Advances in Applied Sociology
2012. Vol.2, No.4, 313-319
Published Online December 2012 in SciRes (http://www.SciRP.org/journal/aasoci) http://dx.doi.org/10.4236/aasoci.2012.24041
Predictors of Depressive Symptoms: What Are the Roles of
Geography and Informal Social Support?
Timothy S. Killian, Megan Penfield
Department of Human Environmental Sciences, University of Arkansas, Fayetteville, USA
Received August 23rd, 2012; revised September 25th, 2012; accepted October 9th, 2012
Using data from the 2004 wave of the Health and Retirement Study (HRS), three central research ques-
tions were examined. First, are there variations in depressive symptoms by geographic region? Second,
are variations in depressive symptoms related to informal social support? Third, are there interactions
between geography and informal social support in regard to predicting depressive symptoms? Results
from this study found a small, but significant difference in depressive symptoms by geography region.
Also, informal social support from children and friends were predictive of lower levels of depression.
Furthermore, informal social support interacted with region to explain additional variation in depressive
symptoms. Participants’ perceptions that they could rely on their adult children to meet their needs was
more salient in regard to reducing depressive symptoms for exurban than non-exurban participants, and
increased limitations in mobility, strength, and fine motor skills were more influential in explaining de-
pressive symptoms in exurban than other persons.
Keywords: Depression; Aging; Rurality; Social Support; Older Adults
Predictors of Depressive Symptoms: What Are
the Roles of Geography and Informal Social
Support?
There are many social and personal costs associated with de-
pressive symptoms in late life. In terms of social costs, older
persons’ depressive symptoms are positively related to the
utilization of both formal and informal caregiving. It has been
estimated that the caregiving costs associated with depressive
symptoms in older persons in the US are about $9 billion per
year (Langa, Valenstein, Fendrick, Kabeto, & Vijan, 2004).
Callahan, Kesterson, and Tierney (1997) found that, when con-
trolling for physical health, older persons who report a larger
number of depressive symptoms are more likely to use emer-
gency departments and be hospitalized, as well as incur greater
diagnostic test charges over a one year period. Ambulatory
costs and inpatient hospital costs have been estimated to be
between 43 percent to 51 percent higher and 47 percent to 53
percent higher for depressed than for nondepressed older per-
sons (after adjusting for chronic medical illness) (Katon, Lin,
Russo, & Unutzer, 2003). The same study found that increased
medical costs were not significantly different between those
actually diagnosed with depression and those reporting high
depressive symptomatology; in either case, medical costs were
significantly higher than those exhibiting less symptoms of
depression.
In addition to the social costs associated with depressive
symptoms, there are many personal costs. Older persons who
report a greater number of depressive symptoms experience
physical decline at faster rates than those who report lesser
symptoms of depression, even after adjusting for health status
and other sociodemographic factors (Penninx et al., 1998).
Additionally, research has found that depression, along with
poor social integration and dementia, are the most consistent
predictors of death in older persons, especially for those over
75 years of age (Endegal, 1996). There are significant personal
costs to family members that provide care for older persons
reporting symptoms of depression. For instance, even after
controlling for sociodemographic factors, disease, and disease
symptoms, caregivers of older persons reporting greater symp-
toms of depression experience more strain and stress due to
caregiving, which has been related to increased caregiver mor-
tality (Schulz & Beach, 1999).
There is increasing evidence that place of residence make
one more likely to experience depressive symptoms and less
able to cope with depression. Rathbone-McCuan and Bane
(2003) indicated that limited mental health resources, geo-
graphic distance to services, attitudes toward areas. Further-
more, older rural persons might be more reluctant to use mental
health services even when they are available because utilization
of those services is perceived to be at odds with individualism
and self-reliance that are especially valued in rural areas (Gatz
& Smyer, 1992). For these reasons, in addition to the more
rapid pace of population aging in rural areas (Rogers, 2002a), it
is critical to understand variation in depressive symptoms
across place of residence.
The purpose of this paper was to address the following ques-
tions. First, are there variations in depressive symptoms across
place of residence? Second, are there variations in the provision
of informal social support across place of residence? Finally,
does informal social support mitigate depressive symptoms and,
if so, does mitigation of informal social support vary by resi-
dence? Then, predictive models of depressive symptoms were
developed that focused on variations in both residence and
informal social support while controlling for physical health
and sociodemographic variables likely related to depressive
symptoms. mental health problems, the fear of stigmatization,
and lack of affordable care were all barriers to mental health
service utilization in rural.

T. S. KILLIAN, M. PENFIELD
Literature Review and Background
The Center for Epidemiologic Studies Depression Scale
(CES-D) (Radloff, 1977) is a widely-used measure of depres-
sion with well-established reliability and validity estimates
when used with older adults (Himmelfarb & Murrell, 1983;
Kohout, Berkman, Evans, & Cornoni-Huntley, 1993). Recent
research on depressive symptoms using the CES-D has shown
that depressive symptoms significantly affect quality of life in
older persons. In a study of persons aged 71 years old and older,
Penninx et al. (1998) found that depressive symptoms predicted
declines in physical functioning across a span of four years,
even after adjusting for baseline functioning scores, health
scores, and sociodemographic variables. Also, using the CES-D,
Wilson, Mendes de Leon, Bennett, Bienias, and Evans (2004)
found a significant positive relationship between initial levels
of depressive symptoms and cognitive decline during a five-
year-period. Because of the social and personal costs associated
with depressive symptoms among older adults, it is important
to develop a greater understanding of variations in the experi-
ence of depressive symptoms that include variations in the con-
text within which they are experienced and coped with.
Impairments in functioning and poor self-reported health
have been frequently studied in association with depressive
symptoms (e.g., Killian, Turner, & Cain, 2005). It is clear that a
limited ability to perform Activities of Daily Living (ADLs)
and Instrumental Activities of Daily Living (IADLs) increases
the likelihood of depression (Probst et al., 2006; Yang &
George, 2005). A recent study on over 22,000 individuals from
11 different European countries (Braam et al., 2005) found a
positive relationship between physical health and depression in
each country, although the magnitude of the relationship dif-
fered across countries. In particular, the association between
physical health and depressive symptoms was more pronounced
in the UK and Ireland than in western continental European and
Scandinavian countries. Braam and colleagues hypothesized
that this might be due to greater access to health services in
western continental European and Scandinavian countries as
compared to the western isles (UK and Ireland). Analogously,
given the increased barriers in accessing health care services for
rural US populations relative to persons living in more urban
regions (Rogers, 2002b), it would be reasonable to expect that
the association between physical health and depression would
be greater in rural populations than nonrural. Poor self-rated
health has been found to be a robust predictor of depression in
older adults (Killian, Turner, & Cain, 2005), with reported odds
ratios as high as 5.6 for older individuals and 6.8 among the
oldest old (those over 85 years old) (Djernes, 2006).
Beyond physical health, gender and marital status are also
important predictors of depressive symptoms in older persons.
In a review of literature on depression in older adults, Djernes
(2006) found that older women were significantly more likely
to have depressive symptoms and disorders than older men
(with odds ratios as high as 3.4). However, other studies have
suggested that this gender gap is reduced with aging (Ried &
Planas, 2002). Being married is associated with less depressive
symptoms (Zunzunegui, Beland, Llacer, & Leon, 1998). Bier-
man, Fazio, and Milkie (2006) suggested that married individu-
als report less depression than nonmarried individuals because
of their greater access to socioeconomic and psychosocial re-
sources. Additionally, research has shown that the transition to
widowhood, especially when recent, doubles the likelihood that
a person will be depressed (Djernes, 2006).
In their review of the literature on social relationships and
health, House, Landis, and Umberson (1988) highlighted re-
search which indicated that less socially integrated individuals
suffer more physical and mental health problems and are more
likely to die. In a study examining the relations between func-
tional disability, depression, and perceived social support, Yang
(2006) found that functional disability alone does not necessar-
ily lead to increased depressive symptoms. Rather, it seems that
functional disability, in conjunction with deficits in high-qual-
ity social relationships, contributed to higher levels of depress-
sion. Yang contended that improving the quality of social rela-
tionships in the elderly is a critical component to reducing risks
of depression in older disabled persons. There is reason to be-
lieve that this might be especially true for rural older persons
who often face significant barriers in accessing formal support
(Rogers, 2002b) potentially making informal support more
important for them than their nonrural counterparts.
Often left out of predictive models of depressive symptoms
among older persons is geographic region. This is likely a key
variable in understanding depressive symptoms for several
reasons. First, as already noted, there may be more significant
barriers to accessing rural mental health services in rural areas
as compared to less rural areas (Rathbone, McCuan, & Bane,
2003). Also, in spite of strong normative obligations of family
members to provide assistance to aging family members
(Killian & Ganong, 2002), out migration of adult children may
reduce the availability of informal caregiving assistance more
for rural than nonrural aging persons (Johnson, 1998). Whereas
technical interventions have demonstrated some success for
improving the physical health of older persons living in rural
areas (Buckwalter, Davis, Wakefield, Kienzle, & Murray,
2002), there is only preliminary evidence for the effectiveness
of telehealth strategies for improving psychological well being
(Montani et al., 1997; Sumner, 2001). The purpose of this study,
therefore, was to examine variations in depressive symptoms
and informal social support across geographic region and de-
velop predictive models of depressive symptoms with region as
a predictor of depression.
Methods
Sample
Participants in this study were drawn from the 2004 wave of
the Health and Retirement Study (HRS). HRS data were first
collected in 1992 from persons born between the years 1931
and 1941, and have been collected biannually since then. Addi-
tionally, since 1998, data from three additional cohorts were
also added: Aging and Health Dynamics (born prior to 1923),
Children of the Depression (1923-1930), and War Babies
(1942-1947). For the purposes of this study, we were inter-
ested in individuals at least 65 years old in 2004, and therefore
persons from the War Babies cohort were not included. How-
ever, older persons from all other cohorts participated.
We chose to only include data collected in 2004 because it
was in this wave of data collection that a so-called leave-behind
questionnaire was piloted with 4000 participants. This leave-
behind lifestyle questionnaire focused on psychosocial vari-
ables that dealt with attitudes towards life, family, friends, and
work. Overall, 76.8 percent of those selected to complete the
survey returned the questionnaire (Clarke, Fisher, House, Smith,
& Weir, 2007). In our analyses, only those respondents that
314

T. S. KILLIAN, M. PENFIELD
were selected to fill-out the psychosocial questionnaire, com-
pleted the questionnaire, and were at least 65 years old were
included. As a result, our analyses were based on data collected
from 1250 individual participants. Additionally, we also cor-
rected our estimates using a specific weight that was developed
by the HRS to adjust for selection and non-response to the
questionnaire because it was not administered to all partici-
pants.
As can be seen in Table 1, the mean age of participants was
about 73 years old, of which about 42 percent were male. Most
of the participants were White, and while most were married,
about 23 percent were widowed and 9 percent were unmarried.
The average number of children of participants was about 3.4.
Most participants were retired, although roughly 22 percent
were still working.
Measures
Depression. In order to assess depressive symptoms, we used
an abbreviated form of the CES-D available in the HRS. The
CES-D is a widely-used measure of depression that has demon-
strated high internal consistency and adequate discriminant
validity (Radloff, 1977). Research has indicated that both long
and short versions of the CES-D are reliable for usage in older
populations (Himmelfarb & Murrell, 1983; Kohout et al., 1993).
Due to interview time constraints, the HRS employed a short-
ened version of the CES-D that consisted of 8-items. Specifi-
cally, participants were asked whether or not they had experi-
enced 8 different indicators of depression over the past week
(coded so 1 = yes; 0 = no). These were summed and the result
was a single estimate of the number of depressive symptoms
participants had experienced in the previous week. Total scores
could range from 0 - 8 with higher scores indicating more
symptoms of depression. In a report on affective functioning
measures in the HRS, Steffick (2000) indicated that the 8-item
version of the CES-D has acceptable estimates of reliability
(Cronbach’s αs ranging from .78 - .81). Additionally, construct
validity was established by comparing the shortened version of
the CES-D to self-rated emotional health as well as examining
expected relations between the CES-D and known predictors of
depression. From these findings, Steffick concluded that the
CES-D is an acceptable measure of depressive symptoms. In
addition, this shortened version was acceptable for this study
because we were attempting to examine variation in depressive
symptoms, and not differentiating between those who were
clinically depressed or not.
Rurality. In this study, rural status was assigned on the basis
of the 10 category Beale Rural-Urban Continuum codes (Butler
& Beale, 1994) as provided by HRS. These codes are assigned
based on participants’ counties of residence at the time that they
participated in the study. HRS collapses the scale into three
variables exurban (codes 3 - 9), suburban (codes 1 and 2), and
urban (code 0). It should be noted that the label exurban used
by HRS reflects the collapsing of the codes in a way that does
not fully correspond to traditional definitions of rurality. How-
ever, given the limitations of the data set, it is the best ap-
proximation of rurality available in the non-restricted data.
Moreover, in this study, the scale was further collapsed so that
rural corresponded to Beale codes 3 - 9 and nonrural repre-
sented codes 0 - 2.
Informal social support. The operationalization of informal
social support was based on several questions about relation-
ships with friends, family, and children. In terms of examining
Table 1.
Weighted means and standard errors of variables by rural status (N =
1250).
Rural Nonrural Full sample
(N = 459) (N = 791) (N = 1250)
Variable M SE M SE M SE
CES-D1 1.37* .09 1.15 .06 1.23 .05
Age 73.04 .32 73.29 .23 73.2 .19
Male .38* .02 .45 .02 .42 .01
Black .02 .01 .04 .01 .03 .01
Other minority .02 .01 .02 .01 .02 0
Hispanic .02* .01 .04 .01 .03 .01
Unmarried .07* .01 .11 .01 .09 .01
Widowed .26 .02 .22 .01 .23 .01
Number of kids 3.43 .1 3.31 .08 3.36 .06
Income/10,000 5.34 .38 6.27 .38 5.93 .28
Assets/10,000 40.09 4.37 51.01 4.67 46.99 3.37
Working .25* .02 .2 .01 .22 .01
Mobility difficulty
index2 2.69** .12 2.25 .09 2.41 .07
Number of close friends6.97*** .45 5.37 .2 5.96 .21
Rely on friends 3.1 .04 3.03 .03 3.06 .03
Meet with friends 3.18 .05 3.2 .04 3.19 .03
Number of close
family3 4.8 .21 4.37 .21 4.53 .15
Rely on family3 3.13* .05 3 .04 3.05 .03
Meet with family3 2.56* .06 2.38 .04 2.45 .04
Number of children
close with 2.76*** .08 2.4 .06 2.53 .05
Rely on children 3.58*** .04 3.38 .04 3.45 .03
Meet with children 3.03* .05 2.87 .04 2.93 .03
Child living nearby .59 .02 .54 .02 .56 .01
Note: Significance reported for t-tests between rural and nonrural means.
1CES-D = Center for Epidemiologic Studies Depression Scale. 2Number of
difficulties in performing mobility, strength, and fine motor skills. 3Family =
any family member not living in the household. *p < .05. **p < .01. ***p
< .001. ****p < .0001.
friendships, three different questions from HRS were used per-
taining to: 1) the number of close friends; 2) the extent to which
one can rely on friends; and 3) the frequency of meeting with
friends. Specifically, the number of close friends variable asked
respondents to indicate how many friends they would say they
have a close relationship with. The extent to which one can rely
on friends was assessed by asking respondents how much they
can rely on their friends if a serious problem occurs (1 = a lot to
4 = not at all). Responses were reverse-coded so that higher
responses indicated greater confidence in friends. Last, the

T. S. KILLIAN, M. PENFIELD
frequency of meeting with friends was evaluated through a
single question that asked how often respondents met up with
their friends. Responses were coded so that higher numbers
indicated a greater frequency of meeting with friends (1 = only
once or twice a year or never; 2 = every few months; 3 =
monthly; 4 = weekly). The same three questions were used to
evaluate family relationships. The only difference was that,
instead of friends, the questions asked about relationships with
family members that were not residing in the same household
with them at the time of the survey. The same questions were
also used to examine relationships with children plus an addi-
tional question which asked whether or not the respondent had
a child living nearby (i.e., within 10 miles). Responses were
coded so that 1 = child living nearby and 0 = no child living
nearby.
Physical well-being. In order to assess physical well-being,
the mobility, strength, and fine motor skills measure available
in the HRS was used. This measure asked participants about
their abilities to perform 12 different activities (including things
like walking a block, climbing stairs, stooping, picking up a
dime, and others). The activities are more challenging than
ADLs and having difficulties in these activities are thought of
as precursors to limitations in ADLs (Fonda & Herzog, 2004).
Respondents were asked if difficulty was present in performing
the given activity, and responses were coded using binary codes
(1 = yes; 0 = no). After being coded, the responses were
summed making the variable equal to the number of difficulties
(0 - 12) in mobility, strength, and fine motor skills. Research
has indicated that this measure is a valid measure of physical
functioning and has good internal consistency (α = .87) (Fonda
& Herzog).
Other predictors. In addition to the variables above, several
other predictor variables—age, gender, race, ethnicity, marital
status, and socioeconomic status—were included in our analy-
ses because previous research has indicated that they are related
to levels of depression. Age was calculated by subtracting par-
ticipants’ years of birth from 2004, the year data was collected.
Gender was included by coding male participants as 1 and fe-
male participants as 0. For race, Black was coded as 1 if par-
ticipants self-identified as Black, otherwise 0. Although there
are several categories of race that participants may identify with
in the HRS, there were too few participants in any single of
these categories to be included in multivariate analyses. As a
result, we collapsed those persons into an other minority cate-
gory coded as 1 for participants who identified with a minority
group other than Black, and otherwise coded as 0. In terms of
ethnicity, participants who identified as Hispanic were coded as
1 and all other participants were coded as 0. In regard to marital
status, participants who were divorced, separated, or never-
married were coded as 1 for unmarried. Widowed participants
were coded as 1, otherwise 0. The implied comparison category
was married. Three indicators were used to measure socioeco-
nomic status. Estimates of annual income and net worth were
two of these indicators and are based on imputed HRS data.
These data are calculated and imputed based on a large number
of questions asked in the HRS survey. A third indicator of so-
cioeconomic status, working, was binary and coded as 1 for
participants who indicated that they were working at a full-time
job and 0 for all other participants.
Approach. Given our first research question was to examine
potential differences in depressive symptoms between rural and
nonrural persons, the first step in data analyses was a simple
t-test that examined whether or not the mean of depressive
symptoms varied between rural and nonrural persons. Next,
predictive models of depressive symptoms were developed to
more fully examine our second and third research questions. To
examine our second question, a linear regression model was
used to predict depressive symptoms while controlling for other
known predictors of depression that included age, gender, race,
and socioeconomic status. An additional regression model was
developed that added interaction terms to assess how rurality
and social support were related to depressive symptoms. These
variables were created by taking the product of rurality and
each social support indicator.
Results
As can be seen in Table 1, rural participants experienced sig-
nificantly more depressive symptoms (M = 1.37) than their
nonrural counterparts (M = 1.23), although the difference was
small (t = −2.09, p < .05). This result gave some, albeit weak,
support for the hypothesis that rural persons experience more
depressive symptoms than nonrural persons. It may be, how-
ever, that regardless of little difference in the number of de-
pressive symptoms, the experience of those symptoms and the
context in which those symptoms are coped with vary across
rurality.
As expected of a sample of older persons, participants were
more likely to be female than male, but rural participants were
more likely to be female than nonrural participants (t = 2.48, p
< .05). Rural participants were less likely than nonrural partici-
pants to be Hispanic (t = 2.20, p < .05), less likely to be unmar-
ried (t = 2.20, p < .05), and more likely to be working (t =
−1.96, p < .05). In addition, rural participants, as compared to
others, were more likely to experience limitations in mobility,
strength and fine motor skills (t = −3.02, p < .01) and have
more close friends (t = 6.97, p < .001). In spite of having more
close friends, there was no difference in the perceived ability to
rely on friends by rural status (t = −1.42, p = n.s.).
There was no difference by rurality in the number of family
members with whom participants had close relationships (t =
–1.37, p = n.s.), but rural participants were slightly more likely
than others to perceive that they could rely on family members
(t = –2.11, p < .05). Rural persons were more likely to meet
with family members than their nonrural counterparts (t =
−2.51, p < .05). Rural participants were substantively more
likely than nonrural participants to have close relationships with
adult children (t = −3.74, p < .001), perceive that they could
rely on adult children (t = −3.62, p < .001), and meet with adult
children (t = −2.35, p < .05). There was no difference by rural-
ity in the likelihood of living close to at least one adult child (t
= −1.84, p = n.s.).
Two regression models were constructed. The first model in-
cluded only sociodemographic variables and estimates of limi-
tations in mobility, strength, and fine motor skills. As can be
seen in Table 2, there was a small negative relationship be-
tween age and depressive symptoms, Black participants ex-
perienced more depressive symptoms than other participants,
widowed participants experienced more depressive symptoms
than married participants, and limitations in mobility, strength,
and fine motor skills were positively related to depressive
symptoms. None of the other variables included in the regres-
sion model were significant. As a whole, this model predicted
19 percent of the variance in depressive symptoms.
The second regression model included the same variables as
316

T. S. KILLIAN, M. PENFIELD
Table 2.
Hierarchical regression predicting depression (N = 1250).
Model 1 Model 2
Variable B  B 
Sociodemographic
Age –.02* –.07 –.02* –.06
Male .18 .05 .08 .02
Black .88** .09 .94*** .09
Other minority .44 .04 .29 .02
Hispanic .24 .02 .08 .01
Unmarried .29 .05 .21 .03
Widowed .61**** .14 .64**** .15
Number of children .02 .03 .06* .08
Income/10,000 0 –.01 0 0
Assets/10,000 0 –.02 0 –.01
Working –.24* –.06 –.26* –.06
Mobility difficulty index1 .28**** .38 .23**** .31
Living in a rural area .12 .03 1.1 .29
Informal social support
Number of close friends - - –.01 –.02
Rely on friends - - –.18* –.09
Meet with friends - - –.15* –.08
Number of close family2 - - 0 –.01
Rely on family2 - - –.04 –.02
Meet with family2 - - .1 .07
Number of close children - - –.11* –.1
Rely on children - - .09 .05
Meet with children - - –.1 –.07
Child living nearby - - .02 .01
Rural by social support variables
Rural x number of close friends - - .01 .04
Rural x rely on friends - - –.1 –.09
Rural x meet with friends - - .03 .03
Rural x number of close family - - .01 .02
Rural x rely on family - - –.02 –.02
Rural x meet with family - - .02 .02
Rural x number of close children - - .03 .03
Rural x rely on children - - –.34* –.34
Rural x meet with children - - .03 .03
Rural x child living nearby - - .02 .01
Rural x mobility difficulty index1 - - .08* .09
Intercept 1.68 0 2.65 0
R-square .19 .24
Note: Regression analyses were conducted using weighted data. 1Number of difficulties in performing mobility, strength, and fine motor skills.
2Family = any family member not living in the household. *p < .05. **p < .01. ***p < .001. ****p < .0001.

T. S. KILLIAN, M. PENFIELD
included in the first. However, additional variables were added
to examine how participants’ social contexts were related to
variation in depressive symptoms, as well as how those social
contexts interacted with rurality.
In regard to relationships with friends, this study found that
the number of friends with whom participants had close rela-
tionships was not related to depressive symptoms. However,
the ability of participants to rely on friends and the frequency of
meeting with friends were both related to lower levels of de-
pressive symptoms. Relationships with family members were
not significantly related to depressive symptoms with the single
exception of the number of children with whom participants
had close relationships.
Of the interactions with rurality, only two variables were
significantly related to variation in depressive symptoms. First,
rural participants who perceived that they could rely on adult
children were likely to experience fewer depressive symptoms
than nonrural participants who perceived that they could rely on
adult children. Second, limitations in mobility, strength, and
fine motor skills were related to higher levels of depressive
symptoms for rural, as compared to nonrural, participants. As a
whole, the additional variables in the second model increased
the amount of variance explained in depressive symptoms to 24
percent.
Disscussion
The purpose of this study was to examine variation in de-
pressive symptoms among older persons focusing on how that
variation may be related to rurality and informal social support.
Three central research questions were examined. Are there
variations in depressive symptoms by rurality? Are variations
in depressive symptoms related to informal social support? And
are there interactions between rurality and informal social sup-
port in regard to predicting depressive symptoms?
Regarding the first research question, results from this study
found a small, but significant difference in depressive symp-
toms by rurality. Although small, this significant difference is
likely substantive to older persons themselves, as well as the
rural communities that they live in. As noted by Callahan and
colleagues (1997), depressive symptoms are related to a greater
likelihood of the use of emergency departments, hospitaliz-
tions, and diagnostic tests. These are likely to increase the
strains placed on rural health agencies that are already con-
strained by rurality (Rogers, 2002b). There is also some evi-
dence that, in rural areas, these risks may not be alleviated by
mental health services.
Mental health services can be important resources as people
cope with depressive symptoms in their lives. Again, however,
this may place rural persons at increased risk in comparison to
nonruralpersons. As noted by Rathbone McCuan and Bane
(2003), there are significant barriers to accessing mental health
services in rural areas. These included limited resources and
geographic distance to services, as well as fear of stigmatization
and negative attitudes toward mental health services. Although
mental health services may be important resources to help older
persons cope with depressive symptoms, this research suggests
that there may be significant barriers for rural persons accessing
those services. However, one limitation of this study is that
there were no direct measures of the availability of mental
health services or participants’ attitudes toward those services,
an indication that more research is needed.
Given those barriers, the focus of this study was on interper-
sonal relationships with friends and family members and how
those relationships interacted with rurality to predict depressive
symptoms. Again, this study found significant, but small, rela-
tionships between interpersonal relationships and rurality.
These findings are not surprising and well established by pre-
vious studies. For example, Oxman, Berkman, Kasl, Freeman,
and Barret (1992) found that functional disability and changes
in social networks were the largest contributors to variance
explained in depressive symptoms, even after controlling for
baseline depression levels. In particular, this study found that
children making weekly visits to participants were predictive of
lower levels of depressive symptoms. It was not surprising that
relationships with friends were related to lower levels of de-
pressive symptoms.
Whereas few interactions with rurality were significant, the
two significant variables are indicative of an important experi-
ence in the ecology of rural aging. Specifically, this study found
that participants’ perceptions that they could rely on their adult
children to meet their needs was more salient in regard to re-
ducing depressive symptoms for rural than nonrural participants.
It is not clear why. However, it may be that rural persons are
relying on the informal assistance of adult children to provide
them with assistance that nonrural persons may be able to seek
from formal health care providers.
This interpretation is complemented by the finding that in-
creased limitations in mobility, strength, and fine motor skills
were more influential in explaining depressive symptoms in
rural than urban persons. Again, one interpretation of this find-
ing is that barriers associated with accessing formal healthcare
services for rural persons as compared to nonrural persons re-
sults in a disadvantage in regard to coping with physical health
issues and limitations in functioning.
In the end, this study suggests that rural older persons may
face barriers to accessing rural health services. Therefore, it is
not surprising that depressive symptoms are exacerbated by
physical limitations for rural persons as compared to others. It
also follows that, in this context, the importance of being able
to rely on adult children is more important to the psychological
well-being of older rural, than nonrural, persons.
It is important to note several limitations of the study. First,
the effect sizes reported in this study were significant, but small.
Consequently, only 24 percent of the variance was explained in
the full model. Nevertheless, even small differences can be
substantive and this study does provide important evidence for
guiding practice and future research.
A second limitation of this study is that the data do not allow
for a direct examination of participants’ access or lack of access
to formal health care services. Instead, this is inferred based on
previous studies of rural aging and health (Rathbone-McCuan
& Bane, 2003). More research is needed to clearly explicate the
hypothesized link between rurality, access to formal health care
services, and depressive symptoms.
In spite of those limitations, this study does point to some
important implications. First, this study suggests that intergen-
erational relationships are likely more salient for rural persons
as compared to nonrural persons. That is, rural persons may be
more likely to rely on adult children to meet their needs than
nonrural persons. It is important, therefore, that health care
providers and social workers develop an understanding of the
availability and willingness of adult children to provide assis-
tance to older persons. The quality of intergenerational rela-
318

T. S. KILLIAN, M. PENFIELD
tionships, in general, have been found to be key for informal
caregiving (Ganong & Coleman, 1999; Killian & Ganong,
2002), and this study suggests that those relationships may be
more important for rural than nonrural persons.
REFERENCES
Bierman, A., Fazio, E. M., & Milkie, M. A. (2006). A multifaceted
approach to the mental health advantage of the married: Assessing
how explanations vary by outcome measure and unmarried group.
Journal of Family Issues, 27, 554-582.
doi:10.1177/0192513X05284111
Braam, A. W., Prince, M. J., Beekman, A. T. F., Delespaul, P., Dewey,
M. E., Geerlings, S. W. et al. (2005). Physical health and depressive
symptoms in older Europeans. British Journal of Psychiatry, 187,
35-42. doi:10.1192/bjp.187.1.35
Buckwalter, K. C., Davis, L. L., Wakefield, B. J., Kienzle, M. G., &
Murray, M. A. (2002). Telehealth for elders and their caregivers in
rural communities. Family and C o m m u n i t y Health, 25, 31-40.
doi:10.1097/00003727-200210000-00007
Butler, M. A., & Beale, C. L. (1994). Rural-urban continuum codes for
metro and non-metro counties, 1993 (Staff Report No. 9425). Wash-
ington DC: US Department of Agriculture, Agriculture and Rural
Economy Division, Economic Research Service.
Callahan, C., Kesterson, J., & Tierney, W. (1997). Association of
symptoms of depression with diagnostic test charges among older
adults. Annals of Interna l Medicine, 126, 426-433.
Clarke, P., Fisher, G., House, J., Smith, J., & Weir, D. (2007). Guide to
content of the HRS psychosocial leave-behind participant lifestyle
questionnaires: 2004 & 2006.
http://hrsonline.isr.umich.edu/docs/userg/HRS2006LBQscale.pdf
Djernes, J. K. (2006). Prevalence and predictors of depression in popu-
lations of elderly: A review. Acta Psychiatrica Scandinavica, 113,
372-387. doi:10.1111/j.1600-0447.2006.00770.x
Engedal, K. (1996). Mortality in the elderly—A 3-year follow-up of an
elderly community sample. International Journal of Geriatric Psy-
chiatry, 11, 467-471.
doi:10.1002/(SICI)1099-1166(199605)11:5<467::AID-GPS338>3.0.
CO;2-S
Fonda, S., & Herzog, A. R. (2004). Documentation of physical func-
tioning measured in the Health and Retirement Study and Asset and
Health Dynamics among the Oldest Old Study.
http://hrsonline.isr.umich.edu/docs/userg/dr-008.pdf
Ganong, L. H., & Coleman, M. (1999). Changing families, changing
responsibilities: Family obligations following divorce and remar-
riage. Mahwah: Lawrence Erlbaum Associates.
Gatz, M., & Smyer, M. A. (1992). The mental health system and older
adults in the 1990s. American Psychologist, 47, 741-751.
doi:10.1037/0003-066X.47.6.741
Himmelfarb, S. & Murrell, S. A. (1983). Reliability and validity of five
mental health scales in older persons. Journal of Gerontology, 38,
333-339. doi:10.1093/geronj/38.3.333
House, J. S., Landis, K. R., & Umberson, D. (1988). Social relation-
ships and health. Science, 241, 540-545.
doi:10.1126/science.3399889
Johnson, J. E. (1998). Stress, social support, and health in frontier eld-
ers. Journal of Gerontological Nursing, 24, 29-35.
Katon, W. J., Lin, E., Russo, J., & Unutzer, J. (2003). Increased medi-
cal costs of a population-based sample of depressed elderly patients.
Archives of General Psychiatry, 60, 897-903.
doi:10.1001/archpsyc.60.9.897
Killian, T. S., & Ganong, L. H. (2002). Ideology, context, and obliga-
tions to assist older persons. Journal of Marriage and the Family, 64,
1080-1088. doi:10.1111/j.1741-3737.2002.01080.x
Killian, T.S., Turner, J., & Cain, R. (2005). Depressive symptoms of
caregiving women in midlife: The role of physical health. Journal of
Women & Aging, 17, 115-127. doi:10.1300/J074v17n01_09
Kohout, F. J., Berkman, L. F., Evans, D. A., & Cornoni-Huntley, J.
(1993). Two shorter forms of the CES-D depression symptoms index.
Journal of Aging and Health, 5, 179-193.
doi:10.1177/089826439300500202
Langa, K. M., Valenstein, M. A., Fendrick, A. M., Kabeto, M. U., &
Vijan, S. (2004). Extent and cost of informal caregiving for older
Americans with symptoms of depression. The American Journal of
Psychiatry, 161, 857-863. doi:10.1176/appi.ajp.161.5.857
Montani, L, Billaud, N., Tyrrell, J., Fluchaire, I., Malterre, C., Lauver-
nay, N., Couturier, P., & Franco, A. (1997). Psychological impact of
a remote psychometric consultation with hospitalized elderly people.
Journal of Telemedicine and Telecare, 3, 140-145.
doi:10.1258/1357633971931048
Oxman, T. E., Berkman, L. F., Kasl, S., Freeman, D. H., & Barret, J.
(1992). Social support and depressive symptoms in the elderly.
American Journal of Epide m io l o g y , 135, 356-368.
Penninx, B. W. J. H., Guralnik, J. M., Ferrucci, L., Simonsick, E. M.,
Deeg, D. J. H., & Wallace, R. B. (1998). Depressive symptoms and
physical decline in community-dwelling older persons. The Journal
of the American Medical Association, 279, 1720-1726.
doi:10.1001/jama.279.21.1720
Probst, J. C., Laditka, S. B., Moore, C. G., Harun, N., Poweel, P., &
Baxley, E. G. (2006). Rural-urban differences in depression preva-
lence: Implications for family medicine. Family Medicine, 38, 653-
660.
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale
for research in the general population. Applied Psychological Meas-
urement, 1, 385-401. doi:10.1177/014662167700100306
Rathbone-McCuan, E., & Bane, S. (2003). Rural mental health: A
discussion of service capacity building for rural elders. Journal of
Gerontological Soc i a l Work, 41, 301-312.
doi:10.1300/J083v41n03_07
Ried, L., & Planas, L. (2002). Aging, health, and depressive symptoms:
Are women and men different? Journal of Women’s Health, 11, 813-
824. doi:10.1089/15409990260430963
Rogers, C. C. (2002a). Rural health issues for the older population.
Rural America, 17 , 30-36.
Rogers, C. C. (2002b). The older population in 21st century rural
America. Rural America, 17, 1-10.
Schulz, R., & Beach, S. R. (1999). Caregiving as a risk for mortality:
The caregiver health effects study. Journal of the American Medical
Association, 282, 2215-2219. doi:10.1001/jama.282.23.2215
Steffick, D. E. (2000). Documentation of affective functioning measures
in the Health and Retirement Study.
http://hrsonline.isr.umich.edu/docs/userg/dr-005.pdf
Sumner, C. R. (2001). Telepsychiatry: Challenges in rural aging. The
Journal of Rural Healt h, 17, 370-373.
doi:10.1111/j.1748-0361.2001.tb00291.x
Wilson, R., Mendes de Leon, C., Bennett, D., Bienias, J., & Evans, D.
(2004). Depressive symptoms and cognitive decline in a community
population of older persons. Journal of Neurology, Neurosurgery
and Psychiatry, 75 , 126-129.
Yang, Y. (2006). How does functional disability affect depressive
symptoms in late life? The role of perceived social support and psy-
chological resources. Journal of Health and Social Behavior, 47,
355-372. doi:10.1177/002214650604700404
Yang, Y., & George, L. K. (2005). Functional disability, disability
transitions, and depressive symptoms in late life. Journal of Aging
and Health, 17, 263-292. doi:10.1177/0898264305276295
Zunzunegui, M. V., Beland, F., Llacer, A., & Leon, V. (1998). Gender
differences in depressive symptoms among Spanish elderly. Social
Psychiatry and Psychiatric Epidemiology, 33, 195-205.
doi:10.1007/s001270050043