The Impact of Social Resources on Depressive Symptoms in Racially and Ethnically Diverse Older Adults

Abstract

The objectives of this study were (1) to develop an empirical typology of physical health risks in racially and ethnically diverse older adults and (2) to examine whether the impact of social resources on depressive symptoms differs across the identified health risk groups (low, moderate, and high risks). The data source was the Survey of Older Floridians, a statewide survey of older adults aged 65 and older (n = 1,432). Latent profile analysis with multiple indicators of physical health (chronic conditions, functional disability, and self-rated health) was used to identify three health risk groups (low, moderate, and high risks). The direct and interactive effects of the health risk group membership and social resources (social support and religious service attendance) on depressive symptoms were found. Of particular interest was that the positive impact of social support was most pronounced in the moderate health risk group.

Keywords

health risks social resources depressive symptoms older adults

Depression is one of the most common mental health problems among older adults (Lebowitz et al., 1997; Wu, Schimmele, & Chappell, 2012). A recent report, for example, suggests that the diagnosis of depression in older community-dwelling adults nearly doubled from 3.2 to 6.3% between 1992 and 2005 (Akincigil et al., 2011). Further, research indicates that approximately 10–30% of community-dwelling older adults are affected by such symptoms of depression as depressive mood and somatic complaints (Bekman et al., 1986; Black, Markides, & Miller, 1998; Blazer, Landerman, Hays, Simonsick, & Saunders, 1998; Gonzalez, Haan, & Hinton, 2001). Late-life depressive symptoms are associated with increased risks for the onset of major depression, physical disabilities, poor quality of life, accelerated mortality, and suicide (Blazer, 2003; Blazer, Hybels, & Pieper, 2001; Lebowitz et al., 1997; Sun et al., 2012).

The prevalence of depressive symptoms and related problems in older adults has prompted extensive research into the etiology of these problems. Social resources such as participation in social activities and having diverse social connectedness have received particular attention. Beneficial effects of social resources include decreased rates of mortality (Agahi & Parker, 2008; Maier & Klumb, 2005), slowing of the disablement process (Buchman et al., 2009; James, Boyle, Buchman, & Bennett, 2011; Mendes de Leon, Glass, & Berkman, 2003), decreased risk of cognitive impairment (Barnes, Mendes de Leon, Wilson, Bienias, & Evans, 2004; Seeman et al., 2011; Zunzunegui, Alvarado, Del Ser, & Otero, 2003), and fewer depressive symptoms (Glass, Mendes de Leon, Bassuk, & Berkman, 2006; Hong, Hasche, & Bowland, 2009; Johnson, 1991). The gerontological literature has generally identified two major categories: emotional support (i.e., support making people feeling cared for and trusted) and instrumental support (i.e., providing tangible support such as assistance with transportation and shopping). Of these, emotional support or perceived support may be most closely related to satisfaction with support and individual well-being (Antonucci, 2001; Carstensen, Fung, & Charles, 2003). Aside from social support, religiosity and religious involvement have been positively associated with individual well-being through a proposed mechanism of promoting good health habits and providing the opportunities to be socially connected (Koenig, McCullough, & Larson, 2001; Park et al., 2008).

In addition to the direct effects of social resources, a substantial body of evidence suggests these resources may exert a buffering effect on mental health (Moak & Agrawal, 2010; Thoits, 2011). Pearlin and Bierman’s (2013) stress process theory posits that the impact of the stressors on mental health varies according to the availability of personal and social resources. Social resources such as social support and religious involvement, for example, have the potential to mitigate the negative effects of life stresses and promote mental health (Dickinson, Potter, Hybels, McQuoid, & Steffens, 2011; Jang, Chiriboga, Kim, & Phillips, 2008; Koenig et al., 2001; Krause, 1987; Krause & Bastida, 2011; Seeman, 2000). Conversely, lower levels of social resources are linked with risks for the detrimental effects of the stressor on depression and other mental health problems (Turner & Turner, 2013).

The adverse effects of physical health constraints such as chronic health conditions, functional disability, and perceived health on mental health have been reported in studies of diverse populations (Bekman et al., 1986; Bruce, 1999; Chiriboga, Black, Aranda, & Markides, 2002; Jang, Haley, Small, & Mortimer, 2002; Mezuk, Edwards, Lohman, Choi, & Lapane, 2012). Although multiple domains of physical health constraints are highly correlated (August & Sorkin, 2010; Salomon, Nordhagen, Oza, & Murray, 2009; Seeman, Merkin, Crimmins, & Karlamangla, 2010), most studies have focused on the individual effects of these variables. This focus on discrete variables may yield inconsistent findings across studies and make it difficult to capture full understanding of one’s health (Fiori & Jager, 2012; Schmiege, Meek, Bryan, & Petersen, 2012).

As an alternative to the focus on discrete variables, the “person-centered” or typological approach, employed in the present study, identifies subgroups of individuals who share a similar profile on key criterion variables; in the present study, these key variables are related to health constraints (Hair & Black, 2000; B. O. Muthén & Muthén, 2000). Empirically based typologies identify “mutually exclusive, easily identified subgroups” (Davis et al., 2000, p. 34) and have proved to be useful tools to identify regularities or patterns in diverse groups (Schmiege et al., 2012). The advantages of using the person-centered approach over variable-centered approaches include the following: (1) interaction or spurious effects among variables can be dealt with at least in part by classifying persons into meaningful subgroups; (2) findings can be generalized to previously unidentified groups of people; and (3) information can be organized according to subjects with shared profiles on key variables, thus yielding a more holistic perspective (Everitt, Landau, & Leese, 2001; Hair & Black, 2000). In other words, instead of looking at how variables relate to each other, the approach examines how people relate to each other. Because empirically derived typologies take commonalities among people into account, they also can provide insights into synergy effects where an individual’s scores on particular variables could have a larger impact than the sum of the effects taken separately. Thus, the person-centered approach does not treat physical health variables as discrete domains but rather identifies groups of individuals who share similar profiles of physical health.

Although depression affects many individuals, some share characteristics that can make them more vulnerable to the condition, such as being members of disadvantaged racial/ethnic minority groups, being women, being unmarried, and having lower socioeconomic status. In particular, racial/ethnic minorities generally experience higher levels of depressive symptoms than non-Hispanic Whites, yet depression as a disorder remains underrecognized and undertreated among older racial/ethnic minorities (Akincigil et al., 2012; Cochran, Brown, & McGregor, 1999; Dunlop, Song, Lyons, Manheim, & Chang, 2003). The racial/ethnic minorities are also more likely to have physical health constraints such as functional disability than their White counterparts (Louie & Ward, 2011; Mezuk et al., 2010). Although findings on social resources across racial/ethnic groups are mixed (Mendes de Leon & Glass, 2004), racial/ethnic minorities tend to have smaller social networks than non-Hispanic Whites, and their networks are likely to be concentrated on family/kin relationships (Barnes, Mendes de Leon, Bienias, & Evans, 2004; Fiori, Antonucci, & Cortina, 2006; McPherson, Smith-Lovin, & Brashears, 2006). On a positive note, the beneficial effects of religious attendance on individual well-being have been widely reported across racial/ethnic groups (Krause & Bastida, 2011; Reyes-Ortiz, Pelaez, Koenig, & Mulligan, 2007); minority elders, Blacks in particular, tend to be highly involved in religious activities and reap benefits from such involvements (Krause, 2002, 2006).

The goals of the present study were two-fold. The first goal was to develop an empirical typology of health risks of older adults using multiple physical health indicators (e.g., chronic conditions, functional disability, and self-rated health). The second goal was to determine whether the impact of social resources (e.g., social support and religious attendance) on depressive symptoms differs across the identified health risk groups. We hypothesized that (1) there would be subgroups of individuals with varying degrees of health risk profiles and (2) the effects of social resources on depressive symptoms would vary across the identified risk groups. It was expected that individuals in higher health risk groups would receive greater benefits from social resources. In regression analyses, we controlled sociodemographic variables (e.g., age, gender, race/ethnicity, marital status, education, and income), which were identified to be related to physical health and depressive symptoms.

Method

Participants

Data for the present study came from the Survey of Older Floridians (SOF; Jang et al., 2008; Zayac, Salmon, & Chiriboga, 2005). Data were collected through computer-assisted telephone interviews between 2004 and 2005. Using a statewide sampling frame complemented by an oversampling of racial/ethnic minorities, the study included a total sample of 1,432 respondents: 503 Whites, 360 Blacks, 328 Cubans, and 241 non-Cuban Hispanics. Respondents were selected through random digit dialing. Eligibility criteria included being age 65 or older and free from cognitive impairment. Individuals with more than five errors on the Short Portable Mental Status Questionnaire (Pfeiffer, 1975) were excluded. The response rate for the overall statewide sample was 62%. More information on the SOF is available elsewhere (Jang et al., 2008; Park, Jang, Lee, & Chiriboga, 2013).

Measures

Dependent Variable

A short form of the Center for Epidemiologic Studies Depression scale (CES-D; Andresen, Malmgren, Carter, & Patrick, 1994; Radloff, 1977) was used to assess depressive symptoms. The instrument contains eight negatively worded items and two positively worded items (reverse coded). The items ask how often symptoms, such as loneliness, feelings of fearfulness, and restless sleep, were experienced during the past week. Responses were coded on a 4-point scale (0 = rarely; 3 = most of the time). The summative score of the 10 items, with a range of 0–30, was used with higher scores suggesting greater depressive mood. A score of 10 or higher was used for the cutoff score for depressive symptoms (Andresen et al., 1994). In the present analysis, the scores were used in a continuous format. Internal consistency for the sample was α = .80.

Physical Health Indicators

Chronic health conditions were measured with a checklist asking participants whether they had ever experienced or been diagnosed with nine specific conditions or diseases: heart attack, stroke, high blood pressure, Parkinson’s disease, respiratory problems, cancer, diabetes, osteoporosis, and arthritis. A total count of the reported diseases and conditions was used in the analyses. Functional disability was measured with a composite score of 6 items of activities of daily living (ADLs; Katz, 1983) and 3 items of instrumental activities of daily living (IADL; Lawton & Brody, 1969). Using a yes/no response format, participants were asked whether they need help with each of the nine activities. The potential range of total scores was 0 (functional independence) to 6 (functional dependence on all items) for ADLs and 0 (functional independence) to 3 (functional dependence on all items) for IADLs. Self-rated health (Ware, 2004) was measured with a single item on how they would rate their health, with possible responses being poor (1), fair (2), good (3), and excellent (4). Self-rated health has been shown as a valid and reliable measure of general health in numerous population-based studies (August & Sorkin, 2010; Cummings & Jackson, 2008; Idler & Benyamini, 1997).

Social Resources

Social resources included social support and religious attendance. Social support was measured with a single question, “in times of trouble, can you count on at least some of your family or friends?” with three response options: (1) hardly ever, (2) some of the time, and (3) most of the time. A single-item measure of religious attendance asked how often the individual attended mass or religious services on a 5-point scale (1 = never or almost never; 5 = more than once a week).

Covariates

Sociodemographic characteristics included four racial/ethnic categories (non-Hispanic White, Black, Cuban, and non-Cuban Hispanic), age (in years), gender (0 = man, 1 = woman), education (0 = high school or less, 1 = beyond high school), marital status (0 = not married, 1 = married), and annual household income (0 = less than $30,000; 1 = beyond $30,000). Although marital status may be used as an indicator of social resources, it was treated as a control variable because perceived social support captures the overall satisfaction with relationships, and perceived quality is more important than the structural component.

Data Analysis

Analysis followed a three-step procedure. First, latent profile analysis (B. O. Muthén, 2001; B. O. Muthén & Shedden, 1999; Nylund, Asparouhov, & Muthén, 2007) was performed to determine the optimum number of clusters of physical health risks. The underlying assumption was that an unobserved heterogeneity of health risks exists and can be manifested through variability in multiple physical health indicators (e.g., chronic conditions, ADL, IADL, and self-rated health). The characteristics of the identified groups were then examined using an analysis of variance and χ² tests. Finally, we performed hierarchical multiple regression analyses to explore direct and interactive effects of health risk groups and social resources on depressive symptoms. Each block of variables (sociodemographics, health risk groups, social resources, and an interaction term between health risk groups and social resources) was entered in sequential order, and the model’s R ² change was calculated for the blocks of the variables. In hierarchical regression analyses, the White group was used as a reference group. Each group of interaction terms (social support and religious attendance) was separately entered into the direct effect model, and its statistical significance was evaluated. Analyses were performed using Mplus, version 6 (L. K. Muthén & Muthén, 1998–2010) and IBM Statistical Package for the Social Sciences, version 20.

Results

Subgroups With Varying Physical Health Risk Profiles

Latent profile analysis was conducted on the total sample (N = 1,432), using chronic conditions, ADL, IADL, and self-rated health as criterion variables. To identify the optimal number of clusters, we compared different model solutions ranging from a two-cluster through a five-cluster model. We evaluated several model fit criteria including the Bayesian Information Criterion (BIC), entropy, the Lo–Mendell–Rubin likelihood ratio test, and the bootstrap likelihood ratio test. The optimal cluster solutions could be achieved with lower BIC and higher entropy (i.e., an index of classification quality) values (B. O. Muthén & Muthén, 2000). The two likelihood ratio tests compare two adjacent models: the (c − 1)-cluster model versus the c-cluster model; a significant p value suggests the current model performs better than the prior model. Based on the selection criteria and model parsimony, we identified a three-cluster model as the best fitting model.

Table 1 presents profiles of the three health risk groups. The majority of older adults (80.6%) belonged to a low-risk group, which scored lowest on all domains of health constraints: On average, they had two chronic conditions, almost no ADL and IADL impairments and rated their health as good. The moderate-risk group (14.2%) had greater health constraints than the high-risk group; individuals in the group had a mean of 2.4 chronic conditions, minimal ADL impairment, and approximately one IADL impairment. A majority of the low-risk group rated their health as good (43.5%), whereas a majority of individuals in the moderate risk (39.9%) and high risk (42.9%) groups rated as fair and poor, respectively. The high-risk group consisted of individuals (5.2%) who had most health constraints; on average, they had 2.66 chronic conditions, approximately two ADL and IADL impairments, respectively. As noted, they were more likely to rate their health as fair to poor.

Table 1.

Profiles of Health Risks (N = 1,432).

Cluster/variables^a	Cluster 1	Cluster 2	Cluster 3	Test statistic and group comparisons^b
	Low risk	Moderate risk	High risk
	(n = 1,154)	(n = 203)	(n = 75)
	M (SD)	M (SD)	M (SD)
Chronic conditions (range 0–9)	2.0 (1.25)	2.40 (1.43)	2.66 (1.32)	F = 15.89***; (1:2, 3)^c
ADL (range 0–6)	0.09 (0.38)	0.51 (1.01)	1.97 (1.91)	F = 286.08***; (1:2, 3); (2:3)^d
IADL (range 0–3)	0.00 (0.00)	1.00 (0.00)	2.23 (0.43)	F = 25329.15***; (1:2, 3); (2:3)^d
Self-rated health (range 1–4)	2.23 (0.85)	2.72 (0.94)	3.16 (0.90)	F = 59.49***; (1:2, 3); (2:3)^d
Poor (%)	7.0	22.2	42.9
Fair (%)	29.3	39.9	35.7
Good (%)	43.5	25.8	15.7
Excellent (%)	20.2	12.1	5.7

Note. ADL = activity of daily living; IADL = instrumental activity of daily living.

^aHigher scores indicate more chronic conditions, ADL/IADL impairments, and poorer self-rated health.

^bStatistically different groups in the post hoc Tukey’s comparison at p < .05 are listed in parentheses.

^cGroup 1 is different from Groups 2 and 3, but Groups 2 and 3 are not significantly different.

^dGroup 1 is different from Groups 2 and 3, and Groups 2 and 3 are different.

***p < .001.

Group differences were further examined across sociodemographic characteristics, social resources, and depressive symptoms as shown in Table 2. All sociodemographic variables were statistically significant except being Black. Whites were most represented in the low-risk group, whereas Cubans and non-Cuban Hispanics were more likely to be placed in either the moderate-risk or high-risk groups. In addition, the moderate- and high-risk groups were older, more likely to be women, not married, and to have lower levels of education and income. Regarding social resources, the low-risk group had the highest level of social support followed by the moderate-risk and high-risk groups; there were no significant group differences on religious attendance. Finally, and consistent with our expectations, the high-risk group was associated with highest level of depressive symptoms followed by the moderate-risk and low-risk groups. Using the CES-D cutoff score (10 or higher), 17.3%, 46.6%, and 59.6% of individuals in the low-, moderate-, and high-risk groups, respectively, had a probable depression.

Table 2.

Cluster Differences by Sociodemographic Characteristics, Social Resources, and Depressive Symptoms.

	% or M (SD)			Test tatistic and group comparisons^a
	Low risk	Moderate risk	High risk	Test tatistic and group comparisons^a
Sociodemographic characteristics
Whites	40	16	16	χ² = 54.30***
Blacks	24	28	35	χ² = 4.94
Cubans	21	31	28	χ² = 10.73**
Non-Cuban Hispanics	15	25	21	χ² = 13.59**
Age	73.35 (6.44)	75.03 (7.67)	75.93 (7.63)	F = 9.70***; (1:2, 3)^b
Women	64	80	76	χ² = 21.43***
Married	45	30	31	χ² = 18.52***
Education (>high school)	48	33	38	χ² = 17.59***
Income (>$30,000)	29	11	11	χ² = 30.98***
Social resources
Social support	2.79 (0.54)	2.66 (0.65)	2.39 (0.76)	F = 18.75***; (1:2, 3); (2:3)^c
Religious attendance	3.16 (1.54)	3.26 (1.46)	2.90 (1.53)	F = 1.44
Depressive symptoms	5.18 (5.08)	10.35 (7.14)	11.06 (6.65)	F = 82.32***; (1: 2, 3)^b

Note. ^aChi-square tests for categorical variables show only test statistics. For continuous variables, statistically significant analysis of variance led to multiple comparisons; statistically different groups in the post hoc Tukey’s comparison at p < .05 are listed in parentheses.

^bGroup 1 is different from Groups 2 and 3, but Groups 2 and 3 are not significantly different.

^cGroup 1 is different from Groups 2 and 3, and Groups 2 and 3 are different.

p < .05. **p < .01. ***p < .001.

Multiple Regression Analyses

Results from hierarchical regression analyses are summarized in Table 3. In these analyses, the lowest health risk group was used as a reference group; each of the higher risk groups was therefore contrasted with the lowest risk group. In the initial model with sociodemographic variables, 13% of the total variance in depressive symptoms was explained. Greater levels of depressive symptoms were associated with being Cubans, being non-Cuban Hispanics, women, not being married, and having lower levels of education and income. The addition of the moderate- or high-risk groups increased explained variance by 9%. Those who were in moderate-risk or high-risk groups were each more likely to report depressive symptoms (as noted earlier, this was in contrast to the low health risk group). When social resources variables were included, an additional 3% of the variance was explained. Greater levels of social support and religious attendance led to lower levels of depressive symptoms. In the final model with interaction terms, the interaction between the moderate-risk group and social support was statistically significant, adding 1% of explained variance. Overall, this final model explained 26% of the total variance. The statistically significant interaction effect (b = −1.74, p < .05) between moderate health risk grouping and social support indicated that an increase in social support was associated with a significant reduction in depressive symptoms in the moderate-risk group compared to the low-risk group. This finding points out the greater benefits of social support for the moderate-risk group compared to the low-risk group.

Table 3.

Hierarchical Regression Models of Depressive Symptoms.

Variable	Model 1		Model 2		Model 3		Model 4
Variable	b	SE	b	SE	b	SE	b	SE
Sociodemographic characteristics
Blacks^a	0.63	0.48	–0.01	0.46	0.24	0.47	0.25	0.47
Cubans^a	1.95***	0.51	1.22*	0.49	1.13*	0.48	1.16*	0.48
Non-Cuban Hispanics^a	1.55**	0.56	0.73	0.54	0.54	0.54	0.53	0.54
Age	–0.04	0.03	–0.06*	0.03	–0.06*	0.03	–0.06*	0.03
Women	1.09**	0.40	0.86*	0.38	1.07**	0.37	1.10**	0.37
Married	–1.18**	0.40	–0.90*	0.38	–0.80*	0.38	–0.79*	0.38
Education (>high school)	–1.32**	0.38	–1.28***	0.37	–1.32***	0.36	–1.29***	0.36
Income (>$30,000)	–2.32***	0.46	–2.11***	0.44	–1.86***	0.43	–1.88***	0.43
Health risk groups
Moderate-risk group			4.66***	0.52	4.44***	0.51	9.11***	2.21
High-risk group			5.28***	0.92	4.20***	0.92	5.20	2.80
Social resources
Social support					–1.84***	0.31	–1.49***	0.37
Religious attendance					–0.23*	0.12	–0.22	0.12
Interactions
Moderate Risk × Social Support							–1.74*	0.80
High Risk × Social Support							–0.36	1.14
Overall R ²	.13		.22		.25		.26
ΔR ²	.13***		.09***		.03***		.01*

Note. ^aIn racial/ethnic categories, the White group was used for a reference and was omitted from the analysis.

*p < .05. **p < .01. ***p < .001.

Discussion

The present study explored associations among physical health risks, social resources, and depressive symptoms in diverse community-dwelling populations. One goal of the study was to develop an empirical typology of health risk based on multiple physical health indicators and to examine the moderating effects of social resources in the relation between health risk groupings and depressive symptoms. Using latent profile analysis, typologies of health risks were developed under the assumption that there would be natural groupings of individuals with varying degrees of health constraints. We also intended to examine whether these natural groupings differed in terms of the impact of social resources on depressive symptoms.

As hypothesized, we found multiple groups of individuals with different profiles of health risks. All solutions suggested that most of these older adults fell into a general low-risk category; of note is that the higher risk groups generally included more racially/ethnically diverse participants. Review of the latent profile analyses indicated that of the five models, a three-group solution provided the best fit with the data. This typology identified low-risk (n=1,154), moderate-risk (n=203), and high-risk (n=75) groups. In the relationships between health risk groups and study variables, the moderate- and high-risk groups were more likely than the low-risk group to be Cubans or non-Cuban Hispanics, older, women, not married, and to have lower levels of education and income; they also had less social support and higher levels of depressive symptoms than the low-risk group. These findings are consistent with evidence in the literature regarding risk factors associated with the status of physical and mental health (Akincigil et al., 2012; Centers for Disease Control and Prevention, 2010; Chiriboga et al., 2002; Hewitt, Turrell, & Giskes, 2012; Mezuk et al., 2010).

In hierarchical regression models, health risk groups and social resources variables remained significant after controlling for demographic variables. When the interaction terms entered, however, the religious attendance variable was no longer significant. The interactive effects of health risk groups and social support suggest that social support reduced depressive symptoms for the moderate-risk group, but not for the high-risk group, in reference to the low-risk group.

These findings increase our understanding of the role of health factors. Decline in physical health is one of the most common concerns and sources of stress in older populations (Bekman et al., 1986; Bruce, 1999; Jang et al., 2008; Mezuk et al., 2012). The strong association between health risk groupings and depressive symptoms found in the present study is consistent with previous studies supporting the critical role of physical health on determining mental health status of diverse populations (Chiriboga et al., 2002; Jang et al., 2002). These results are also similar to previous studies where findings suggest that although health constraints generally have negative mental health consequences, some individuals may be more vulnerable to the risks and gain benefits from social resources (Ganster & Victor, 1988). The present study not only identified groups of individuals with higher risks of health constraints but also provided evidence of moderating effects of social resources in the relation between health risks and depressive symptoms in a racially/ethnically diverse population.

As hypothesized, the present findings support the premise that social resources may reduce the impact of physical health on mental health: Members of the three risk profiles were affected differently as a function of social resources. In particular, those with moderate health risks were shown to benefit the most from social resources. Thus, the findings are not completely consistent with the buffering effect hypothesis for social resource (i.e., the extent to which the health of an individual has consequences for depressive symptoms is moderated by his or her social resources; Uchino, Cacioppo, & Kiecolt-Glaser, 1996). The mechanism behind the moderate-risk group being most likely to benefit deserves some consideration.

One possible explanation is that health conditions of this group—in between those of the low- and high-risk groups by definition—may be at a level where the presence of social support is more likely to play a role. The social support may not be particularly important for those in relatively good health and may be inadequate to buffer the effect of health for those whose health problems are most serious; also the insignificant finding for the lower risk group may simply result from the fact that healthier individuals generally have fewer depressive symptoms and therefore manifest less variance. Further research on those with moderate health risks is clearly called for. Such research may also wish to test the intervening mechanism of social resources on mental health where the stressor diminishes personal and social resources, which influences mental health outcomes (Pearlin & Bierman, 2013; Thoits, 2011; Yang, 2006). For example, using a population-based longitudinal study of older adults, Yang (2006) found that perceived social support mediated the detrimental effects of functional disability on depressive symptoms as playing the role of a stressor reducer. Another explanation is that the moderate-risk group had higher proportions of both Cuban Americans and Hispanics from other groups, in contrast to the low-risk group. Due to factors such as immigration history and levels of acculturation, social support may take on a more important value for Hispanic populations (Mendes de Leon & Glass, 2004; Rodríguez-Galán & Falcón, 2010).

Of the two social resources variables included in the study, social support had direct as well as indirect effects on depressive symptoms, whereas religious attendance had only direct effects. Past research supports the idea that of the several domains of social resources, functional social support (e.g., perceived social support, emotional social support) is consistently associated with mental health outcomes, whereas the association with structural social support (e.g., the size of social network, frequency of social contacts) is weaker (Thoits, 2011; Uchino et al., 1996; Yang, 2006). Although a buffering effect of religiosity on mental health has been reported in numerous studies (e.g., Jang, Bergman, Schonfeld, & Molinari, 2006; Koenig et al., 2001), the single question on religious attendance employed in the present study may not be sensitive to detect this effect. In line with the buffering effect of perceived social support, measures more directly linked to religiosity, including the social and emotional aspects of attendance (e.g., church-based support), could be used in future studies. Also, considering the differing role of religiosity across racial/ethnic groups (Krause, 2006), the group-specific buffering role of religious support could then be examined more conclusively.

The present study has several limitations. First, the study used cross-sectional data that did not allow the drawing of causal inference. For example, as much as health constraints affect mental health status, depressive symptoms may adversely affect the course and outcome of common physical conditions such as physical performance, arthritis, diabetes, and cancer (Bekman et al., 1986; Centers for Disease Control and Prevention, 2010; Lee et al., 2012; Mezuk et al., 2012). Second, the study sample was drawn in the state of Florida, which may limit generalizing the findings to other geographic areas. Third, social resources such as social support and religiosity were measured using single items. Considering that multifaceted nature of social resources command distinct influences on mental health (Berkman & Glass, 2000; Thoits, 2011), future studies need to employ validated multiple-item instruments to examine specific associations and mechanisms (Uchino et al., 1996).

It is important to note that one of the strengths of the study was its inclusion of diverse racial/ethnic minority older adults that allowed comparisons that went beyond Black and White differences. The finding that racial/ethnic minorities (particularly Cubans and non-Cuban Hispanics) manifest significantly more depressive symptoms is consistent with other findings (Akincigil et al., 2012). The finding may be further explored in future studies in conjunction with consideration of the roles of poor physical health (Liang et al., 2010) and fewer social resources (e.g., Mendes de Leon & Glass, 2004) in racial/ethnic minority older adults. Understanding the constellations of disadvantages could help practitioners provide specific services (such as building community-based or church-based support systems) to at-risk populations to increase their social resources.

In closing, the current study identified subgroups of older adults with varying health risks and examined how specific groups of individuals are associated with social resources and depressive symptoms. Using the person-centered approach, the present study in part addressed the limitations of variable-centered approach regarding inconsistent findings associated with the isolated effects of health risk variables and multicollinearity of the related variables. The findings on interactions suggest the varying levels of impact of social support on mental health in older adults with different health risk profiles.

Footnotes

Declaration of Conflicting Interests

The author(s) 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 project was supported by the Administration on Aging Research (grant 90AM2750; Dr. Jennifer Salmon, principal investigator; Dr. David A. Chiriboga, coprincipal investigator).

References

Agahi

Parker

M. G.

(2008). Leisure activities and mortality: Does gender matter? Journal of Aging and Health, 20, 855–871.

Akincigil

Olfson

Siegel

Zurlo

K. A.

Walkup

J. T.

Crystal

(2012). Racial and ethnic disparities in depression care in community-dwelling elderly in the United States. American Journal of Public Health, 102, 319–328. doi:10.2105/ajph.2011.300349

Akincigil

A., Olfson, M., Walkup, J. T., Siegel, M. J., Kalay, E., Amin, S., . . . Crystal, S.

(2011). Diagnosis and treatment of depression in older community-dwelling adults: 1992–2005. Journal of the American Geriatrics Society, 59, 1042–1051. doi:10.1111/j.1532-5415.2011.03447.x

Andresen

E. M.

Malmgren

J. A.

Carter

W. B.

Patrick

D. L.

(1994). Screening for depression in well older adults: Evaluation of a short form of the CES-D (Center for Epidemiologic Studies Depression Scale). American Journal of Preventive Medicine, 10, 77–84.

Antonucci

T. C.

(2001). Social relations: An examination of social networks, social support, and sense of control. In Birren

J. E.

Schaie

K. E.

. (Eds.), Handbook of the psychology of aging (5th. ed., pp. 427–453). San Diego, CA: Academic Press.

August

K. J.

Sorkin

D. H.

(2010). Racial and ethnic disparities in indicators of physical health status: Do they still exist throughout late life? Journal of American Geriatrics Society, 58, 2009–2015. doi:10.1111/j.1532-5415.2010.03033.x

Barnes

L. L.

Mendes de Leon

Bienias

J. L.

Evans

D. A.

(2004). A longitudinal study of black-white differences in social resources. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 59, S146–S153.

Barnes

L. L.

Mendes de Leon

Wilson

R. S.

Bienias

J. L.

Evans

D. A.

(2004). Social resources and cognitive decline in a population of older African Americans and whites. Neurology, 63, 2322–2326. doi:10.1212/01.wnl.0000147473.04043.b3

Bekman

L. F., Berkman, C. S., Kasl, S., Freeman, D. H., Leo, L., Ostfeld, A. M., . . . Brody, J. A.

(1986). Depressive symptoms in relationship to physical health and functioning in the elderly. American Journal of Epidemiology, 124, 372–388.

10.

Berkman

L. F.

Glass

T. A.

(2000). Social integration, social network, social support, and health. In Berkman

L. F.

Kawachi

. (Eds.), Social epidemiology (pp. 137–173). New York, NY: Oxford University Press.

11.

Black

S. A.

Markides

K. S.

Miller

T. Q.

(1998). Correlates of depressive symptomatology among older community-dwelling Mexican Americans: The Hispanic EPESE. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 53B, S198–S208. doi:10.1093/geronb/53B.4.S198

12.

Blazer

D. G.

(2003). Depression in late life: Review and commentary. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 58, M249–M265. doi:10.1093/gerona/58.3.M249

13.

Blazer

D. G.

Hybels

C. F.

Pieper

C. F.

(2001). The association of depression and mortality in elderly persons. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 56, M505–M509. doi:10.1093/gerona/56.8.M505

14.

Blazer

D. G.

Landerman

L. R.

Hays

J. C.

Simonsick

E. M.

Saunders

W. B.

(1998). Symptoms of depression among community-dwelling elderly African-American and white older adults. Psychological Medicine, 28, 1311–1320.

15.

Bruce

M. L.

(1999). The association between depression and disability. American Journal of Geriatric Psychiatry, 7, 8–11.

16.

Buchman

A. S.

Boyle

P. A.

Wilson

R. S.

Fleischman

D. A.

Leurgans

Bennett

D. A.

(2009). Association between late-life social activity and motor decline in older adults. Archives of Internal Medicine, 169, 1139–1146. doi:10.1001/archinternmed.2009.135

17.

Carstensen

L. L.

Fung

H. H.

Charles

S. T.

(2003). Socioemotional selectivity theory and the regulation of emotion in the second half of life. Motivation and Emotion, 27, 103–123. doi:10.1023/a:1024569803230

18.

Centers for Disease Control and Prevention. (2010). Morbidity and mortality weekly report (Vol. 59/No. 38). U.S. Department of Health and Human Services: Atlanta, GA.

19.

Chiriboga

D. A.

Black

S. A.

Aranda

Markides

(2002). Stress and depressive symptoms among Mexican American elders. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 57, P559–P568. doi:10.1093/geronb/57.6.P559

20.

Cochran

D. L.

Brown

D. R.

McGregor

K. C.

(1999). Racial differences in the multiple social roles of older women: Implications for depressive symptoms. The Gerontologist, 39, 465–472.

21.

Cummings

J. L.

Jackson

P. B.

(2008). Race, gender, and SES disparities in self-assessed health, 1974–2004. Research on Aging, 30, 137–167. doi:10.1177/0164027507311835

22.

Davis

K. J., Sloane, P. D., Mitchell, C. M., Preisser, J., Grant, L., Hawes, M. C., . . . Koch, G. G.

(2000). Specialized dementia programs in residential care settings. The Gerontologist, 40, 32–42. doi:10.1093/geront/40.1.32

23.

Dickinson

W. J.

Potter

G. G.

Hybels

C. F.

McQuoid

D. R.

Steffens

D. C.

(2011). Change in stress and social support as predictors of cognitive decline in older adults with and without depression. International Journal of Geriatric Psychiatry, 26, 1267–1274. doi:10.1002/gps.2676

24.

Dunlop

D. D.

Song

Lyons

J. S.

Manheim

L. M.

Chang

R. W.

(2003). Racial/ethnic differences in rates of depression among preretirement adults. American Journal of Public Health, 93, 1945–1952. doi:10.2105/ajph.93.11.1945

25.

Everitt

B. S.

Landau

Leese

(2001). Cluster analysis (4th ed.). New York, NY: Oxford University Press.

26.

Fiori

K. L.

Antonucci

T. C.

Cortina

K. S.

(2006). Social network typologies and mental health among older adults. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 61, P25–P32.

27.

Fiori

K. L.

Jager

(2012). The impact of social support networks on mental and physical health in the transition to older adulthood. International Journal of Behavioral Development, 36, 117–129. doi:10.1177/0165025411424089

28.

Ganster

D. C.

Victor

(1988). The impact of social support on mental and physical health. British Journal of Medical Psychology, 61, 17–36. doi:10.1111/j.2044-8341.1988.tb02763.x

29.

Glass

T. A.

Mendes de Leon

Bassuk

S. S.

Berkman

L. F.

(2006). Social engagement and depressive symptoms in late life. Journal of Aging and Health, 18, 604–628. doi:10.1177/0898264306291017

30.

Gonzalez

H. M.

Haan

M. N.

Hinton

(2001). Acculturation and the prevalence of depression in older Mexican Americans: Baseline results of the Sacramento Area Latino Study on Aging. Journal of American Geriatrics Society, 49, 948–953.

31.

Hair

J. F.

Jr. Black

W. C.

(2000). Cluster analysis. In Grimm

L. G.

Yarnold

P. R.

. (Eds.), Reading and understanding more multivariate statistics (pp. 147–205). Washington, DC: American Psychological Association.

32.

Hewitt

Turrell

Giskes

(2012). Marital loss, mental health and the role of perceived social support: Findings from six waves of an Australian population based panel study. Journal of Epidemiology and Community Health, 66, 308–314. doi:10.1136/jech.2009.104893

33.

Hong

Hasche

Bowland

(2009). Structural relationships between social activities and longitudinal trajectories of depression among older adults. The Gerontologist, 49, 1–11. doi:10.1093/geront/gnp006

34.

Idler

E. L.

Benyamini

(1997). Self-rated health and mortality: A review of twenty-seven community studies. Journal of Health and Social Behavior, 38, 21–37.

35.

James

B. D.

Boyle

P. A.

Buchman

A. S.

Bennett

D. A.

(2011). Relation of late-life social activity with incident disability among community-dwelling older adults. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 66A, 467–473. doi:10.1093/gerona/glq231

36.

Jang

Bergman

Schonfeld

Molinari

(2006). Depressive symptoms among older residents in assisted living facilities. International Journal of Aging and Human Development, 63, 299–315.

37.

Jang

Chiriboga

D. A.

Kim

Phillips

(2008). Depressive symptoms in four racial and ethnic groups: The Survey of Older Floridians (SOF). Research on Aging, 30, 488–502. doi:10.1177/0164027508316617

38.

Jang

Haley

W. E.

Small

B. J.

Mortimer

J. A.

(2002). The role of mastery and social resources in the associations between disability and depression in later life. The Gerontologist, 42, 807–813.

39.

Johnson

T. P.

(1991). Mental health, social relations, and social selection: A longitudinal analysis. Journal of Health and Social Behavior, 32, 408–423.

40.

Katz

(1983). Assessing self-maintenance: Activities of daily living, mobility, and instrumental activities of daily living. Journal of American Geriatrics Society, 31, 721–727.

41.

Koenig

H. G.

McCullough

M. E.

Larson

D. B.

(2001). Handbook of religion and health. New York, NY: Oxford University Press.

42.

Krause

(1987). Satisfaction with social support and self-rated health in older adults. The Gerontologist, 27, 301–308. doi:10.1093/geront/27.3.301

43.

Krause

(2002). Church-based social support and health in old age. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 57, S332–S347. doi:10.1093/geronb/57.6.S332

44.

Krause

(2006). Exploring the stress-buffering effects of church-based and secular social support on self-rated health in late life. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 61, S35–S43.

45.

Krause

Bastida

(2011). Religion, suffering, and self-rated health among older Mexican Americans. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 66B, 207–216. doi:10.1093/geronb/gbq086

46.

Lawton

M. P.

Brody

E. M.

(1969). Assessment of older people: Self-maintaining and instrumental activities of daily living. The Gerontologist, 9, 179–186. doi:10.1093/geront/9.3_Part_1.179

47.

Lebowitz

B. D., Pearson, J. L., Schneider, L. S., Reynolds, C. F., Alexopoulos, G. S., Bruce, M. L., . . . Katz, I. R.

(1997). Diagnosis and treatment of depression in late life: Consensus statement update. JAMA: The Journal of the American Medical Association, 278, 1186–1190. doi:10.1001/jama.1997.03550140078045

48.

Lee

C.-T.

Leoutsakos

J.-M.

Lyketsos

C. G.

Steffens

D. C.

Breitner

J. C. S.

Norton

M. C.

(2012). Latent class-derived subgroups of depressive symptoms in a community sample of older adults: The Cache County Study. International Journal of Geriatric Psychiatry, 27, 1061–1069. doi:10.1002/gps.2824

49.

Liang

Quiñones

A. R.

Bennett

J. M.

Shaw

B. A.

Ofstedal

M. B.

(2010). Evolving self-rated health in middle and old age: How does it differ across Black, Hispanic, and White Americans? Journal of Aging and Health, 22, 3–26. doi:10.1177/0898264309348877

50.

Louie

G. H.

Ward

M. M.

(2011). Socioeconomic and ethnic differences in disease burden and disparities in physical function in older adults. American Journal of Public Health, 101, 1322–1329. doi:10.2105/ajph.2010.199455

51.

Maier

Klumb

P. L.

(2005). Social participation and survival at older ages: Is the effect driven by activity content or context? European Journal of Ageing, 2, 31–39. doi:10.1007/s10433-005-0018-5

52.

McPherson

Smith-Lovin

Brashears

M. E.

(2006). Social isolation in America: Changes in core discussion networks over two decades. American Sociological Review, 71, 353–375.

53.

Mendes de Leon

Glass

T. A.

(2004). The role of social and personal resources in ethnic disparities in late-life health. In Anderson

N. B.

Bulatao

R. A.

Cohen

. (Eds.), Critical perspectives on racial and ethnic differences in health in late life. Washington, DC: National Academic Press.

54.

Mendes de Leon

Glass

T. A.

Berkman

L. F.

(2003). Social engagement and disability in a community population of older adults. American Journal of Epidemiology, 157, 633–642. doi:10.1093/aje/kwg028

55.

Mezuk

Edwards

Lohman

Choi

Lapane

(2012). Depression and frailty in later life: A synthetic review. International Journal of Geriatric Psychiatry, 27, 879–892. doi:10.1002/gps.2807

56.

Mezuk

B., Rafferty, J. A., Kershaw, K. N., Hudson, D., Abdou, C. M., Lee, H., . . . Jackson, J. S.

(2010). Reconsidering the role of social disadvantage in physical and mental health: Stressful life events, health behaviors, race, and depression. American Journal of Epidemiology, 172, 1238–1249. doi:10.1093/aje/kwq283

57.

Moak

Z. B.

Agrawal

(2010). The association between perceived interpersonal social support and physical and mental health: Results from the national epidemiological survey on alcohol and related conditions. Journal of Public Health, 32, 191–201. doi:10.1093/pubmed/fdp093

58.

Muthén

B. O.

(2001). Latent variable mixture modeling. In Marcoulides

G. A.

Schumacker

R. E.

. (Eds.), New developments and techniques in structural equation modeling (pp. 1–33). Mahwah, NJ: Lawrence Erlbaum.

59.

Muthén

B. O.

Muthén

L. K.

(2000). Integrating person-centered and variable-centered analyses: Growth mixture modeling with latent trajectory classes. Alcoholism: Clinical and Experimental Research, 24, 882–891. doi:10.1111/j.1530-0277.2000.tb02070.x

60.

Muthén

B. O.

Shedden

(1999). Finite mixture modeling with mixture outcomes using the EM algorithm. Biometrics, 55, 463–469.

61.

Muthén

L. K.

Muthén

B. O.

(1998–2010). Mplus user’s guide (6th ed.). Los Angeles, CA: Author.

62.

Nylund

K. L.

Asparouhov

Muthén

B. O.

(2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal, 14, 535–569.

63.

Park

N. S.

Jang

Lee

B. S.

Chiriboga

D. A.

(2013). Racial/ethnic differences in predictors of self-rated health: Findings from the Survey of Older Floridians. Research on Aging, 35, 201–219. doi:10.1177/0164027512440572

64.

Park

N. S.

Klemmack

D. L.

Roff

L. L.

Parker

M. W.

Koenig

H. G.

Sawyer

Allman

R. M.

(2008). Religiousness and longitudinal trajectories in elders’ functional status. Research on Aging, 30, 279–298. doi:10.1177/0164027507313001

65.

Pearlin

L. I.

Bierman

(2013). Current issues and future directions in research into the stress process. In Aneshensel

C. S.

Phelan

J. C.

Bierman

. (Eds.), Handbook of the sociology of mental Health (2nd ed., pp. 325–340). New York, NY: Springer.

66.

Pfeiffer

(1975). A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. Journal of American Geriatrics Society, 23, 433–441.

67.

Radloff

(1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401.

68.

Reyes-Ortiz

C. A.

Pelaez

Koenig

H. G.

Mulligan

(2007). Religiosity and self-rated health among Latin American and Caribbean elders. International Journal of Psychiatry in Medicine, 37, 425–443.

69.

Rodríguez-Galán

M. B.

Falcón

L. M.

(2010). Patterns of social activity engagement among older Hispanics and their relationship to sociodemographic and health variables. Activities, Adaptation & Aging, 34, 251–275. doi:10.1080/01924788.2010.523868

70.

Salomon

J. A.

Nordhagen

Oza

Murray

C. J. L.

(2009). Are Americans feeling less healthy? The puzzle of trends in self-rated health. American Journal of Epidemiology, 170, 343–351. doi:10.1093/aje/kwp144

71.

Schmiege

S. J.

Meek

Bryan

A. D.

Petersen

(2012). Latent variable mixture modeling: A flexible statistical approach for identifying and classifying heterogeneity. Nursing Research, 61, 204–212 doi:10.1097/NNR.0b013e3182539f4c

72.

Seeman

T. E.

(2000). Health promoting effects of friends and family on health outcomes in older adults. American Journal Health Promotion, 14, 362–370. doi:https://dx-doi-org.web.bisu.edu.cn/10.4278/0890-1171-14.6.362

73.

Seeman

T. E.

Merkin

S. S.

Crimmins

E. M.

Karlamangla

A. S.

(2010). Disability trends among older Americans: National Health and Nutrition Examination Surveys, 1988-1994 and 1999-2004. American Journal of Public Health, 100, 100–107. doi:10.2105/AJPH.2008.157388

74.

Seeman

T. E.

Miller-Martinez

D. M.

Stein Merkin

Lachman

M. E.

Tun

P. A.

Karlamangla

A. S.

(2011). Histories of social engagement and adult cognition: Midlife in the U.S. Study. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 66B, i141–i152. doi:10.1093/geronb/gbq091

75.

Sun

F., Park, N. S., Roff, L. L., Klemmack, D. L., Parker, M., Koenig, H. G., . . . Allman, R. M.

(2012). Predicting the trajectories of depressive symptoms among southern community-dwelling older adults: The role of religiosity. Aging & Mental Health, 16, 189–198. doi:10.1080/13607863.2011.602959

76.

Thoits

P. A.

(2011). Mechanisms linking social ties and support to physical and mental health. Journal of Health and Social Behavior, 52, 145–161. doi:10.1177/0022146510395592

77.

Turner

J. B.

Turner

R. J.

(2013). Social relations, social integration, and social support. In Aneshensel

C. S.

Phelan

J. C.

Bierman

. (Eds.), Handbook of the sociology of mental health (pp. 341–356). New York, NY: Springer.

78.

Uchino

B. N.

Cacioppo

J. T.

Kiecolt-Glaser

J. K.

(1996). The relationship between social support and physiological processes: A review with emphasis on underlying mechanisms and implications for health. Psychological Bulletin, 119, 488–531.

79.

Ware

J.,

Jr . (Ed.) (2004). SF-36 health survey update. New Jersey, NY: Lawrence Erlbaum.

80.

Schimmele

C. M.

Chappell

N. L.

(2012). Aging and late-life depression. Journal of Aging and Health, 24, 3–28. doi:10.1177/0898264311422599

81.

Yang

(2006). How does functional disability affect depressive symptoms in late life? The role of perceived social support and psychological resources. Journal of Health and Social Behavior, 47, 355–372. doi:10.1177/002214650604700404

82.

Zayac

Salmon

J. R.

Chiriboga

D. A.

(2005). Florida’s minority elders: An assessment of physical, mental health, and long-term care service needs. Tampa, FL: Florida Policy Exchange Center on Aging.

83.

Zunzunegui

M.-V.

Alvarado

B. E.

Del Ser

Otero

(2003). Social networks, social integration, and social engagement determine cognitive decline in community-dwelling Spanish older adults. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 58, S93–S100. doi:10.1093/geronb/58.2.S93