Abstract
Chronic illness with its accompanying physical stressors poses a risk factor for loneliness and depression in later life. Testing a model of stress and coping, we examined the effects of three types of coping resources (religious coping; Selection, Optimization, and Compensation [SOC] adaptive strategies; and perceived social support) on the deleterious effects of chronic illness among older women. Community-dwelling older women (N = 138) with at least one chronic illness (M = 3.9, SD = 2.1) completed mailed questionnaires. Respondents reported multiple morbidities and 90% reported interference with daily life. Social support was associated with less loneliness and depression and mediated the relationship between physical health and loneliness. Our study demonstrates two distinct pathways to depressive symptoms: one through physical symptoms, pain, and disability, and another through the experience of loneliness. Findings support intervention approaches that address disability-related issues and loneliness, and assist older women with chronic illness in identifying and marshaling social support.
Introduction
In the United States, nearly 80% of all older adults have one or more chronic illness, while 66% of elders have two or more comorbidities (Centers for Disease Control and Prevention, 2013). Due to longer life expectancies than men, women spend more years living alone, suffering from more debilitating chronic illness than men in later life. Disability and functional impairments often accompany chronic illness, leading to reductions in activity and increased social isolation, known risk factors for loneliness and depression among older adults. Older women living in the community with chronic illness are therefore at increased risk for loneliness and depressive symptoms. Despite this, little research has focused specifically on the mental health outcomes for these older women with chronic illness and the roles of key coping resources that may mediate this relationship between chronic illness-related physical stressors and mental health.
The purpose of this study is to describe a sample of older women with chronic illnesses and to examine the effects of sociocultural factors, measures of physical health and functioning, and several potential coping resources (religious, adaptive, and supportive) on their mental health. A stress and coping model (Folkman, 2010; Lazarus & Folkman, 1984; Pearlin, Mullan, Semple, & Skaff, 1990; Thoits, 1986), with its focus on the stressor–outcome relationship, is commonly used to conceptualize the effects of primary stressors upon mental health outcomes and informs a conceptual model of the interplay of key resources, adaptation strategies, and any mediating effects on loneliness and depressive symptoms for chronically ill older women. In this model, chronic illnesses and physical impairments comprise the primary physical health stressors. We specifically examine the mediating effect of three types of coping resources—religious coping; the adaptive coping strategies of Selection, Optimization, and Compensation (SOC); and subjective social support—on the deleterious effects of those primary stressors on two mental health outcomes, loneliness and depressive symptoms.
Literature Review
Chronic Illness, Loneliness, and Depression in Later Life
The association of various definitions of physical health with mental health has been widely demonstrated in the gerontology literature. Particular findings with the strongest support include the link between depressive symptoms and health variables such as one or more chronic illnesses (Huang, Dong, Lu, Yue, & Liu, 2010), health-related functional impairments, or disability (Blazer, 2003; Cole & Dendukuri, 2003; Gallo, Rebok, Tennstedt, Wadley, & Horgas, 2003; Jones, Amtmann, & Gell, 2015). A number of studies have concluded that poor health, disability, and pain contribute to depression among older adults (Cole & Dendukuri, 2003; Cummings, Neff, & Husaini, 2003; Parmelee, Katz, & Lawton, 1991; Schoevers et al., 2000). Other studies have concluded that elders’ experience of depression (including subthreshold, minor, or major depression) is associated with greater functional impairment or disability (Fiske, Gatz, & Pedersen, 2003; Hybels, Blazer, & Pieper, 2001; Seidlitz, Lyness, Conwell, Duberstein, & Cox, 2001). Researchers using longitudinal data in the Health and Retirement Study examined the direction of influence between self-rated health and depressive symptoms and found that physical health problems were more likely to affect depression than vice versa (Kosloski, Stull, Kercher, & Van Dussen, 2005).
Poor physical health in older adults is associated with loneliness (Luanaigh & Lawlor, 2008; Pennix et al., 1999). Common age-related losses and transitions, such as changes in health that necessitate changes in routine, may also promote loneliness (Alpass & Neville, 2003; Cohen, 2000). In a review of 149 studies examining loneliness in older persons, women, those above age 80, and those with lower incomes were more likely to be lonely (Pinquart & Sorensen, 2001), suggesting that older women may be uniquely vulnerable to loneliness.
In turn, loneliness may increase vulnerability to depressive symptoms and disorders among older adults (Adams, Sanders, & Auth, 2004; Blazer, 2002; Cohen-Mansfield & Parpura-Gill, 2007; Luanaigh & Lawlor, 2008). Qualitative evidence suggests that the experience of loneliness is actually very closely related to the depressive experience for many older adults (Barg et al., 2006). Recent longitudinal studies have shown different relationships between loneliness and depression. Heikkinen and Kauppinen (2004) concluded that loneliness preceded the onset of depression, while Cacioppo, Hughes, Waite, Hawkley, and Thisted (2006) demonstrated reciprocal effects between loneliness and depression over time.
Adaptive Resources
Although the above discussion of the influence of physical health stressors on older women’s mental health is well established in the literature, our conceptual model simultaneously examines the effects of three selected coping resources in the pathway from the stress of chronic illness to these undesirable mental health outcomes. Gerontological literature suggests the potential importance of coping resources, including (a) religious coping (e.g., drawing upon spiritual beliefs and practices), (b) adaptive coping to functional limitations, which we operationalize as SOC behavioral adaptations, and (c) the perception of available social support.
Religious Coping
The role of religion and spirituality in the lives of older adults and the influence of “religiosity” on health has become an increasing focus of gerontological research (Koenig, King, & Carson, 2012; Koenig, McCullough, & Larson, 2001). The importance of religion and spirituality as a context for understanding and managing health is receiving more attention in health sciences research (Harvey & Silverman, 2007; Koenig et al., 2001; Thune-Boyle, Stygall, Keshtgar, & Newman, 2006). Older adults often report using religion as a coping strategy for challenges such as chronic illness (Emery & Pargament, 2004). Health research about religion/spirituality suggests that many people rely on religion, spirituality, and faith to cope with illness, and that use of these coping techniques is associated with less depression (see Koenig et al., 2001, for a review). A study by Han and Richardson (2010) identified spirituality as a coping strategy to mitigate depression and loneliness in their sample of homebound elders.
The interconnections between religiosity and health may be especially crucial for older adults with chronic illness (Ardelt & Koenig, 2006). Qualitative studies have described how vulnerable groups of older adults, including nursing home residents (Choi, Ransom, & Wyllie, 2008), older immigrants (Lee & Chan, 2009), and homeless elders residing in extended-stay hotels (Lewinson, Hurt, & Hughes, 2015), report using their religion to cope with depression and other negative emotions.
Due to the fact that religion is embedded within cultural and family values, attitudes toward using religiosity as a coping mechanism to reduce loneliness and depression may differ by gender and ethnicity. Religion and spirituality may be more important to older women than to older men (McInnis-Dittrich, 2009), as women are more likely to use religious coping as a resource in dealing with stressful situations than men (Chatters, Taylor, Jackson, & Lincoln, 2008). Variations by membership within diverse ethnic groups suggest that African Americans and Latinas, compared with White women, are more likely to use religious coping (Haley et al., 1996; Lee, Czaja, & Schulz, 2010; McIlvane, 2007). However, less research has focused on how older women’s religious or spiritual coping with their chronic illness and physical impairment impact their mental health.
Researchers on religious coping have identified two coping functions of religion that they label positive and negative religious coping (Pargament & Abu Raiya, 2007; Pargament, Feuille, & Burdzy, 2011; Pargament, Smith, Koenig, & Perez, 1998). As such, the use of positive religious coping suggests “a secure relationship with a transcendent force, a sense of spiritual connectedness with others, and a benevolent world view,” whereas negative religious coping “reflects underlying spiritual tensions and struggles with oneself, with others, and with the divine,” (Pargament et al., 2011, p. 51). Positive religious coping involves working with God or a divine being as a partner, with the expectation that God will help to provide comfort, guidance, and strength to cope with stressful situations. Beliefs associated with negative religious coping may be more frightening or shaming, therefore exacerbating stressors through beliefs of abandonment by God or the church, or punishment for sinful behaviors. Therefore, positive religious coping strategies such as seeking a stronger connection with God may result in alleviating stress related to declining physical health, while negative religious coping may increase stress. The distinctive functions of positive and negative religious coping strategies and mental health of older women with chronic illnesses are less known. In our study, we expected to see an association between higher levels of positive religious coping and lower levels of negative religious coping with less loneliness and depressive symptoms, suggesting a mediating role for religious coping between the challenges from the health stressors facing these older women with chronic illness and their mental health.
SOC
SOC (Baltes & Baltes, 1990; Freund & Baltes, 1998) is a set of related adaptations to cope with age-related or health-related changes in functioning from a successful aging perspective. SOC is based on the proposition that reserve capacities generally decline with aging and to cope with these losses, older adults may use the adaptive coping strategies of selection, optimization, and compensation. In the selection component of the model, the older person selects tasks or activities as priorities and drops others from his or her repertoire. Optimization is a concerted effort, resource allocation, or refinement of skills to make the best use of one’s remaining strengths, skills, and energies to reach goals. Compensation refers to finding and using alternative means to reach a given goal when skills or functioning are affected by diminished capacities. Compensation includes use of assistive devices and the help of other people (Jopp & Smith, 2006).
In early studies, higher rates of all three modes of adaptation comprising the SOC model were associated with indicators of successful aging (less loneliness, greater positive affect, and subjective well-being) among older adults, even after controlling for other demographic and health characteristics (Freund & Baltes, 1998, 2002). Some researchers consider SOC primarily a theory of resource use. When resources are scarce (e.g., through disability or functional impairment), the theory suggests that SOC strategies are more important in managing remaining resources (Freund, 2008), such as a study in which positive effects of SOC adaptation and coping were especially strong for participants with fewer personal and social resources (Jopp & Smith, 2006). Among older women with chronic illness, this literature informed the current study where we anticipated that greater use of SOC strategies in everyday tasks to cope with physical health stressors would result in reduced loneliness and depressive symptoms.
Social Support
Perceived or subjective social support is considered one of the main resources in the stress and coping model as it has been applied to aging and family caregiving (Pearlin et al., 1990). Social support is an important negative correlate of loneliness, although does not fully explain it (Dykstra & de Jong Gierveld, 2004). As a resource for coping with physical health stressors, social support has received much attention in the research literature (Gallant, 2003), appearing to serve a protective function with respect to depression or anxiety conditions among individuals dealing with chronic disease (Ferreira & Sherman, 2007; Keyes et al., 2005; Thomas, Jones, Scarinci, & Brantley, 2007). Thus, in this study, we anticipated that greater endorsement of subjective social support would be associated with less loneliness and fewer depressive symptoms.
The Current Study
The above literature suggests the conceptual model for the current study (see Figure 1), whereby two physical health stressors (the number of chronic illnesses and a composite score of self-assessed general health, pain, and physical impairment) may increase vulnerability to both loneliness and depressive symptoms. Further, as described above, our model incorporates the importance of three types of coping resources (i.e., religious coping, SOC adaptive strategies, and social support) as mediators that may decrease two mental health outcomes of loneliness and depressive symptoms among older women with chronic illness. In the model, loneliness and depression are theorized to be interrelated in such a way that loneliness may be an intermediate outcome and also a predictor for depression, so we also anticipate that loneliness will mediate the relationship between coping resources and depressive symptoms. The hypothesized relationships among the variables, based on the theoretical and empirical evidence summarized above, are represented in the conceptual model (see Figure 1). In this cross-sectional model, we test three specific related hypotheses, as follows:
Conceptual model.
Method
Sample
This study uses data from the first wave of a longitudinal study on the health of older women with chronic illnesses. Women were eligible to participate in the study if they were 65 years of age or older, had one or more medically diagnosed chronic conditions, and lived in the community. Potential participants were recruited by convenience sampling strategies in a Midwestern state. Over 15 agencies serving older adults, including senior transportation services, senior housing, social services, and free medical clinics, helped with sample recruitment. Fliers were posted in public areas such as libraries, bus stops, and so on. An article about the study was also released in the local paper and appeared in an online news release and Blog. A total of 173 potential participants contacted our office to express interest in participating; seven were deemed ineligible (younger than 65, male gender, lived in assisted living facilities, and/or cognitively impaired), 10 did not respond to a follow-up screening phone call, and five refused. A total of 151 women were deemed eligible and agreed to participate; however, 12 did not return questionnaires and one was removed from analysis because the respondent was 55 years old, leaving a final N of 138. Characteristics of the sample are shown in Table 1. This study received Institutional Review Board approval (IRB Protocol # 20090641).
Sample Characteristics (N = 138).
Note. Age range = 65 to 88 years, M = 72.7, SD = 6.2.
Measures
Sociocultural factors
Demographic variables including age in years, racial identification (0 = White, 1 = people of color), marital status (0 = single, 1 = married/partnered), and income adequacy (0 = inadequate or barely adequate, 1 = adequate) were examined to determine their effects on the subsequent variables in the study model. Religious affiliation and educational attainment were used as descriptors only.
Independent variables
Physical health stressors
Two measures were used to represent the physical health stressors associated with having a chronic illness.
Number and severity of chronic illnesses
Physical comorbidity was measured with the Charlson Comorbidity Score, a weighted index of medical comorbidities that considers the number and severity of medical conditions (Charlson, Pompei, Ales, & MacKenzie, 1987). We used this measure as a summative index.
Physical health
Health, pain, and social impairments in functioning were measured with the Physical Component of the SF-12v2 (Ware, Kosinski, & Keller, 1996). The physical component of the SF-12v2 instrument is a composite measure of the effects of health and pain on functioning, and also a self-rating of health. This summary score of the physical component of the SF includes the following subscales and corresponding items: Physical Functioning (two items), Role Physical (two items), General Health Perception (one item), Vitality (one item), Pain (one item), and Social Functioning (one item). Response options include three ratings of degree of limitation, and five ratings of frequency of limitations. The measure has well-established norms for many age groups and both genders (Ware et al., 1996). Higher scores on this scale indicate worse physical health stressors.
Coping
Religious coping
Religious coping was measured by the Brief R-COPE (Pargament et al., 1998) that consists of two subscales with seven items each that focus on two types of religious coping: positive and negative. Previous research has demonstrated that each subscale has good discriminant validity and demonstrated high internal consistency across diverse samples, with the positive religious coping median alpha value of .92 and the negative religious coping median alpha value of .81 (Pargament et al., 2011). Participants were asked to think about how they try to understand and deal with major problems in their life and to what extent was each item involved in the way they coped. Item responses ranged from 0 (not at all) to 3 (a great deal). Items in the positive subscale included “sought God’s love and care,” “tried to put my plans into action together with God,” and “focused on religion to stop worrying about my problems.” The positive subscale in this study had an alpha reliability of .86. Some items in the negative subscale included “questioned God’s love for me,” “wondered what I did for God to punish me,” and “questioned the power of God.” The Cronbach’s alpha was .81 in the sample.
SOC adaptive strategies
We used a brief (12 item) measure of SOC strategies adapted from Ziegelmann and Lippke (2007). Those authors in turn derived their measure from the original SOC Questionnaire, a set of 48 items developed by the theory’s originators (Freund & Baltes, 2002). Both the original longer measure and the shorter version consist of four subscales: Elective Selection (ES), Loss-Based Selection (LBS), Optimization (O), and Compensation (C), but published findings report only the total scale scores.
The SOC items used in this study focus on adaptations made in undertaking daily tasks and activities, through selecting from among activities (selection), trying harder with activities (optimization), or finding another means to accomplish something (compensation). They included such statements as the following: “In order to keep going, I have looked for new methods to do things,” “When it became harder to do the same things, I have increased my efforts to accomplish them,” and “When I could no longer do something in my usual way, I’d carefully consider how I might do this under current conditions.” Items were scored with five response choices ranging from strongly disagree to neither agree nor disagree, to strongly agree. Exploratory factor analysis (EFA) with the four theorized subscale scores did not find sufficient evidence to use the subscales, so like earlier authors, we report only the total SOC score. The 12-items SOC scale obtained an internal consistency of .89 in the current study.
Social support
Subjective social support, measuring the perceptions of social support available if needed, was assessed using a subscale of the Duke Social Support Index (Hughes, Blazer, & Hybels, 1990), a well-established measure used in the Epidemiological Catchment Area surveys. The nine-item subscale for subjective social support (rated 1, hardly ever, to 3, most of the time) includes questions such as, “In times of trouble, how often can you count on at least some of your family and friends?” “Do you feel useful to your family and friends?” and “Do you feel you have a definite role in your family and among friends?” The Cronbach’s alpha for subjective social support was .90 in this study.
Dependent variables
Depressive symptoms were measured with the Center for Epidemiological Studies–Depression (CES-D) scale (Radloff, 1977). The 20 scale items are scored 0 (rarely) to 3 (most or all of the time) and summed, with higher scores indicating more depressed mood. While the continuous variable was used as one of the two mental health outcomes in this study, this scale can be used to identify those at risk for developing clinically significant depression (e.g., a score of 16 or higher). Additionally, Blazer, Hybels, and Pieper (2001) have defined subthreshold depression as a range of subcriterion scores on the CES-D. The CES-D is widely used in clinical and research settings with demonstrated validity and reliability in samples of community-dwelling elders. Questions included, “Did you feel depressed?” “Did you have trouble keeping your mind on what you are doing?” and “Did you feel fearful?” Four out of the 20-items are positively worded and were reverse-scored before summation. In our sample, the scale ranged from 0 to 54 with a mean total score of 13.8 (SD = 10.4). The Cronbach’s alpha in this study was .88.
Loneliness was measured by the 11-item de Jong-Gierveld Loneliness Scale (de Jong-Gierveld, 1987). There are two components of the scale, designed to measure the absence of a larger social network (social loneliness) and the absence of an intimate relationship, such as a partner or best friend (emotional loneliness), but are meant to be considered together (de Jong-Gierveld, 1987). Participants were asked to indicate their agreement using yes, no, or more or less responses to statements such as “There is always someone I can talk to about my day-to-day problems,” “I miss having people around,” and “I find my circle of friends and acquaintances too limited.” The scale includes both positively and negatively worded items, has been used in a number of studies, and has good psychometric properties. In our study, the scale had a range of 0 to 12 with a mean score of 5.18 (SD = 4.01) and the internal consistency was .90.
Analysis
Using IBM (2014, SPSS 23.0), we examined descriptive statistics and the zero-order correlations of all the variables in the model. Tests for the primary underlying assumptions of regression and structural equation modeling (SEM) were conducted; assumptions of linearity, reliability of measurement, homoscedasticity, and normality were met. Factor analyses were run to examine the factor structure of scales and similar constructs (e.g., social support and loneliness, chronic conditions and physical health). Power analysis was conducted using methods from MacCallum, Browne, and Sugawara (1996), MacCallum and Hong (1997), MacCallum, Browne, and Cai (2006). The minimum sample size required for the study was 127 participants, and our study surpassed this number and recruited 138 participants.
We used SEM with AMOS 23.0 (Arbuckle, 2015) to test the hypothesized model of the associations between coping resources (religious coping, SOC adaptive strategies, and subjective support) and loneliness on depression, after controlling for sociodemographic characteristics (race, age, marital status, and income) and health stressors (number of chronic illnesses and SF-12 physical health). Model fit was assessed with the comparative fit index (CFI), the Tucker–Lewis index (TLI), the root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR; Hu & Bentler, 1999). CFI and TLI values of .95 or greater and RMSEA and SRMR values of .08 are considered to indicate good fitting models (Hu & Bentler, 1999).
Findings
Descriptive results of major variables in the study appear in Table 2. The number of chronic illnesses reported by the participants ranged from 1 to 10 with a mean of 3.9 illnesses (SD = 2.1). Seventy-five percent of the sample reported three or more chronic illnesses. The most common chronic illnesses were hypertension, arthritis, heart disease, diabetes, and digestive disorders. Of the sample, 90.6% of the women indicated that their chronic illnesses caused interference with daily living. Over one half of the sample endorsed some loneliness on the de Jong-Gierveld Loneliness Scale, while 37.7% of the women were at-risk for depression (CES-D ≥ 16), as measured by the CES-D. We contacted women whose scores were 16 or higher to notify them of their elevated scores and to recommend that they consult with their primary care physician or social worker. Additional resources such as the phone number for United Way’s First Call for Help were provided.
Descriptive Information on Continuous Predictors and Depression (N = 138).
Note. Adaptive strategies (SOC) was centered at 0 = “neither agree nor disagree” so that negative scores (−2 or −1) indicate disagreement and positive scores (1, 2) indicate agreement. CES-D = Center for Epidemiological Studies–Depression; SOC = Selection, Optimization, and Compensation; SF-12=12 item Short Form Health Survey.
Several notable bivariate relationships help to describe our unique sample and variables of interest (see Tables 2 and 3). Women of color with inadequate incomes had higher numbers of chronic illnesses, yet overall, women with better incomes reported worse physical health in this sample. With regard to depression, the women of color reported more depressive symptoms, while the unmarried women reported more loneliness. We found a positive relationship between the SOC strategies and the positive religious coping variables; women who used more SOC strategies also tended to use more positive religious coping. Additionally, there was a negative relationship between subjective social support and negative religious coping. Those women with higher reported subjective social support relied less on negative religious coping.
Pearson Correlations of Bivariate Relationships Between Sociocultural Variables, Health Stressors, Chronic Illnesses, Coping, and Mental Health Outcomes (N = 138).
Note. CES-D = Center for Epidemiological Studies–Depression; SOC = Selection, Optimization, and Compensation.
SEM
To examine our theoretical model, we began with an initial fully saturated model that included all paths. After eliminating nonsignificant paths, we achieved an excellent-fitting SEM for our conceptual model (χ2 = 37.86, df = 44, p = .731; CFI = 1.0, TLI = 1.0, RMSEA = .00). Figure 2 illustrates our final empirical model of the path analysis and Table 4 shows the detailed results of our structural model.
The Structural Equation Model of significant associations between sociocultural demographic characteristics, physical health stressors, coping resources, and loneliness on depressive symptoms (N = 138).
Maximum Likelihood Estimates for a Recursive Path Model of Sociocultural Demographic Characteristics, Physical Health Stressors, Coping Resources, and Loneliness on Depressive Symptoms (N = 138).
Note. For clarity, only statistically significant associations appear in the table. Bootstrapping procedures were used to test the potential mediating effect of the coping resources that were significantly related to loneliness and depression. Results are presented in text. R-COPE = religious coping; CES-D = Center for Epidemiological Studies–Depression; SOC = Selection, Optimization, and Compensation.
p < .05. **p < .01. ***p < .001.
Of our hypotheses, H1 was partially supported. Significant pathways in the model show that the two physical health stressors were both directly and positively associated with depressive symptoms, but not with loneliness. Older women with more physical health impairment and a higher number of chronic health conditions reported more depressive symptoms.
Findings supported H2, confirming that loneliness was positively associated with depressive symptoms. As anticipated, the degree of loneliness had a positive, significant relationship with depressive symptoms, independent of the demographic and health variables within the model.
In the final hypothesis (H3), we anticipated that the selected coping resources of religious coping, adaptive coping (SOC), and social support would mediate the relationship between the two primary health-related stressors and both loneliness and depression. For depressive symptoms, contrary to expectation, we found no significant direct relationships between either religious coping or adaptive coping (SOC) and the depression outcome. For loneliness, positive religious coping was directly related to loneliness, but was associated with more loneliness, rather than less, and there was no relationship between adaptive coping (SOC) and loneliness. Greater subjective social support was associated with both less loneliness and fewer depressive symptoms.
Subjective social support showed preliminary promise as a mediator between physical health and loneliness because of the strong association between the social support and loneliness variables and the lack of direct association between physical health impairment and loneliness. To examine whether social support was a mediator between physical health and the loneliness outcome, we used the bootstrapping procedures in AMOS 23.0, with 5,000 bootstrap samples and bias-corrected confidence estimates (Hayes, 2009) for simultaneous testing (Arbuckle, 2015) to obtain the adjusted standard errors for indirect effects. Without subjective social support included in the model, physical health had a significant direct effect on loneliness (−.35***). Adding in subjective social support, and including both direct and indirect paths, the direct effect from physical health was reduced to .10 and was no longer significant. A two-sided bias-corrected bootstrap confidence interval ([−.336, −.146]) for the standardized indirect (mediated) effect of physical health on loneliness was significantly different from zero at the .001 level (p = .001 two-tailed). A value of .047 was the bootstrap estimate of the standard error of the standardized indirect (mediated) effect. From these analyses, we conclude that subjective social support is a particularly strong mediator of the relationship between physical health and loneliness. Thus, H3 was partially supported.
Discussion
This study described a sample of community-dwelling older women living with one or more chronic illness, and used SEM to test a cross-sectional conceptual model positing relationships among two physical health stressors (number of chronic conditions and degree of physical impairment), three coping strategies (positive and negative religious coping, SOC, and social support), and two outcomes commonly associated with aging and chronic illness: loneliness and depressive symptoms.
The study’s first hypothesis posited that the health stressors would be positively associated with both loneliness and depressive symptoms. Here, the analyses found positive significant direct pathways from both of the physical health stressors to greater depression, but not to loneliness. Other differences in the experience of these two outcomes within our sample included that women of color reported more depressive symptoms, whereas unmarried women reported more loneliness.
H2 in the study was confirmed: There was a significant, direct pathway of association between loneliness and depression. This relationship has been noted in previous studies suggesting that loneliness is strongly associated with depression and may also be a precursor to depression among older adults (Adams & Moon, 2009; Adams et al., 2004; Barg et al., 2006; Blazer, 2002; Cohen-Mansfield & Parpura-Gill, 2007).
Thus, a clear finding of our final SEM model is the demonstration of two distinct pathways to depressive symptoms for older women with chronic illness: one through physical health stressors and another through the experience of loneliness. Both physical disability (Fiske, Wetherell, & Gatz, 2009) and loneliness (Ong, Uchino, & Wethington, 2015) are putative risk factors for late life depression, but this study suggests these risks may operate somewhat independently. Physical health stressors were directly linked to depressive symptoms, but not to loneliness in this sample of older women, but loneliness, in turn, was also linked to depression. Our findings echo Rillamas-Sun and colleagues’ (2016) work on the Women’s Health Initiative Study where they examined functional and social determinants of depression and concluded that greater depressive symptoms were associated with poorer physical functioning in community-dwelling women with multiple morbidities.
In the study’s third hypothesis, we anticipated three coping strategies would be significantly related to the participants’ reported loneliness and depressive symptoms and that the coping strategies would mediate the relationship between the two physical health stressors and those mental health outcomes. As noted, our findings partially supported this hypothesis. Of the coping strategies, only greater subjective social support, not religious coping or SOC, was significantly related to less reported loneliness and fewer depressive symptoms. Social support also mediated the effects of physical health impairment on loneliness, suggesting that greater social support is the mechanism through which loneliness is reduced.
In terms of the religious coping strategy, an interesting finding was the statistically significant relationship between non-White racial identification, positive religious coping, greater loneliness, and more depressive symptoms. This finding suggests that the make-up of this specific convenience sample of older women with chronic illness makes racial comparisons difficult, as the women of color tend to have lower socioeconomic status, greater levels of depressive symptoms, and more endorsement of positive religious coping. Although not as common in the literature, this finding of an inverse relationship between positive religious coping and depressive symptoms has been reported in other studies (Pargament et al., 2011).
There were some anecdotal comments by participants that the R-COPE questions, written from a Judeo-Christian perspective, did not represent their beliefs—they deemed themselves to be “spiritual,” not “religious,” or they did not believe in a “Higher Power.” This finding is consistent with Tarakeshwar et al.’s (2006) study that utilized the negative R-COPE measure to examine religious coping and quality of life among cancer patients. Additionally, the Brief R-COPE may work best in samples with individuals who have recently experienced traumatic events. Women living with longstanding illnesses may have accepted and/or managed their illness and corresponding functional limitations and differ from those who have a sudden or unexpected health event. Our findings should be considered within the context of an ever-increasing aging society that will consist largely of older women managing multiple chronic illnesses.
Greater adherence to the SOC strategies was associated with higher levels of the physical health stressors, as it was in a similar study of community-dwelling older adults with multimorbidity in which greater disability led to greater use of SOC strategies (Yuen & Vogtle, 2016). The SOC strategies were unrelated to the women’s loneliness or depressive symptoms in this study. In a longitudinal study by Lang, Rieckmann, and Baltes’ (2002), participants who were resource-rich were more likely to use SOC adaptive strategies than resource-poor participants. In our study, there was a range of “resources.” Three fourths of the women were not married or partnered, over half of the sample deemed their incomes inadequate of barely adequate, and 84% were not employed. In general, our sample was not extremely lonely and had good cognitive functioning yet the women were managing chronic illness.
Although SOC (Baltes & Baltes, 1990) is one of several dominant theories that define and operationalize the concept of “successful aging,” additional theoretical work in this area is recommended to more fully incorporate concepts of disability (Martin et al., 2015). Furthermore, measurement of the SOC construct is not yet standardized which may be an issue in the results of our study as well as the study by Donnellan, Hevey, Hickey, and O’Neill (2012). We found that there was little variance for many of the items in our measure of SOC among our respondents who had scores with higher levels of endorsement of these behaviors. Perhaps, the SOC type of adjustments measured are simply the routine actualization of the strengths and resources of these older women—a natural and perhaps necessary adaptation to aging for women with chronic illness.
Donnellan and colleagues (2012) also attempted to measure SOC behaviors among stroke patients and obtained similar results. In that study, SOC was seen among the stroke victims, but overall was not significantly associated with positive outcomes. Thus, as we have been able to measure SOC, our measures of health stressors were associated with more SOC type adaptations, but the adaptations alone were not associated with depressive symptoms. Further, the purported subscales that were to capture specific elements of SOC (selection, optimization, and compensation) were not clearly distinguished in this study. The women in our sample may use some of these strategies, but not all. Other researchers have examined the subscales of SOC and these are applied to specific contexts (like working) or losses (e.g., driving cessation). Zacher, Chan, Bakker, and Demerouti (2015) found that “selection” was not related to their outcome of work engagement; however, “optimization” and “compensation” were significantly related.
Finally, of the coping resources examined, study results highlight the importance of subjective social support in the experience of loneliness—and by virtue of the effects of loneliness on depressive symptoms, also on depression among older women with chronic illness. Greater subjective social support had a strong negative association with both mental health outcomes. This study’s conceptual model suggests that depression may have its origins in emotional or social loneliness, or in the symptoms and impairments of chronic illness. In either case, although cross-sectional, this study demonstrates that greater social support is associated with fewer symptoms of loneliness or depressive symptoms and was a mediator between the physical health variable and loneliness, specifically.
As Blazer (2005) noted, social support stands out in aging research on depression and its interventions:
Social support is a most important factor in preventing both the onset and progression of depression in later life. Recent advances in our understanding of the biological underpinnings of these depressive states must not blind us to the importance of social factors. (Blazer, 2005, p. 499)
Our findings are similar to Tanner, Martinez, and Harris (2014) who examined depressive symptoms, loneliness, functioning, and social support among community-dwelling homebound older adults with high levels of comorbidity and found that satisfaction with family support, functional limitations, and loneliness were useful predictors of depression. The current study’s findings offer potential implications for future research and treatment involving social support for older adults whose physical health renders them vulnerable to social isolation, loneliness, and depression.
Study Strengths and Limitations
This study makes a contribution to the literature on the health and mental health of older women with chronic illness. The most innovative aspect of this study is the examination of a conceptual model including two very different types coping resources—religious coping and SOC—along with a well-established coping resource, social support, with the same sample of older women with chronic illnesses. Additionally, the inclusion of loneliness and its relationship to the physical and mental health of older women with chronic illness warrants attention. Hawkley and Cacioppo (2010) maintain the necessity of examining the consequences and mechanisms of loneliness in research to effectively inform interventions to reduce loneliness and improve health.
While a number of researchers have reported positive effects of social support and religious/spiritual coping for older adults dealing with chronic illness, few studies have examined the lesser known SOC adaptive strategies. A strength of the study is the use of SEM with adequate power to test pathways from health-related stressors to mental health outcomes simultaneously. Findings provide additional insights into the interplay of loneliness and depressive symptoms for women who are living with physical disease and disability.
Another unique contribution of this study is the descriptive data that provide a snapshot of community-dwelling older women coping with chronic illness (see Tables 1 and 2). The high comorbidity of chronic illness was remarkable among these women, who reported an average of close to four chronic conditions each. Our sample may be compared with a national study of the multimorbidity of Medicare beneficiaries that described that over two thirds of the sample had two or more co-occurring chronic conditions (Salive, 2013). The increased prevalence of multimorbidity in older age was associated with higher risk of negative outcomes, including disability (Salive, 2013). A high rate of depressive symptoms, affecting more than one third of the sample, was also noteworthy, compared with estimates of 15% for community-dwelling older adults with clinically significant depressive symptoms (Blazer, 2003).
A few important limitations of this study must be noted. This was a pilot study conducted with convenience sampling in a limited geographic area. Although a factor analysis indicated distinct constructs from the measures of social support and loneliness, the reader should be aware of potential confounding effects with conceptually related variables. For instance, low-quality social interactions may be responsible for lower social support, which contributes to more loneliness. Analysis of cross-sectional data provides information about associations at one point in time among the variables in our conceptual model. We can determine directional influence but cannot evaluate autoregressive effects or time lags with this design; therefore, longitudinal analyses with SEM could strengthen the ability to make causal inferences.
Another limitation is the limited variability in responses to the measure designed to capture SOC adaptations; a majority of the women in the sample “agreed” with most of the items. The sample may simply have been quite homogeneous with regard to these behaviors. Furthermore, our abbreviated measure of SOC, adapted to reduce the survey length and respondent burden, may have been inadequate to tap into this construct comprehensively.
Implications for Future Research and Practice
This cross-sectional study revealed significant paths of our proposed model for prediction of depressive symptoms among older women with chronic illness. Longitudinal analyses on future waves with this and other sample populations will be needed to confirm these findings. Furthermore, future studies should examine the likely reciprocal effects between loneliness and depressive symptoms over time.
At the most fundamental level, these study findings identify the high degree of depressive symptoms among older women living with chronic illnesses and the major beneficial role of subjective social support. Practitioners should concentrate their efforts on identification and recognition of the frequent occurrence of mental health problems such as loneliness and elevated depressive symptoms among older women with chronic illness, particularly those who may be more socially disengaged or disconnected due to their physical limitations.
Older women in the community coping with chronic conditions might benefit from intervention approaches that address disability-related issues and loneliness, and assist them to identify, maintain, and strengthen social and emotional support. Associations found among low-income/non-White race, physical health stressors, loneliness, and depression point to several possible areas for psychosocial intervention development. Future interventions targeting women with elevated loneliness and low perceived social support might include group support, telephone support, identifying sources of support among friends, family and neighbors, increasing women’s social skills and assertiveness, and cognitive restructuring of their appraisals about current available sources of support. Further, intervention approaches to address depressive symptoms should incorporate consideration of common disability and pain-related issues affecting the daily lives of these women and strategies to help women with very real physical limitations remain socially engaged.
Footnotes
Acknowledgements
The authors thank Christopher J. Burant and Carol M. Musil for their insights on earlier drafts of this paper.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors received grant funding by the Case Western Reserve University Center on Aging and Health, President’s Initiative Fund, sponsored by the McGregor Foundation.