter distinct risk factors, interventions should be tailored and prioritized accordingly, particularly in resource-limited settings [12].
One limitation of this cross-sectional observational study is that the findings related to late elderly women cannot be interpreted causally in terms of reducing depressive symptoms. However, despite adjusting for socioeconomic and physical health factors associated with depressive symptoms, a significant protective effect in late elderly women remained. We hypothesized a potential explanation for this finding.
First, late elderly women may experience a reduction in and simplification of their social roles. Adapting to these changes may effectively alleviate depressive symptoms [25]. Late elderly women (75 years and older) may have already been freed from the burden of caregiving, such as losing a spouse or having independent children. Research suggests that women’s vulnerability to depression is not due to a lack of coping ability or social support compared to men, but rather, because women tend to be more emotionally involved in the lives of others, leading to greater emotional costs, which contribute to higher stress and suffering [26]. In the late elderly time of life, women’s emotional investments in important individuals may be more limited. Consequently, the incidence of depressive symptoms in late elderly women also lowers. Several studies have highlighted that women tend to be more emotionally affected by others’ lives. Research indicates that women are more psychosocially flexible and tend to react more sensitively to situations by utilizing both problem- and emotion-focused coping strategies more effectively [27]. These characteristics may positively impact the reduction in depressive symptoms. In contrast, older elderly men may experience increased loneliness and social isolation following the death of their spouse, which could contribute to higher depressive symptoms. This finding has been reported to show a stronger association among depressive symptoms, disabilities, and mortality in men than in women [28].
Second, late elderly women seem to maintain a sense of self-control in the face of the natural health decline associated with aging. Research has shown that an increased sense of self-control leads to greater subjective well-being [29] and life satisfaction [30], and even elderly individuals in long-term care facilities can regain autonomy [31]. Mental health outcomes do not consistently align with these risks, suggesting the potential role of resilience. Among aging populations, resilience has been conceptualized as the process of adapting to challenges through individual and social resources. Resilience refers to the ability to maintain or recover mental well-being despite adversity [32,33]. Studies indicate that resilience can buffer the negative effects of stress and protect against mental health deterioration [34,35].
This study focused on a local culture unique to Jeju Province in Korea. As such, our results may not be generalizable to other populations. Further research in other populations is required to verify the generalizability of our findings.
One of the study’s strengths is its population-based approach, which enables a more representative analysis of older adults than studies that rely on clinical or convenience samples. Additionally, by differentiating between early and late elderly age groups, this study provided a more nuanced understanding of mental health outcomes in older women, a factor often overlooked in previous research. Furthermore, the use of a comparative framework that included men and women offered valuable insights into sex-based disparities and highlights, in particular, how mental health outcomes differ between men and women across various early and late elderly age groups. Our findings underscore the need for longitudinal studies that can account for temporal associations and causality alongside repeated cross-sectional studies that explore descriptive statistics from continuous observations.
This study highlighted the vulnerability and potential resilience among the elderly living in Jeju. The finding that late elderly women did not experience the expected increase in depressive symptoms challenges the assumption that accumulated disadvantages lead to worsened mental health. A previous meta-analysis of 300 studies suggested that women are less happy or satisfied than men at all ages, with this gap widening after middle age [36]. In contrast, our findings raised the possibility that older women may maintain or experience better mental health as they enter the late elderly stage of life. However, this interpretation warrants cautious consideration. Further longitudinal research is required to explore the hypothesis that resilience strengthens with advancing age among older adults. Interventions informed by these findings are crucial for shaping targeted strategies. Importantly, intentions should focus on risk reduction, as well as reinforce resilience factors among elderly women.
This study found that although elderly women in Jeju, Korea, exhibited greater socioeconomic and physical health vulnerabilities, yet late elderly women showed lower than expected levels of depressive symptoms. Additionally, difficulty chewing was associated with depressive symptoms in women. In elderly men, no risk factors for depressive symptoms were identified. Notably, being in the late elderly stage of life was protective against depressive symptoms. Overall, given the distinct risk factors faced by different population subgroups, interventions should be tailored and prioritized accordingly, particularly in resource-limited settings.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
