Loneliness Among Older Veterans in the United States: Results from the National Health and Resilience in Veterans Study

Abstract

This study examined the current prevalence, and demographic, military, health, and psychosocial correlates of loneliness in a contemporary nationally representative sample of older U.S. veterans.
Two thousand twenty-five veterans aged 60 years and older participated in the National Health and Resilience in Veterans Study. Loneliness was assessed using a questionnaire adapted from the Revised UCLA Loneliness Scale. A broad range of demographic, military, health, and psychosocial variables was also assessed.
44% of veterans reported feeling lonely at least some of the time (10.4% reported often feeling lonely). Greater age, disability in activities of daily living, lifetime traumas, perceived stress, and current depressive and post-traumatic stress disorder symptoms were positively associated with loneliness, and being married/cohabitating, higher income, greater subjective cognitive functioning, social support, secure attachment, dispositional gratitude, and frequency of attending religious services were negatively associated with loneliness. The largest magnitude associations were observed for perceived social support, secure attachment style, and depressive symptoms.
Loneliness is prevalent among older veterans in the United States, and associated with several health and psychosocial variables. These results suggest that multifactorial interventions that emphasize bolstering of social support and reduction of depressive symptoms may help mitigate loneliness in the rapidly growing population of older veterans.

Full text

Loneliness Among Older
Veterans in the United
States: Results from
the National Health
and Resilience in
Veterans Study
Philipp Kuwert, M.D.,
Christine Knaevelsrud, Ph.D.,
Robert H. Pietrzak, Ph.D., M.P.H.
Objectives: This study examined the current preva-
lence, and demographic, military, health, and psycho-
social correlates of loneliness in a contemporary
nationally representative sample of older U.S. veterans.
Methods: Two thousand twenty-five veterans aged 60
years and older participated in the National Health
and Resilience in Veterans Study. Loneliness was
assessed using a questionnaire adapted from the
Revised UCLA Loneliness Scale. A broad range of
demographic, military, health, and psychosocial vari-
ables was also assessed. Results: 44% of veterans re-
ported feeling lonely at least some of the time (10.4%
reported often feeling lonely). Greater age, disability in
activities of daily living, lifetime traumas, perceived
stress, and current depressive and post-traumatic stress
disorder symptoms were positively associated with
loneliness, and being married/cohabitating, higher
income, greater subjective cognitive functioning,
social support, secure attachment, dispositional grati-
tude, and frequency of attending religious services
were negatively associated with loneliness. The largest
magnitude associations were observed for perceived
social support, secure attachment style, and depressive
symptoms. Conclusions: Loneliness is prevalent
among older veterans in the United States, and asso-
ciated with several health and psychosocial variables.
These results suggest that multifactorial interventions
that emphasize bolstering of social support and
reduction of depressive symptoms may help mitigate
loneliness in the rapidly growing population of older
veterans. (Am J Geriatr Psychiatry 2013; -:-e-)
Key Words: Loneliness, veterans, old age, PTSD, war,
trauma, correlates
Leading a socially connected life is a core aspect of
human nature.
1
Some individuals, however, may
experience difficulties in forming meaningful social
relations, and others may lose them during the course
of their lives through divorce/separation, widow-
hood, or death of relatives and close friends. These
individuals are consequently at greater risk for lone-
liness, which has been defined as perceived social
isolation
2
or a discrepancy between a person’s desired
and actual social relationships.
3
Feeling lonely is a common experience in older
persons, with 40% of adults aged 65 years or older
reporting feeling lonely at least sometimes, with this
proportion increasing with age.
2,4,5
Despite the high
prevalence of loneliness in the elderly population,
contemporary geriatric mental health research has
typically focused on characterizing the prevalence
and determinants of major psychiatric disorders, and
has only recent begun to explore factors associated
with broader, more pervasive psychosocial issues
such as loneliness.
4
Several health and psychosocial variables associated
with loneliness have been identified.
2,4
These include
risk variables such as greater medical burden, depres-
sion, and cognitive dysfunction; as well as putative
protective variables, such as greater social support. A
number of other positive psychosocial factors, such as
optimism, resilience, and religiosity, which have been
Received June 28, 2012; revised December 7, 2012; accepted February 15, 2013. From the Ernst-Moritz-Arndt-University Greifswald,
Department of Psychiatry and Psychotherapy at the HELIOS Hansehospital Stralsund (PK), Germany; Treatment Centre for Torture Victims,
and Free University Berlin, Department of Clinical Psychology and Psychotherapy (CK), Berlin, Germany; U.S. Department of Veterans Affairs
National Center for Posttraumatic Stress Disorder, VA Connecticut Healthcare System, and Department of Psychiatry, Yale University, School of
Medicine (RHP), New Haven, CT. Send correspondence and reprint requests to Philipp Kuwert, M.D., Department of Psychiatry and Psycho-
therapy, Ernst-Moritz-Arndt University Greifswald, Rostocker Chaussee 70, D-18437 Stralsund, Germany. e-mail: kuwert@uni-greifswald.de
Supplemental digital content is available for this article in the HTML and PDF versions of this article on the journal’s Web site (www.
ajgponline.org).
Ó2013 American Association for Geriatric Psychiatry
http://dx.doi.org/10.1016/j.jagp.2013.02.013
Am J Geriatr Psychiatry -:-,-2013 1

identified as protective against psychopathology,
6
may also be negatively related to loneliness, although
no study of which we are aware has examined this
possibility. Characterization of potentially modifiable
factors associated with loneliness may help identify
targets for preventive interventions, which may ulti-
mately help reduce the deleterious effect of loneliness
on morbidity and mortality in older persons.
1,4,7
Given their increased risk for psychiatric disorders
and negative behavioral outcomes,
8
older veterans
may be at increased risk for loneliness in later life.
Although data are lacking on the prevalence and
correlates of loneliness in older veterans, available
research has underscored the importance of loneliness
as a predictor of health-related outcomes in this pop-
ulation. For example, one study found that loneliness
is one of the most prominent factors associated with
the reactivation of combat-related post-traumatic
stress disorder (PTSD) in older combat veterans.
9
Another study found that loneliness mediated the
relationship between PTSD and marital adjustment in
elderly combat veterans and former prisoners of war.
10
Although these studies offer some insight into the role
of loneliness in older veterans, they are based on small
and highly selective samples. Research on population-
based samples of older veterans is needed to evaluate
more basic questions regarding the prevalence, and
demographic, military, health, and psychosocial
correlates of loneliness in this population. Character-
ization of factors associated with loneliness may help
identify targets for prevention efforts that may help
mitigate risk for more severe health-related outcomes.
The purpose of the current study was to examine the
prevalence and correlates of loneliness in a large,
contemporary, nationally representative sample of
older U.S. veterans. Correlates examined in relation to
loneliness included those identified as being differen-
tially related to this outcome in prior research,
2,4
as
well as additional factors that have been found to be
protective against psychopathology, such as opti-
mism, resilience, and religiosity.
6
METHODS
Sample
Data were drawn from the National Health and
Resilience in Veterans Study (NHRVS), a nationally
representative study of 3,157 U.S. veterans that was
conducted in October to December 2011. Participants
in the NHRVS completed an anonymous 60-minute
Web survey. The majority of this sample (N ¼2,025;
64.1%) consisted of veterans aged 60 years or older
(mean age: 71.0 years; SD: 7.1; range: 60e96); data
from these respondents were analyzed in this study.
The NHRVS sample was drawn from a research panel
of more than 80,000 households that is developed and
maintained by Knowledge Networks, Inc., a survey
research firm based in Menlo Park, CA. Knowledge
Networks maintains KnowledgePanel, a probability-
based, online non-volunteer access survey panel of
a nationally representative sample of U.S. adults that
covers approximately 98% of U.S. households,
including cell-phoneeonly households. Knowledge
Networks, who first began recruiting survey panelists
in 1999, established the first online research panel
based on probability sampling that covered both
online and offline populations in the United States.
Panel members are recruited through national
random samples, originally by telephone and now
almost entirely by postal mail. Households are
provided with computer hardware and Internet access
if needed. Unlike Internet convenience panels, also
known as “opt-in”panels, that include only individ-
uals with Internet access who volunteer themselves
for research, KnowledgePanel recruitment uses dual
sampling frames that includes both listed and unlisted
telephone numbers, telephone and non-telephone
households, cell-phoneeonly households, and house-
holds with and without Internet access. Only persons
sampled through these probability-based techniques
are eligible to participate on KnowledgePanel. Unless
invited to do so as part of these national samples, no
one on their own can volunteer to be part of the panel.
A total of 3,188 individuals in the Knowledge
Networks panel answered “Yes”to an initial screening
question that confirmed veteran status: “Have you
ever served on active duty in the U.S. Armed Forces,
Military Reserves, or National Guard?”Of these, 3,157
(99.0%) completed the survey. To permit generaliz-
ability of study results to the entire population of U.S.
veterans, post-stratification weights were applied
based on demographic distributions (i.e., age, gender,
race/ethnicity, education, Census region, and metro-
politan area) from the most contemporaneous
(October 2010) Current Population Survey (U.S.
Census Bureau, 2010).
2 Am J Geriatr Psychiatry -:-,-2013
Loneliness Among Older Veterans in the U.S.

Demographic characteristics of the NHRVS sample
were highly consistent with those observed in prior
population-based surveys of veterans, such as the U.S.
Census Bureau American Community Survey (ACS
11
).
For example, the median age of male veterans in the
NHRVS sample versus the ACS sample were both 64
years; the median age of female veterans was 50 versus
49 years. The proportion of white, non-Hispanic male
veterans was 77.4% versus 82.0%; proportion of white,
non-Hispanic female veterans was 64.6% versus 69.6%.
The proportion of male veterans who completed some
college or higher education was 65.5% versus 60.3%;
proportion of female veterans who completed some
college or higher education was 79.0% versus 77.4%.
Finally, the proportion of male veterans who were
married was 72.3% versus 67.7%; the proportion of
female veterans who were married was 54.7% versus
47.4%. All participants provided informed consent and
the study was approved by the institutional review
board of the VA Connecticut Healthcare System.
Measures
Veterans completed a comprehensive survey con-
taining measures of demographics, military charac-
teristics, and health and psychosocial variables. The
main outcome measure in the current study was
a three-item loneliness questionnaire adapted from the
Revised UCLA Loneliness Scale.
3
This measure
assesses three components of loneliness, including the
extent to which an individual feels left out, isolated, or
that he or she lacks companionship.
12
Each item is
scored as “1”(hardly ever); “2”(some of the time); or
“3”(often). These items are summed to yield
a summary loneliness measure (Cronbach’s
a
in the
current sample: 0.84). Variables examined in relation to
loneliness are shown in Table 1 and described in detail
in the Supplemental Digital Content (available online).
Data Analysis
Distributions of all variables were examined to
ensure suitability for parametric analysis; non-
normally distributed data (e.g., psychiatric symptom
variables) were transformed using logarithmic base 10
transformations. There were minimal missing item-
level data (<5% across variables); the expectation-
maximization algorithm was used to impute these
missing data prior to analysis. Pearson correlations
were computed to examine associations between
explanatory variables and scores on the summary
measure of loneliness. Explanatory variables associ-
ated with summary loneliness scores at the p less than
0.05 level in bivariate analyses were then entered into
a stepwise linear regression analysis with backward
elimination to identify variables independently asso-
ciated with these scores.
RESULTS
The mean age in the sample was 71.0 years (SD:
7.1, range: 60e96); the majority was male (96.2%);
white, non-Hispanic (84.7%); completed high school
or higher level of education (95.3%); were married/
cohabitating (77.7%) and retired (80.3%); resided in
a metropolitan area (80.8%); and had a household
income less than $60,000/year (60.4%). The majority
of the sample (78.7%) enlisted in the military; served
in the Army (41.2%), Navy (24.9%), or Air Force
(23.9%); and 37.9% were combat veterans.
A total of 929 (weighted percentage: 44.0%)
veterans endorsed a response of “some of the time”or
“often”for one or more of the questions on the three-
item loneliness questionnaire that assessed current
frequency of feeling left out, isolated, or lacking
companionship; 210 (10.4%) endorsed a response of
“often”for one or more of these questions.
Results of bivariate correlations and linear regres-
sion analysis are shown in Table 1. In the regression
model, 50.9% of the variance in loneliness scores was
explained. Greater age, activities of daily living (ADL)
disability, number of lifetime traumatic events,
perceived stress, and current depressive and PTSD
symptoms were positively associated with loneliness
scores, whereas being married/living with partner,
higher income, greater subjective cognitive func-
tioning, greater number of close friends, greater
perceived social support, secure attachment, disposi-
tional gratitude, and frequency of attending religious
services. The largest magnitude associations were
observed for perceived social support, secure attach-
ment style, and depressive symptoms.
DISCUSSION
Results of this study suggest that loneliness is
prevalent, and associated with several health and
Am J Geriatr Psychiatry -:-,-2013 3
Kuwert et al.

psychosocial variables among older veterans in the
United States. Consistent with previous work in
nonveteran samples,
1,2,4
a total of 44.0% of this sample
reported feeling lonely at least sometimes, and 10.4%
reported feeling lonely often. Further, greater age,
ADL disability, and depressive symptoms were posi-
tively associated with loneliness, whereas greater
subjective cognitive functioning and perceived social
support was negatively associated with this outcome.
Results further suggested that, after adjustment for
these known correlates of loneliness, a greater number
of traumatic events, perceived stress, and PTSD
symptoms were positively associated with loneliness,
whereas secure attachment, dispositional gratitude,
and greater frequency of engagement in religious
services were negatively associated with this outcome.
TABLE 1. Results of Bivariate Correlation and Multivariate Regression Analysis Examining Variables Associated with Loneliness in
Older Veterans (N [2,025)
Scale Used to
Measure Construct
Bivariate Analyses Regression Analysis
r
b
tp
Demographic and military variables
Age e0.07** 0.07*** 4.18 <0.001
Male sex e0.03 eee
White, non-Hispanic ethnicity e0.04 eee
High school or higher education e0.01 eee
Married/living with partner e0.28*** 0.11*** 6.09 <0.001
Currently working e0.01 eee
Household income >$60K e0.10*** 0.04* 2.47 0.013
Metropolitan residence e0.03 eee
Enlisted into military e0.01 eee
Combat veteran e0.05* eee
Medical and cognitive variables
Number of medical conditions e0.15*** eee
Current somatic symptoms BSI-18-Somatization 0.28*** eee
Any ADL disability e0.08** 0.05** 2.90 0.004
Any IADL disability e0.21*** 0.02 0.73 0.47
Cognitive functioning MOS-CFS-R 0.38*** 0.08** 3.04 0.002
Psychiatric variables
Number of lifetime traumatic events THS 0.28*** 0.07*** 3.70 <0.001
Perceived stress PSS-4 0.42*** 0.12*** 5.72 <0.001
Current depressive symptoms PHQ-4 0.49*** 0.18*** 7.99 <0.001
Current PTSD symptoms PCL 0.41*** 0.09* 4.01 <0.001
Current GAD symptoms PHQ-4 0.40*** eee
Alcohol consumption severity AUDIT-C 0.01 eee
Protective psychosocial variables
Resilience CD-RISC-10 0.29*** eee
Optimism LOT-R 0.28*** eee
Number of close friends e0.23*** 0.04* 2.16 0.030
Perceived social support MOS-SSS-5 0.55*** 0.26*** 12.25 <0.001
Secure attachment e0.48*** 0.19*** 9.39 <0.001
Curiosity/Exploration CEI-II 0.21*** eee
Dispositional gratitude GQ-6 0.36*** 0.07** 3.44 0.001
Community Integration e0.34*** eee
Purpose in life PIL-SF 0.38*** eee
Frequency of attending religious activities DUREL 0.19*** 0.04* 2.30 0.021
Frequency of engagement in private spiritual activities DUREL 0.10*** eee
Intrinsic religiosity DUREL 0.13*** eee
Notes: ADL: activity of daily living; IADL: instrumental activity of daily living; PTSD: posttraumatic stress disorder; GAD: generalized
anxiety disorder.
BSI-18-Somatization: Brief Symptom Inventory-18 Somatization subscale; MOS-CFS-R: Medical Outcomes Survey Cognitive Functioning
Scale-Revised; THS: Trauma History Screen; PSS-4: Perceived Stress Scale-4; PHQ-4: Patient Health Questionnaire-4; PCL: PTSD Checklist;
AUDIT-C: Alcohol Use Disorders Identification Test-Consumption; CD-RISC-10: Connor-Davidson Resilience Scale-10; LOT-R: Life Orien-
tation Test-Revised; MOS-SSS-5: Medical Outcomes Survey Social Support Survey-5; CEI-II: Curiosity and Exploration Inventory-II; GQ-6:
Gratitude Questionnaire-6; PIL-SF: Purpose in Life Test-Short Form; DUREL: Duke University Religion Index. Details regarding covariates,
including how they were operationalized and scored, are provided in Supplemental Digital Content (available online). Degrees of freedom
(df) for Pearson r’s: 1,809; df for t tests in regression model: 1,708.
Significant association with scores on summary measure of loneliness: *p <0.05; **p <0.01; ***p <0.001.
4 Am J Geriatr Psychiatry -:-,-2013
Loneliness Among Older Veterans in the U.S.

Although the directionality of the observed asso-
ciations cannot be ascertained on the basis of cross-
sectional data, these results contribute to a growing
body of research concerning the role of putative
protective factors on psychosocial problems such as
loneliness in older persons.
6
The largest magnitude
associations with loneliness were observed for
perceived social support (
b
¼0.26), secure attach-
ment (
b
¼0.19), and depressive symptoms (
b
¼
0.18). Older veterans who have a secure attachment
style may be better able to form lasting relationships
with others and consequently benefit from greater
social support. Depressive symptoms, which have
been reciprocally linked to loneliness in longitudinal
studies, may act synergistically to affect emotional
well-being.
13,14
Dispositional gratitude and engage-
ment in religious services were additionally nega-
tively related to loneliness, even after adjustment
for perceived social support, attachment style, and
depressive symptoms. Greater dispositional grati-
tude may be associated with greater ability to
appreciate one’s relationships and other positive
aspects of one’s life, whereas greater engagement
in religious services may provide veterans with
a broader social network, which may in turn help
prevent loneliness. That a broad range of health
and psychosocial factors were associated with
loneliness underscores the importance of multifacto-
rial approaches to the prevention and treatment of
loneliness in older veterans. Loneliness may arise in
part from a combination of intrinsic traits, such as
dispositional gratitude, extrinsic factors such as social
support and engagement in religious services, and
factors that may combine both such as attachment
style. Longitudinal studies will be useful in teasing
apart directional associations among these variables,
and in identifying optimal strategies for the preven-
tion and treatment of loneliness.
Methodological limitations of this study must be
noted. First, all assessments were obtained via self-
report, so it is unclear if results would differ if
clinician interviews were conducted. Second, recall
bias may have negatively affected responses to
certain questions (e.g., recollection of past traumatic
events). Third, use of a cross-sectional study design
precludes an understanding of the temporal or causal
associations between health and psychosocial vari-
ables associated with loneliness. Fourth, independent
associations between some explanatory variables
such as depressive symptoms and loneliness may, at
least in part, be related to their assessing overlapping
constructs. Nevertheless, prior research has found
that depressive symptoms and loneliness are distinct
constructs; for example, although lonely persons
often present with depression, depressed persons do
not consistently report feeling lonely. Further, prior
longitudinal work has revealed that loneliness and
depressive symptoms are reciprocally related over
time.
13,14
Notwithstanding these limitations, this study is
the first of which we are aware to characterize the
current prevalence and health and psychosocial
correlates of loneliness in a contemporary, nationally
representative sample of U.S. older veterans. Results
suggest that two out of five older veterans in the
United States experiences at least some loneliness,
and that in addition to well-known risk factors for
loneliness such as physical and mental health diffi-
culties, greater perceived social support, secure
attachment, gratitude, and engagement in religious
services are inversely related with this outcome. The
high prevalence of loneliness observed in this
sample underscores the importance of assessing for
it in healthcare settings, as it may identify older
veterans with or at risk for psychopathology. Addi-
tional research is needed to evaluate temporal and
causal associations among multifarious health and
psychosocial factors associated with loneliness; and
the effect of targeting these factors in psychothera-
peutic interventions
1,15
on mitigation of loneliness in
older veterans, as well as other at-risk elderly
populations.
References
1. Masi CM, Chen HY, Hawkley LC, et al: A meta-analysis of interven-
tions to reduce loneliness. Pers Soc Psychol Rev 2011; 15:219e266
2. Hawkley LC, Cacioppo JT: Loneliness matters: a theoretical and
empirical review of consequences and mechanisms. Ann Behav
Med 2010; 40:218e227
3. Russell D, Peplau LA, Cutrona CE: The revised UCLA loneliness
scale: concurrent and discriminant validity evidence. J Pers Soc
Psychol 1980; 39:472e480
4. Luanaigh CO, Lawlor BA: Loneliness and the health of older
people. Int J Geriatr Psychiatry 2008; 23:1213e1221
Am J Geriatr Psychiatry -:-,-2013 5
Kuwert et al.

5. Cohen GD: Loneliness in later life. Am J Geriatr Psychiatry 2000;
8:273e275
6. Vahia IV, Chattillion E, Kavirajan H, et al: Psychological protec-
tive factors across the lifespan: implications for psychiatry. Psy-
chiatr Clin North Am 2011; 34:231e248
7. Perissinotto CM, Stijacic Cenzer I, Covinsky KE: Loneliness in
older persons: a predictor of functional decline and death. Arch
Intern Med 2012; 172:1078e1083
8. Prigerson HG, Maciejewski PK, Rosenheck RA: Population
attributable fractions of psychiatric disorders and behavioral
outcomes associated with combat exposure among US men.
Am J Public Health 2002; 92:59e63
9. Macleod AD: The reactivation of post-traumatic stress disorder in
later life. Aust NZ J Psychiatry 1994; 28:625e634
10. Solomon Z, Dekel R: The contribution of loneliness and post-
traumatic stress disorder to marital adjustment following war
captivity: a longitudinal study. Fam Process 2008; 47:261e275
11. National Center for Veterans Analysis and Statistics. Data from the
American Community Survey. Profile of Veterans, 2011. http://
www.va.gov/vetdata/docs/SpecialReports/Profile_of_Veterans_
2011.pdf
12. Hughes ME, Waite LJ, Hawkley LC, et al: A short scale for
measuring loneliness in large surveys: results from two
population-based studies. Res Aging 2004; 26:655e672
13. Cacioppo JT, Hughes ME, Waite LJ, et al: Loneliness as a specific
risk factor for depressive symptoms: cross-sectional and longitu-
dinal analyses. Psychol Aging 2006; 21:140e151
14. Luo Y, Hawkley LC, WaiteLJ, et al: Loneliness, health, and mortality
in old age: a national longitudinal study. Soc Sci Med 2012; 74:
907e914
15. Pitkala KH, Routasalo P, Kautiainen H, et al: Effects of socially
stimulating group intervention on lonely, older people’s cogni-
tion: a randomized, controlled trial. Am J Geriatr Psychiatry 2011;
19:654e663
6 Am J Geriatr Psychiatry -:-,-2013
Loneliness Among Older Veterans in the U.S.