Scholarly article on topic 'Structural Equation Modeling to Assess Discrimination, Stress, Social Support, and Depression among the Elderly in South Korea'

Structural Equation Modeling to Assess Discrimination, Stress, Social Support, and Depression among the Elderly in South Korea Academic research paper on "Psychology"

CC BY-NC-ND
0
0
Share paper
Academic journal
Asian Nursing Research
OECD Field of science
Keywords
{ageism / depression / "life stress" / "social support"}

Abstract of research paper on Psychology, author of scientific article — Hung Sa Lee, Chunmi Kim

Summary Purpose The purpose of this study was to find the relationship and conceptual model of discrimination, stress, support, and depression among the elderly in South Korea. Methods This was a cross-sectional descriptive study involving 207 community-dwelling elders. Data were collected through questionnaires from May 5 to May 31, 2014 in community senior centers, and analyzed using descriptive statistics, t test, analysis of variance, Scheffé test, and structural equation modeling. Results There were significant effects of discrimination on stress, support on stress and stress on depression. Moreover, there were two significant indirect effects observed between discrimination and depression, and between support and depression. For each indirect effect, the mediating factor was stress. Additionally, there was no direct effect between discrimination and depression or support. Conclusions This study found that social support and discrimination had indirect effects on depression through stress. More specifically, decreased stress led to a reduction of depression. Therefore, social support based on a thorough understanding of stress is very important for caring elderly who are depressive.

Academic research paper on topic "Structural Equation Modeling to Assess Discrimination, Stress, Social Support, and Depression among the Elderly in South Korea"

Asian Nursing Research xxx (2016) 1—7

Korean Society of Nursing Science

Contents lists available at ScienceDirect

Asian Nursing Research

journal homepage: www.asian-nursingresearch.com

Research Article

Structural Equation Modeling to Assess Discrimination, Stress, Social Support, and Depression among the Elderly Women in South Korea

Hung Sa Lee, PhD, RN, Chunmi Kim, PhD, RN

Department of Nursing, Sunmoon University, Asan, South Korea

ARTICLE INFO

Article history:

Received 25 September 2015 Received in revised form 20 April 2016 Accepted 2 May 2016

Key words: ageism depression life stress social support

SUMMARY

Purpose: The purpose of this study was to find the relationship and conceptual model of discrimination, stress, support, and depression among the elderly in South Korea.

Methods: This was a cross-sectional descriptive study involving 207 community-dwelling elders. Data were collected through questionnaires from May 5 to May 31, 2014 in community senior centers, and analyzed using descriptive statistics, t test, analysis of variance, Scheffe test, and structural equation modeling.

Results: There were significant effects of discrimination on stress, support on stress and stress on depression. Moreover, there were two significant indirect effects observed between discrimination and depression, and between support and depression. For each indirect effect, the mediating factor was stress. Additionally, there was no direct effect between discrimination and depression or support. Conclusions: This study found that social support and discrimination had indirect effects on depression through stress. More specifically, decreased stress led to a reduction of depression. Therefore, social support based on a thorough understanding of stress is very important for caring elderly who are depressive.

Copyright © 2016, Korean Society of Nursing Science. Published by Elsevier. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction

There is a growing interest in the relationship between social support and depression among older adults in South Korea. Depression in late life may be a risk factor for dementia or elderly suicide [1]. Therefore, the Visiting Nurse Associations in Korea has supported depressive elders in the community. The depression rate of senior citizens at age 65 or older exceeded 34.0% in 2009, and it was higher for women than men [2]. According to Statistics Korea [3], this phenomenon has increased medical expenses for the elderly which have risen to more than 30.0% of all South Korean medical expenses. Welfare costs, including financial aids and long term care services, have also increased rapidly since 2008 [4]. In dealing with these rising costs, the budget for elderly facilities may not be used effectively if nursing services are provided without a thorough understanding of the elderlys' needs. Furthermore, healthcare cannot be effectively provided without an

* Correspondence to: Chunmi Kim, PhD, RN, Department of Nursing, Sunmoon University, Asan, #70, Sunmoon-ro 221 beon-gil, Tangjeong-myeon, Asan-si, Chungnam, 336-708, South Korea.

E-mail address: spring4cmk@gmail.com

understanding of older adults' psychological states. Among the psychological problems that afflict elderly individuals, depression is an important issue. Much research on depression among the elderly along with related factors such as stress and social support has been conducted. Social support has been found to mitigate elderly depression while stress has been found to be a risk factor that increased elderly depression [1,5—8]. Elderly stress is believed to come from life-related factors such as the spouse's death and restrictions in the activities of daily living resulting from diseases. On the other hand, it has been reported recently that the elderly women who have experienced discrimination suffer a high level of stress [9—11]. Thus, if experience in discrimination aggravates stress, elderly discrimination is also likely to influence depression in the aged. However, studies on the relationship between discrimination and depression are relatively rare.

Discrimination against the elderly women that isolates individuals from society is one of the most important factors to be considered when developing a social support system for senior citizens [9,12]. Many of the elderly have experiences of discrimination in social institutions that have been built without regard for their needs. Palmore [13] claimed that discrimination against senior citizens, known as ageism, is one of the worst forms of

http://dx.doi.org/10.1016/j.anr.2016.06.003

p1976-1317 e2093-7482/Copyright © 2016, Korean Society of Nursing Science. Published by Elsevier. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

H.S. Lee, C. Kim / Asian Nursing Research xxx (2016) 1—7

prejudice along with racism and sexism. However, while racism and sexism are related to specific groups, any individual can experience ageism if they live long enough [14]. People, who experience discrimination, are driven to feel that society has no place for them, and this negative feeling can manifest as severe stress [10,11]. According to Ezzati et al [15], when an individual is not able to deal with this condition, stress causes diverse symptoms, one of which is depression. In other words, ageism increases the elderly's stress levels and in turn causes depression [1,16]. Since depression is a major cause of social and health problems such as suicide and various psychological disorders, it will worsen as ageism increases. Therefore, it is important to analyze how experiencing discrimination affects stress and depression among the elderly to promote not only the health of senior citizens but also that of the entire society.

Research on depression and its related factors, such as social demographic factors, social support and stress, has been conducted for a long time. However, the relationship between discrimination and depression among South Korean elderly women has been studied only on a limited basis. Moreover, the overall interaction among discrimination, stress, social support and depression has not yet been investigated. Therefore, the current study was conducted to assess not only the interactions between elderly women discrimination (ageism), social support, stress and depression but also the magnitude of the interactions among these factors. In particular, the way in which social support helps reduce depression was highlighted. The specific goals of this research were (a) to gain an understanding of the elderly population's experience of discrimination, stress, social support and depression in terms of social demographic factors; (b) to establish a research model demonstrating the overall paths and interactions among discrimination, stress, support and depression; and (c) to identify causal factors by examining the direct, indirect and total effects of each variable. Based on these goals, the following hypotheses were developed: (a) discrimination increases stress; (b) stress worsens depression; (c) social support reduces stress; and (d) the reduction in stress leads to decreases in depression among the elderly women.

Methods

Design and sample

This cross-sectional survey study was conducted to examine the relation of discrimination, stress, support and depression in elderly women. For this, elderly women, who lived in rural areas and attended community senior centers, were selected by convenient sampling. All of the study participants lived for more than 10 years in K province (Gyeongsangbuk-Do). Eligibility criteria included being at least 65 years old without cognitive impairment, having the ability to communicate verbally in Korean and consenting to participate in this study.

The generally recommended sample size for structural equation modeling analysis is 150—400 [17]. Thus, in consideration of dropouts, this study distributed questionnaires to 220 elders, and used 207 of them as valid data, satisfying the requirement of sample size.

Procedures

For the survey, letters requesting cooperation were sent to the directors of all the senior centers in K Province with the help of the social welfare department of the provincial office. As 23 of the centers agreed to participate in the research, the researcher obtained a prior permission from the director of each senior center by phone and visited the center. The purpose of the research was explained to the elderly members in the senior center, and then

their written consent was obtained. Data were collected by the researchers and 10 assistants from May 5 to May 31, 2014 in community senior centers. The research assistants were nurses, who had worked as home visiting nurses for elderly for more than 3 years. The self-reported questionnaires were completed by the participants. If the participants were illiterate, research assistants helped them. The assistants had been educated about interview methods for 8 hours by the researchers.

Ethical consideration

The study was approved by the Institutional Review Board of the researcher's institution (IRB; SMU-14-04-05), and performed in accordance with the Declaration of Helsinki. To protect the participants' human rights, permission to collect data was granted from the director of the relevant facility. Before the survey was administered, the purpose and process of the study, a guarantee of confidentiality, voluntary participation, anonymity of data and their right to discontinue participation at any time during the process were explained to the director of the community senior centers and potential participants. The researchers explained the participation details and survey questionnaires to the participants who decided to join in the survey and completed a written consent form. For the protection of the participants' personal information, the survey did not ask about information related to personal identity, and collected data were discarded when the study was finished.

Instruments

Discrimination was assessed in terms of the typology of ageism developed by Palmore [13]. The typology was translated into Korean by Won [11] and consisted of 20 items, being divided into three groups that evaluated ignorance, avoidance and negligence. Each item was rated from 1 point to 4 points according to the Likert scale. Respondents indicated how often they experienced a particular event as 1 (never), 2 (rare), 3 (often), and 4 (very often). The index for ignorance consisted of six items on verbal humiliation, ignorance of the individual's words, behavior, and opinions, and a lack of attention to the elderly. The avoidance index contained seven items in terms of conversation and cooperation. The index for negligence had seven items asking whether the participants felt neglected in social settings. When calculating the score for each index, the average item score was used. The higher the score was, the more an individual experienced ageism. As evidence of the instrument's reliability, Cronbach a was .93 in this study.

The level of stress was assessed using a stress recognition scale for the elderly developed by Lee and Lee [6]. The instrument contained a total of 21 items, which were regrouped into four sub-categories by the sources of stress during daily life. The four subcategories were the following: stress derived from family relationships (9 items), economical status (5 items), health-related issues (4 items) and home environment factors (3 items). Each item was rated from 1 to 4 points; 1 (never), 2 (rarely), 3 (sometimes), and 4 (often). The average scores of items were calculated for the four subcategories. Higher average scores indicated greater stress levels. As evidence of reliability of this instrument, Cronbach a was .85 in this study.

Social support was measured by the Multi-dimensional Scale of Perceived Social Support, which was developed by Zimet, Dahlem, Zimet and Farely [18], and translated into Korean by Ko and Seo [7]. The scale includes three subcategories that measure neighborhood support, support from friends and support from family. Responses for items in these subcategories include seven choices ranging from "strongly disagree" (1 point) to "strongly agree" (7 points). The average score for the items in each subcategory was used for data

H.S. Lee, C. Kim/ Asian Nursing Research xxx (2016) 1—7

analysis with higher average scores indicating more social support for the individual. As evidence of reliability of this instrument, Cronbach a was .95 in this study.

To assess depression, the Geriatric Depression Scale Short Form (GDS-SF), developed by Sheikh and Yesavage [19] and translated by Jang, Kim and Chiriboga [20], was used in this study. The Korean translation of GDS-SF was also used for the Lee and Kim's [21] study of Korean-American senior citizens. The GDS-SF contained 15 items with questions such as the following: "Is your life satisfactory?" "Did you quit doing what you have done so far?" and "Do you experience anxiety about whether something bad will happen to you?" When grading the participants' responses, "Yes" was counted as 1 point and "No" as 0. The total score of every item for each participant ranged from 0 to 15. Greater total scores indicated higher levels of depression. As evidence of reliability of this instrument, Cronbach a was .84 in this study.

Data analysis

The data were analyzed using SPSS and AMOS software (IBM, New York, NY, USA) for Windows version 21.0. Descriptive statistics such as frequency, percentage, mean and standard deviation were used to depict the characteristics of the study. To compare differences in measured variables (discrimination, support, stress and depression) by social demographic factors, t test, analysis of variance, and Scheffe test were used. To examine the research model fit and relationship between the variables, a structural equation modeling (SEM) analysis with AMOS was used.

Results

Demographic characteristics and differences of discrimination, stress, support and depression

The mean age of the participants was 77.42 years. Approximately half (50.2%) of the respondents had no official education,

while 43.0% had graduated from elementary school. About a quarter (27.6%) of the respondents was in the low income level and 55.6% were in the middle economic status. In total, 45.9% of the respondents lived alone, 31.9% lived with their spouses only, and 22.2% lived with their offspring. In terms of perceived health status, 51.7% were in poor health, and 23.2% were in good health (Table 1).

Differences of dependent variables according to the respondents' sociodemographic characteristics were shown in Table 1. There were statistical differences in the scores of stress and depression by education and health level. For participants with poor health, the stress score was significantly higher (F = 3.76, p = .025), and the depression score was significantly higher (F = 4.83, p = .009) than those of participants with good health. The discrimination score was significantly lower in the poor health group (F = 5.77, p = .004), but there was no statistical difference in the support score. These results indicated that the perceived health status significantly affected discrimination, stress, and depression.

No difference was found in the discrimination scores grouped by economic status, but there were significant differences in the stress, support and depression scores. The participants in the low economic status had higher stress score (F = 22.40, p < .001), lower support score (F = 9.13, p < .001) and higher depression score (F = 18.55, p < .001) than the participants in the middle or high economic status. The depression score of the participants with no education was higher than the scores of the participants with an elementary or junior high school educational level (F = 5.03, p = .007). This finding demonstrated that the economic status made a difference in the level of stress, support, and depression, while educational level made a difference only in the level of depression.

SEM analysis of research hypothesis

A confirmatory factor analysis (CFA) was conducted for the research variables. Model fitness for the CFA in this study was conducted according to model fit indices, absolute fit indices, and incremental fit indices for the SEM. This included the ratio of chi-

Table 1 Dependent Variables According to Participants' Sociodemographic Characteristics (N=207).

Characteristics n (%) or Mean ± SD Discrimination Stress Support Depression

(Mean ± SD) (Mean ± SD) (Mean ± SD) (Mean ± SD)

Age (yr) 77.42 ± 5.4

Education level

No education 104 (50.2) 1.35 ± 0.39 1.99 ± 0.43 5.38 ± 1.50 6.40 ± 4.16(a)

Elementary school 89 (43.0) 1.46 ± 0.52 2.01 ± 0.46 5.59 ± 1.27 4.89 ± 3.71

Junior high school 14 (6.7) 1.43 ± 0.32 1.83 ± 0.29 6.04 ± 1.19 3.85 ± 2.82 (b)

F (p) 1.35 (.260) 1.02 (.362) 1.60 (.204) 5.03 (.007**) a>b

Economic status

Low 57 (27.6) 1.42 ± 0.39 2.28 ± 0.39 (a) 4.87 ± 1.51 (a) 8.05 ± 3.91 (a)

Middle 115 (55.6) 1.41 ± 0.49 1.91 ± 0.41 (b) 5.78 ± 1.31 (b) 4.86 ± 3.80 (b)

High 35 (16.8) 1.34 ± 0.39 1.77 ± 0.33 (b) 5.67 ± 1.11 (b) 3.94 ± 2.72 (c)

F (p) 0.36 (.698) 22.40 (< .001**) 9.13 (< .001**) 18.55 (< .001**)

a > b a<b a>b>c

Cohabitation

Live alone 95 (45.9) 1.34 ± 0.41 2.00 ± 0.41 5.53 ± 1.32 6.08 ± 4.09

With spouse only 66 (31.9) 1.45 ± 0.50 1.95 ± 0.43 5.65 ± 1.41 5.28 ± 3.82

With children 46 (22.2) 1.45 ± 0.46 2.04 ± 0.49 5.26 ± 1.50 4.97 ± 3.90

F (p) 1.34 (.262) 0.63 (.533) 1.13 (.324) 1.47 (.231)

Perceived health status

Poor 107 (51.7) 1.30 ± 0.30 (a) 2.06 ± 0.40 (a) 5.42 ± 1.43 6.29 ± 4.20 (a)

Middle 52 (25.1) 1.51 ± 0.55 (b) 1.96 ± 0.47 5.69 ± 1.42 5.38 ± 3.61 (b)

Good 48 (23.2) 1.51 ± 0.55 (b) 1.86 ± 0.45 (b) 5.53 ± 1.28 4.20 ± 3.50 (b)

F (p) 5.77 (.004**) 3.76 (.025*) 0.67 (.512) 4.83 (.009**)

a<b a>b a>b

Total 207(100) 1.40 ± 0.45 1.99 ± 0.43 5.51 ± 1.39 5.58 ± 3.97

Note. *p < .05. ** p < .01.

H.S. Lee, C. Kim / Asian Nursing Research xxx (2016) 1—7

square statistic to degree of freedom (c2/df), goodness fit index (GFI), root mean square error of approximation (RMSEA), normed fit index (NFI), Tucker-Lewis index (TLI) and comparative fit index (CFI).The chi-square value for the CFA model was 112.73 (p < .001) and the degree of freedom for the model was 39. The fitness indices for the CFA model demonstrated that the model fit was good. The absolute fit indices were .91 (GFI), 2.89 [minimum value of the discrepancy function (CMIN)/degree of freedom (DF)] and .09 (RMSEA; low 90: .07; high 90: .11), while the incremental fit indices were .91 (NFI), .91 (TLI) and .93 (CFI). As goodness-of-fit is acceptable if GFI, NFI, TLI, and CFI are over .90 and RMSEA is under .10, this study used these indexes for assessing the goodness-of-fit of the model [13].

In terms of the convergent validity of the research constructs, the reliability of each latent variable (discrimination, support and stress) was greater than .70 (.87 for discrimination, .91 for support and .74 for stress). In addition, the average variance extracted of each latent variable was .69 for discrimination, .77 for support and .42 for stress. With regard to the discriminant validity between each research construct, values for all significant correlations between the constructs were under .85 as shown in Figure 1.

Moreover, the correlation between the discrimination and support constructs was not significant (p = .211).

When examining the nomological validity of the constructs, all the statistical significance and signs of the correlations corresponded to the research hypothesis, which were (a) discrimination increases stress in the elderly; (b) stress enhances depression; (c) social support reduces stress; and (d) the reduction in stress leads to decreases in depression. These findings indicate that there is nomological validity between pairs of constructs.

Results of the analysis of the hypothetical model are presented in Figure 2, Table 2, and Table 3. The model fit was statistically examined first. Then, the factor loading values, direct effects, indirect effects and total effects between combinations of constructs were investigated. In terms of the model fit, the chi-square value of the research model was 120.41 (p < .001) and the degree of freedom was 40. Absolute fit indices demonstrated that this research model was statistically well developed: 3.01 (CMIN/DF), .90 (GFI) and .09 (RMSEA; low 90, .07; high 90, .11). These results were also supported by incremental fit indices: .90 (NFI), .93 (CFI) and .90 (TLI).

Figure 1. Confirmatory factor analysis and standardized factor loading values. Note. Model fit analysis by structural equation modeling; df = 39, c2 = 112.73 (p < .001), c2/ df = 2.89, GFI = .91, RMSEA = .09 (low 90: .07; high 90: .11), NFI = .91, CFI = .93, TLI = .91. CFI = comparative fit index; GFI = goodness of fit index; NFI = normed fit index; RMSEA = root mean square error of approximation; TLI = Tucker-Lewis index.

H.S. Lee, C. Kim/Asian Nursing Research xxx (2016) 1—7

Figure 2. Research model and standardized factor loading values. Note. D1 = disturbance of stress; D2 = disturbance of depression; D3 = disturbance of support. Hypothetical model fit analysis by structural equation modeling; df = 40, x2 = 120.41 (p < .001), x2/df = 3.01, GFI = .90, RMSEA = .09 (low 90: .07; high 90: .11), NFI = .90, CFI = .93, TLI = .90. CFI = comparative fit index; GFI = goodness of fit index; NFI = normed fit index; RMSEA = root mean square error of approximation; TLI = Tucker-Lewis index.

The research model that showed the standardized regression weights between constructs was presented in Figure 2. In terms of the direct effects as shown in Table 2, the standardized regression weight of .47 (p < .001) from discrimination to stress was observed. The standardized factor loading value between support and stress was —.43 (p < .001). The standardized regression weight from stress to depression was .64 (p < .001). However, the standardized factor loading value from support to depression was not statistically significant.

As shown in Table 3, there were two indirect effects in this research model: standardized value from discrimination to depression was .34 and standardized value from support to depression was —.28. The mediating factor of each indirect effect path was stress. The total and indirect effects between discrimination and depression were the same, because no direct effect

Table 2 Standardized Path Coefficients between Variables.

Constructs Dependent variables Independent variables Regression weights SE CR P

Depression Stress .64 .11 7.35 < .001

Support -.08 .06 -1.18 .238

Stress Discrimination .47 .07 6.04 < .001

Support -.43 .05 -5.79 < .001

Variables r SE CR P

Discrimination & DS -.92 1.16 -1.22 .223

Note. CR = critical ratio; DS = disturbance of support; SE = standard error.

between discrimination and depression was observed (Table 3). Since there was no statistically significant direct effect between support and depression (Table 2), the real total magnitude of the effect between these variables was —.28. There was an insignificant direct effect in total effect (Table 3). These results imply that discrimination and support affected depression only indirectly through stress in this model. The correlation between discrimination and support (discrimination and D3) was not statistically significance (Table 2).

Discussion

This study found that the elderly women's discrimination increases stress and increased stress causes depression. More importantly, stress is reduced by social support among elderly women. Additionally, reduced stress decreases depression. Since

Table 3 Direct, Indirect, and Total Effects between Each Construct.

Constructs Direct effects Indirect effects Total effects

Dependent Independent

variables variables

Depression Stress .64** .00 .64

Support -.08 -.28** -.36

Discrimination .00 .34** .34

Stress Discrimination .47** .00 .47

Support -.43** .00 -.43

Note. **p < .01.

H.S. Lee, C. Kim / Asian Nursing Research xxx (2016) 1—7

senior citizens live in the current social system, which is mainly designed for younger people with better health and greater economic power, the elderly experience more discrimination and tend to lack social privileges [22]. Discrimination against elderly people increases their stress, leading to higher levels of depression. In particular, women may feel more discrimination than men do in the patriarchal Korean society.

In the descriptive statistical analysis, the discrimination score was 1.40, which was lower than the score reported by Won [11]. The reason for this difference can be that participants of this study were females living in rural areas while those in the investigation by Won [11] were both males and females living in urban areas. There is a greater possibility of experiencing discrimination in urban areas than in rural areas, since the urban social system is more accommodating for young working men in Korea in the 21st century [11]. It is difficult to make further comparisons between the results of the present investigation and those of many other related studies due to the lack of research on discrimination against the elderly in Korea. As Palmore [13] noted that it is difficult to compare his data on ageism with findings from previous publications given the lack of studies on age discrimination in the United States, additional investigations of discrimination against the elderly should be conducted in both countries.

Stress is a factor that greatly affects psychological and physical health [15]. The mean score of stress in this research was 1.99. This score was higher than the stress scores reported by Lee and Lee [6] as well as by Youn and Kim [23]. The reason of the higher stress level in this study than in other studies seems to be participant's sex. This study was conducted with only female elderly while the previous two studies were carried out with both males and females. In the previous studies, stress in females was greater than that in males. In this study, participants living in poorer economic conditions or worse states of health experienced greater stress. This result corresponds to findings from previous studies [6,23,24]. The results imply that the health is threatened if there is no strategy for relieving their stress since stress significantly affects depression.

According to the SEM analysis of the paths of the constructs' interactions based on the research hypotheses, it was found that social support buffers the effect of stress on depression. Given that many studies have demonstrated that depression levels increase as stress levels rise [6,25,26], it is clear that stress is an influential factor for depression. However, the relationship between depression and social support has varied in previous studies. Kim et al [26] along with Lee and Lee [6] claimed that there were negative correlations between social support and depression. Zhu, Hu and Efird [8] as well as Ko and Seo [7] contended that depression had a significant negative relationship with family support but not with the support of neighborhood or friends. In contrast, Brown et al [12] reported that the support of friends rather than family was found to correlate with depression.

Based on these previous studies, it is unclear whether depression and social support have a significant direct relationship. In this study, social support for the elderly was found to indirectly reduce their depression by decreasing stress. In other words, social support targeted at addressing the stress factors of elderly people can effectively reduce depression. In a previous study, support from friends rather than family was discovered to reduce depression more when a person loses his/her spouse because of death [24]. This finding implies that social support based on an understanding of the target population should be offered.

Findings of this study imply that the social support for the elderly should be based on an understanding of stress factors that affect senior citizens. If social support without considering the individual's need is provided to elderly people, their depression is not likely to decline. This is of particular interest since data from this

investigation indicated that social support for the elderly provides a buffer against depression by attenuating stress levels. This consideration is crucial for developing effective nursing care strategy for the elderly.

Conclusion

This study has provided meaningful data for improving the health of senior citizens by defining how discrimination, stress, social support and depression interact with each other among Korean senior citizens. Among these health-related factors, depression was found to influence the health of the elderly women significantly. It was also shown that social support and discrimination have an indirect effect on depression through stress, a mediating factor. Furthermore, it was discovered that social support decreases stress, and decreased stress leads to a reduction of depression. Some of the causal relations between health-related factors were congruent with the findings of previous studies and the others were not because of the characteristics and sampling methods of participants. In conclusion, establishing nursing care system for senior citizens based on a thorough understanding of depression-related factors is very important, when developing strategies and policies for improving the health of the elderly women.

The limitation of this research was that the study was conducted using only elderly South Korean females living in rural regions. Thus, the study findings cannot be applied to all senior citizens. Further studies should therefore be conducted under various conditions taking into account gender, race, and areas of residence. This study has another limitation in that it used a convenience sampling method. Therefore, further research in a broader area, in a wider range of facilities and with a rigorous sampling method is necessary to generalize the results of this study. Nevertheless this study is significant in that it investigated the influence of discrimination and depression in a patriarchal country. The results of this study will be of use to countries that are in the process of developing social support in the nursing system.

Conflict of interest

The authors have no conflict of interest to declare.

Acknowledgments

We gratefully acknowledge financial support from the Academic Research fund of Sunmoon University in 2014.

References

1. Dickinson WJ, Potter GG, Hybels CF, McQuoid DR, Steffens DC. Change in stress and social support as predictors of cognitive decline in older adults with and without depression. Int J Geriatr Psychiatry. 2011;26:1267—74. http://dx.doi.org/10.1002/gps.2676

2. Statistics Korea. Population projections for Korea: 2010—2060 (Based on the 2010 Census). Seoul (Korea): Korean Statistical Information Service; 2012.

3. Statistics Korea. 2012 elderly statistic. Seoul (Korea); 2014 [cited 2014 Jun 14]. Available from: http://kostat.go.kr/portal/korea/kor_nw/2/6/1/index.board? bmode=read&aSeq=260368

4. Kim Y. [A study on the economic effect of the welfare for the aged budget expenditures]. J Welf Aged. 2011;51:101—23. Korean.

5. Nadimpalli SB, James BD, Yu L, Cothran F, Barnes LL. The association between discrimination and depressive symptoms among older African Americans: the role of psychological and social factors. Exp Aging Res. 2015;41(1):1—24. http://dx.doi.org/10.1080/0361073X.2015.978201

6. Lee S, Lee K. [The stress, social support and psychological well-being of the elderly]. Hanguk Nonyonhak. 2002;22(1):1—20. Korean.

7. Ko M, Seo I. [Influence of the elderly's health status upon their stress and depression and moderating effects of social supports]. Korean Public Health Res. 2011;37(1):1—14. Korean.

H.S. Lee, C. Kim/Asian Nursing Research xxx (2016) 1—7

8. Zhu S, Hu J, Efird J. Role of social support in cognitive function among elders. J Clin Nurs. 2012;21:2118—25. http://dx.doi.org/10.1111/j.1365-2702.2012.04178.x

9. Han J, Richardson VE. The relationships among perceived discrimination, self-perception of aging, and depressive symptoms: a longitudinal examination of age discrimination. Aging Ment Health. 2015;19(8):747—55. http://dx.doi.org/10.1080/13607863.2014.962007

10. Graham N, Lindesay J, Katona C, Bertolote J, Camus V, Copeland J, et al. Reducing stigma and discrimination against older people with mental disorders: a technical consensus statement. Int J Geriatr Psychiatry. 2003;18:670—8. http://dx.doi.org/10.1002/gps.876

11. Won Y. [Experiences of elderly discriminations and psychological well-being of the elderly]. Soc Welf Policy. 2005;21:319—39. Korean.

12. Brown SC, Mason CA, Spokane AR, Cruza-Guet MC, Lopez B, Szapocznik J. The relationship of neighborhood climate to perceived social support and mental health in older Hispanic immigrants in Miami, Florida. Aging Health. 2009;21: 431—59. http://dx.doi.org/10.1177/0898264308328976

13. Palmore E. The ageism survey: first findings. Gerontologist. 2001;41(5):572—5. http://dx.doi.org/10.1093/geront/4L5.572

14. McGuire S, Klein D, Chen S. Ageism revisited: a study measuring ageism in east Tennessee. USA Nurs Health Sci. 2008;10:11—6. http://dx.doi.org/10.1111/j.1442-2018.2007.00336.x

15. Ezzati A, Jiang J, Katz M, Sliwinski M, Zimmerman M, Lipton RB. Validation of the perceived stress scale in a community sample of older adults. Int J Geriatr Psychiatry. 2014;29:645—52. http://dx.doi.org/10.1002/gps.4049

16. Barnes L, Mendes de Leon C, Lewis T, Bienias J, Wilson R, Evans D. Perceived discrimination and mortality in a population-based study of older adults. Am J Public Health. 2008;98(7):1241—7. http://dx.doi.org/10.2105/ajph.2007.114397

17. Lee HS, Lim JH. Structural equation modeling with AMOS 18.0/19.0. Seoul (Korea): Jyphyunjae Publishing; 2011. p. 18—37.

18. Zimet G, Dahlem N, Zimet S, Farley G. The multidimensional scale of perceived social support. J Pers Assess. 1988;52(1):30—41. http://dx.doi.org/10.1207/s15327752jpa5201_2

19. Sheikh JL, Yesavage JA. Geriatric depression scale (GDS): recent evidence and development of a shorter version. Clin Gerontol. 1986;5(2):165—72. http://dx.doi.org/10.1300/j018v05n01_09

20. Jang U, Kim G, Chiriboga DA. Acculturation and manifestation of depressive symptoms among Korean American older adults. Aging Ment Health. 2005;9(6):500—7. http://dx.doi.org/10.1080/13607860500193021

21. Lee HS, Kim CM. [Effects of oral health impact profile (OHIP) on depression and quality of life among community-dwelling Korean elderly persons]. J Korean Acad Community Health Nurs. 2012;23(3):338—46. http://dx.doi.org/10.12799/jkachn.2012.23.3338. Korean.

22. Wood G, Wilkinson A, Harcourt M. Age discrimination and working life: perspectives and contestations—a review of the contemporary literature. Int J Manag Rev. 2008;10(4):425—42. http://dx.doi.org/10.1111/j.1468-2370.2008.00236.x

23. Youn H, Kim M. [The stress, coping behavior and well-being of older adults]. Hanguk Nonyonhak. 1994;14(1):17—32. Korean.

24. Utz R, Swenson K, Caserta M, Lund D, deVries B. Feeling lonely versus being alone: loneliness and social support among recently bereaved persons. J Gerontol B Psychol Sci Soc Sci. 2013;69(1):85—94. http://dx.doi.org/10.1093/geronb/gbt075

25. Kaplan HB, Robins C, Martin SS. Antecedents of psychological distress of social support, and life events. J Health Soc Behav. 1983;1:93—115. http://dx.doi.org/10.2307/2136573

26. Kim J, Stewart R, Shin I, Yoon J, Lee H. Lifetime urban/rural residence, social support and late-life depression in Korea. Int J Geriatr Psychiatry. 2004;19: 843—51. http://dx.doi.org/10.1002/gps.1175