Suicidality and Nonsuicidal High-Risk Behavior in Military Veterans

Suicidality and Nonsuicidal High-Risk Behavior in Military Veterans: How
Does PTSD Symptom Presentation Relate to Behavioral Risk?
Nicholas Barr, Sara Kintzle, Kate Sullivan, and Carl Castro
University of Southern California
Considerable attention has been devoted to investigating posttraumatic stress disorder (PTSD) and
suicidality in military veterans, but nonsuicidal high-risk behaviors (HRBs), like reckless driving, are
underexplored and contribute to elevated rates of premature death in this population. This study is the 1st
to employ a structural equation modeling approach to investigate relationships between PTSD symptoms,
suicidality, and HRBs in veterans. Data for this study were drawn from a sample of veterans living in
Southern California. A multipronged sampling strategy was used to recruit 2,422 veterans, who completed an online battery. A structural equation model examining direct and indirect effects between 4
PTSD symptom factors, a suicidality factor, and an HRB factor was fitted to the data. We found positive
direct effects for reexperiencing and negative cognitive–mood symptoms on suicidality and for hyperarousal symptoms on HRBs. We found a negative direct effect for avoidance on suicidality. Suicidality
demonstrated a positive direct effect on HRBs. We detected positive indirect effects for reexperiencing
and negative cognitive–mood symptoms on HRBs and a negative indirect effect for avoidance on HRBs
through the suicidality pathway. Results show that distinct PTSD symptom clusters are associated with
different risk profiles in military veterans. High rates of HRBs with the potential for lethal outcomes
evident in our sample demonstrate the need for nuanced screening procedures. Veterans who do not meet
full clinical criteria for PTSD may be at risk for premature mortality from suicide and HRBs in the
context of reexperiencing, negative cognitive–mood, and hyperarousal symptoms.
Keywords: risk behavior, veterans, PTSD, suicidality
Considerable effort has been devoted to investigating the problem of completed suicide among U.S. military veterans. Less well
understood are elevated rates of potentially lethal but nonsuicidal
high-risk behaviors (HRBs), such as dangerous driving, evident
among some veterans. Both suicidal behaviors and HRBs have
been linked to posttraumatic stress disorder (PTSD; Killgore et al.,
2008; Sareen, Houlahan, Cox, & Asmundson, 2005), but no study
has jointly modeled relationships between suicidal behaviors,
HRBs, and symptoms of PTSD in veterans. Such an investigation
is crucial to improving screening and treatment of veterans presenting with symptoms of PTSD to mitigate risk of early mortality
evident in this population.
Although suicide rates among active-duty service members and
veterans have historically been lower than the civilian national
average, the recent conflicts in Iraq and Afghanistan have coincided with a dramatic shift in this dynamic. A comprehensive
epidemiological review of over 50 million veteran suicide records
collected from 2001 to 2014 indicates that veteran suicide rates
have increased by 32% since 2001, in comparison to a 23%
increase among civilians. After controlling for gender and age,
these data show a 21% greater risk of suicide for veterans in
comparison to civilians since 2001 (Veterans Health Administration [VHA], 2016). Additional evidence has indicated that veterans
are at 41%– 61% greater risk for suicide than those in the general
population (Kang et al., 2014).
Military veterans are also at heightened risk for developing
PTSD (Haugen, Evces, & Weiss, 2012). Evidence has suggested
that the prevalence of PTSD among veterans of the conflicts in
Iraq and Afghanistan is between 13.5% and 15.8% in deployed
veterans and 10.9% in nondeployed veterans (Dursa, Reinhard,
Barth, & Schneiderman, 2014). PTSD in general is associated with
a myriad of negative health outcomes, but research has shown that
specific PTSD symptom clusters predict different risk behaviors.
For example, reexperiencing symptoms are associated with increased risk for suicidal behavior (Barr, Sullivan, Kintzle, &
Castro, 2016; Bell & Nye, 2007), whereas hyperarousal is linked
to behaviors such as alcohol use and violence (McFall & Cook,
2006), which may increase the likelihood of HRBs.
Despite exploration of relationships between PTSD symptoms
and specific behavioral outcomes in veterans, the constellation of
risk factors for veterans’ premature death requires additional investigation. Studies have demonstrated increased mortality among
veterans relative to nonveteran populations due to preventable
This article was published Online First September 7, 2017.
Nicholas Barr, Suzanne Dworak-Peck School of Social Work, University of Southern California; Sara Kintzle, Center for Innovation and Research on Veterans & Military Families, and Suzanne Dworak-Peck School
of Social Work, University of Southern California; Kate Sullivan, Suzanne
Dworak-Peck School of Social Work, University of Southern California;
Carl Castro, Center for Innovation and Research on Veterans & Military
Families, and Suzanne Dworak-Peck School of Social Work, University of
Southern California.
We are indebted to the veterans who participated in the study and
through their self-disclosure enable us to help other veterans.
Correspondence concerning this article should be addressed to Nicholas
Barr, Suzanne Dworak-Peck School of Social Work, University of Southern California, Montgomery Ross Fisher Building, Los Angeles, CA
90089. E-mail: [email protected]
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Traumatology © 2017 American Psychological Association
2018, Vol. 24, No. 1, 55– 61 1085-9373/18/$12.00
http://dx.doi.org/10.1037/trm0000133
55
causes like automobile accidents (Hooper et al., 2006; Kang &
Bullman, 1996; Watanabe & Kang, 1995). Evidence has also
suggested that combat-deployed veterans, particularly those with
multiple combat experiences, engage in more postdeployment
HRBs than do nondeployed veterans and those with fewer combat
experiences (Hooper et al., 2006; Killgore et al., 2008). This
phenomenon has been described as postcombat invincibility, or the
potential for combat veterans exposed to violence and trauma to
experience an alteration in perceived invincibility (Killgore et al.,
2008). Not surprisingly, PTSD has been associated with HRBs and
increased mortality among veterans; in one study examining
causes of death among male veterans with PTSD, two thirds of
veterans’ deaths were attributed to treatable or preventable behavioral causes such as accidents, substance abuse, and suicidality
(Drescher, Rosen, Burling, & Foy, 2003).
We suggest that, consistent with the phenomenon of postcombat
invincibility, habituation to physical and psychological distress
acquired through military training and/or combat exposure may
place veterans at risk for engaging in behavior that, although not
overtly suicidal, nonetheless contributes to elevated risk for premature death. In the present study, we hone in on a subset of
veterans at elevated risk for premature death not only from suicide
but also from HRBs.
In a previous study with a small group of veterans (Barr et al.,
2016), we found significant associations between reexperiencing,
avoidance, suicidal ideation, and suicide plan as well as between
hyperarousal symptoms and a single item querying whether respondents had taken an unnecessary life-threatening risk in the last
year. The current study builds on this empirical foundation by
applying a structural equation model to relationships between
PTSD symptom latent factors, suicidal behaviors, and an HRB
latent factor in a large sample of pre- and post-9/11 veterans. In
addition, the current study employs a four-factor model of PTSD
consistent with diagnostic guidelines of the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM–5; American
Psychological Association, 2013). Our aim was to develop a
nuanced picture of behavioral risk profiles associated with PTSD
symptom clusters in military veterans. Our hypotheses were as
follows:
Hypothesis 1: Reexperiencing, negative cognitive–mood, and
hyperarousal symptom clusters will be significantly associated
with suicidality, but reexperiencing will have the strongest
direct effect.
Hypothesis 2: Reexperiencing, negative cognitive–mood, and
hyperarousal symptom clusters will be significantly associated
with HRBs, but hyperarousal will have the strongest direct
effect.
Hypothesis 3: Suicidal behavior will have a significant direct
effect on HRBs; PTSD symptom clusters will have indirect
effects on HRBs through the suicidality pathway.
Method
Data and Participants
Data for these analyses were drawn from a survey of veterans
residing in two large urban Southern California counties. The
survey was completed in 2015 by a total of 2,422 veterans. Survey
participants were identified through a four-pronged sampling strategy to account for the diversity of the veteran population in the
area. The sampling strategy was designed to target veterans who
are not seen in traditional treatment contexts like Veterans Affairs
clinics.
The first strategy utilized a California state veterans’ organization with contact information of veterans who reported California
residence during their transition out of the military. The agency
identified Los Angeles and Orange County veterans through zip
codes and invited them to complete the survey using an online
survey link (n ! 309). The second sampling strategy used a Los
Angeles County information and referral center by identifying
potential participants through their initial call screening. Callers
who self-identified as former service members were asked permission to be contacted regarding the research study. Those who
agreed to participate were sent either a paper survey copy or the
online survey link (n ! 98).
The third sampling strategy involved partnering with agencies
that serve Los Angeles County veterans as well as college veteran
agencies (n ! 607). Two methods were used to collect agency
data. The first method utilized an online survey approach where
the agency would send out an invitation and survey link to veterans
within their database. The second method used an on-ground
survey approach where agencies would work with the researchers
to organize data collection events within their agencies. The final
sampling strategy used TV and print advertisements, a public
service announcement, and social media to build a presence within
the Los Angeles County veteran community. Avenues such as
Facebook, Twitter, LinkedIn, mass e-mails, and the survey website
promoted the survey opportunity to potential participants (n !
1,408). All study participants who completed the survey received
a $15 gift card for completing the survey, which took approximately 30 –90 min (Castro, Kintzle, & Hassan, 2014). All procedures were reviewed and approved by the University of Southern
California Institutional Review Board. Sample characteristics are
presented in Table 1.
Measures
Suicidality. The survey instrument employed the dichotomous items “In the past 12 months did you ever seriously consider
attempting suicide” and “In the past 12 months did you make a
plan about how you would attempt suicide” to capture suicidal
ideation and plan, respectively.
Nonsuicidal high-risk behavior. The survey instrument included a risk behavior checklist (Adler, Bliese, McGurk, Hoge, &
Castro, 2009) comprising 15 dichotomous items querying risk
behaviors in the last 12 months. The following five risk behavior
items were selected based on substantive and statistical considerations discussed in detail in the Results section: “Have you engaged in any of the following activities in the last 12 months: (1)
driven recklessly, (2) looked to start a fight, (3) carried a weapon
for nonduty purposes, (4) took health risks, (5) took unnecessary
risks to life.”
Posttraumatic stress disorder. The PTSD Checklist—Military version (PCLM; Weathers & Ford, 1996) was used to capture
PTSD symptoms. The PCLM is a 17-item measure that asks
respondents to indicate how much they were bothered by sympThis document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
56 BARR, KINTZLE, SULLIVAN, AND CASTRO
toms in the last 5 days on a 5-point Likert scale ranging from 0 (not
at all) to 5 (extremely). The PCLM includes reexperiencing, avoidance, and hyperarousal subscales. Cronbach’s alpha for the PCLM
was .99 in these data.
Procedure
First, we conducted analyses to examine differences between
participants’ scores on study variables based on sampling strategy;
no significant differences were detected. Next, we examined bivariate correlations between independent and dependent variables
at the item level. Correlations significant at the .05 level or below
were observed between all items. Next, we conducted exploratory
(EFA) and confirmatory factor analysis (CFA) to verify a fourfactor measurement model for the PCLM in the current sample.
We then conducted EFA and CFA of the 15-item risk behavior
checklist to verify the hypothesized unidimensional structure of
HRBs with the potential to cause premature death. The unidimensional measurement model comprising a latent suicidality factor
with observed indicators suicidal ideation and suicide plan was
underidentified, and no measurement modeled was computed.
The final structural equation model comprised PCLM Items 1–5
loading onto the reexperiencing factor, Items 6 – 8 loading onto the
avoidance factor, Items 9 –12 loading onto the negative cognition–
mood factor, and Items 13–17 loading onto the hyperarousal
factor. These latent factors were hypothesized to exert direct
effects on (a) a suicidality latent factor comprising the indicators
suicidal ideation and suicide plan and (b) an HRB factor comprising the five risk behavior indicators. Indirect effects of PTSD
symptom latent factors on high-risk behavior were modeled
through the suicidality pathway.
Results
Data cleaning was conducted in SAS 9.4. All statistical analyses
were conducted in Mplus 7.0 (Muthén & Muthén, 2007). For the
PCLM, the proportion of missing data was .16 with 76% of
missingness at the scale level. For risk behavior items, the proportion of missing data was .17 with 64% of missingness accounted for by nonresponse on single scale items and 34% of
missingness at the scale level. For both suicide items, the proportion of missing data was .13. To account for missing data and
nonnormal distribution of the PCLM scale, we used full information maximum likelihood and maximum likelihood with robust
standard errors (MLR) estimators to generate parameter estimates.
Numerous studies have demonstrated the robustness of MLR for
parameter estimation with nonnormal continuous data (e.g., Li,
2016; Maydeu-Olivares, 2017; Muthén & Muthén, 2007), because
it was developed for this purpose.
We considered, consistent with previous literature (Little, 2013;
Maslowsky, Jager, & Hemkin, 2015), comparative fit index (CFI)
and Tucker–Lewis index (TLI) values above .95 and root-meansquare error of approximation (RMSEA) below .08 to indicate
good model fit. In this context, the CFA of PCLM items derived
from initial EFA results suggested a good fit to the data, “2
(113,
N ! 2,121) ! 1,530.40, p # .001, CFI ! .97, TLI ! .96,
RMSEA ! .08, 95% confidence interval [CI: .07, .08]. Because
these indices were near the boundaries of good fit, we applied the
Lagrange multiplier (LM) test to facilitate model respecification
through inclusion of additional parameters. Based on these results,
correlations were allowed between residuals for PCLM Items 1
and 2 as well as 16 and 17. The respecified model demonstrated
improved fit, “2
(111, N ! 2,121) ! 1,019.22, p # .001, CFI !
.98, TLI ! .97, RMSEA ! .06, 95% CI [.06, .07]. All factor
loadings were significant and ranged from .93 to .79. These fit
indices and factor loadings indicate good reliability for PCLM
items. In addition, identification of a four-factor model for PTSD
symptoms consistent with DSM–5 diagnostic criteria (American
Psychiatric Association, 2013), as well as relationships between
PTSD latent factors and outcomes consistent with study hypotheses, support the validity of PTSD latent factors in these data.
The EFA of the 15-item risk behavior checklist yielded four
distinct factors with the items (a) driving recklessly, (b) looking to
start a fight, (c) carrying a weapon, (d) taking health risks, and (e)
taking unnecessary risks to life loading onto the HRB factor.
Results of a CFA based on this structure yielded “2
(5, N !
2,272) ! 109.55, p # .001, CFI ! .97, TLI ! .94, RMSEA ! .10,
95% CI [.08, .12]. To improve model fit, we examined LM test
results and allowed residuals to correlate for Items RB6 and RB15.
Results of the respecified measurement model suggest good fit to
the data, “2
(4, N ! 2,272) ! 11.99, p # .01, CFI ! .99, TLI !
.99, RMSEA ! .03, 95% CI [.01, .05]; all parameter estimates
were significant and ranged from .73 to .55. These factor loadings
and fit indices indicate that the HRB items are reliable indicators
of an underlying HRB factor in these data. Substantive examination of individual HRB items, each of which convey risk for
causing or experiencing physical harm, as well as relationships
Table 1
Sample Demographic Characteristics and Descriptive Statistics
for PTSD, Suicide Risk, and Nonsuicidal Risk Behavior
Characteristic n %
Age 2,419
18–20 .54
21–25 4.63
26–30 12.03
31–40 18.48
41–50 14.30
51–60 22.41
61–70 19.26
Over 71 8.35
Race 2,402
American Indian/Alaska native 1.90
Asian 4.50
African American 18.00
Hawaiian/Pacific Islander 1.00
White 46.70
Latino 20.60
Other 7.20
Clinical criteria for PTSD met 2,027 34.78
Suicidality
Suicidal ideation 2,113 13.82
Suicide plan 2,117 10.20
Nonsuicidal risk behavior
Drove recklessly 2,271 11.89
Looked to start a fight 2,271 14.75
Carried a weapon 2,270 11.10
Took health risks 2,269 16.04
Unnecessary risk to life behavior 2,005 15.06
Note. PTSD ! posttraumatic stress disorder.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
PTSD, SUICIDALITY, AND RISK BEHAVIOR IN VETERANS 57
between the HRB latent factor, hyperarousal latent factor, and
suicidality latent factor in line with study hypotheses, support
validity of the HRB latent factor in these data.
Although a measurement model was not computed for the
two-item suicidality latent factor due to model underidentification,
dichotomous suicide ideation and suicide plan items are widely
recognized as valid and reliable indicators of suicide risk (Kessler,
Berglund, Borges, Nock, & Wang, 2005; Nock et al., 2008; Posner
et al., 2011). In the current study, standardized loadings for the
suicidality latent factor were .84 for the suicide ideation item and
.73 for the suicide plan item. These items were significantly
correlated (r ! .61) but not to the degree suggestive of collinearity
(Cohen, Cohen, West, & Aiken, 2013). In the context of substantive considerations including theoretical support for the hypothesized relationships among study variables as well as statistical
considerations including very good overall model fit, the latent
suicidality factor demonstrates validity and reliability in these data.
The full model with all significant direct and indirect effects is
presented in Figure 1. The final factor structure is presented in
Figure 2. Results suggest very good fit to the data, “2
(234, N !
2,315) ! 837.99, p # .001, CFI ! .98, TLI ! .98, RMSEA ! .03,
95% CI [.03, .04].
Hypothesis 1 was partially supported; the reexperiencing factor
showed a direct effect on suicidality ($ ! .50, p # .001). Avoidance ($!% .31, p # .001) and negative cognitions–mood ($ !
.35, p # .01) were also significantly associated with suicidality,
but the hyperarousal factor was not. Hypothesis 2 was also partially supported. Only hyperarousal demonstrated a significant
direct effect on HRB ($ ! .47, p # .001).
Consisted with Hypothesis 3, results showed a direct effect of
suicidality on HRB (& ! .21, p # .001) in addition to the significant
direct effect of reexperiencing on suicidality. Reexperiencing ($& !
.10, p # .001), avoidance ($& ! % .07, p # .05), and negative
cognitions–mood ($& ! .07, p # .01) each demonstrated a significant
indirect effect on HRB through the suicidality pathway.
Discussion
This study yielded novel findings related to associations between
PTSD symptoms, suicidality, and HRB among veterans. Results
PCL
M 17
PCL
M 16
PCL
M 15
PCL
M14
PCL
M 13
PCL
M 12
PCL
M 11
PCL
M10
PCL
M9
PCL
M8
PCL
M7
PCL
M1
PCL
M6
PCL
M5
PCL
M4
PCL
M3
PCL
M2
REEXPERIENCING
AVOIDANCE
NEGATIVE
COGNITIONS/
MOOD
HYPERAROUSAL
HRB
SUICIDALITY
Plan
Attempt
RB3
RB4
RB5
RB6
RB15
0.90
0.89
0.89
0.92
0.91
0.91
0.92
0.79
0.89
0.93
0.90
0.81
0.79
0.87
0.88
0.83
0.86
0.01
0.01
0.01
0.00
0.00
0.01
0.00
0.01
0.01
0.00
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.74
0.83
0.02
0.02
0.02
0.02
0.02
0.62
0.67
0.54
0.70
0.74
0.09
-0.20
0.10
0.47
0.21
0.50
-0.31
0.35
-0.10
0.95
0.83
0.85 0.88
0.87
0.92
0.43
0.35 0.40
Figure 1. Full structural equation model of posttraumatic stress disorder (PTSD), suicidality, and high-risk
behavior (HRB) latent factor loadings and relationships between latent factors. Unidirectional arrows indicate
regression relationships; bidirectional arrows indicate correlational relationships. PCLM ! PTSD Checklist—
Military version; RB ! risk behavior.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
58 BARR, KINTZLE, SULLIVAN, AND CASTRO
indicate a strong link between hyperarousal and HRB but not reported
suicidality. This suggests that veterans whose experience of PTSD
symptoms is characterized by hyperarousal and who have an enhanced tolerance for physical and psychological discomfort may be
more likely to evince behavioral disinhibition that facilitates risk
taking with potentially fatal outcomes. In light of veterans’ increased
risk for early mortality from preventable causes (Drescher et al.,
2003), the high rates of HRB found in the current sample are cause for
alarm. Roughly 15% of veterans sampled reported risking their health
or lives unnecessarily, and nearly 12% reported driving recklessly or
carrying a weapon for nonduty purposes. In the context of hyperarousal symptoms and acquired capability for tolerating physical and
psychological distress, this constellation of risk factors may place
veterans in significant danger of early mortality. But without a nuanced approach to PTSD screening, the subset of veterans engaging in
HRB in the context of hyperarousal symptoms may not receive
clinical attention.
In addition, this study is the first to examine relationships
between PTSD symptoms, suicidality, and HRB in a single
model. We found that, consistent with previous literature (Barr
et al., 2016; Bell & Nye, 2007), reexperiencing symptoms
provide the most explanatory power for suicidality, which reinforces the notion that reexperiencing symptoms may be the
most subjectively distressing for veterans. Negative cognitive–
mood symptoms were also linked to suicidality, indicating an
increased burden of risk for veterans whose experience of
PTSD is characterized by these symptoms clusters. In contrast,
avoidance symptoms were negatively associated with suicidality, which suggests that these symptoms, although problematic
for other reasons, may buffer against suicidality, in contrast
with reexperiencing, negative cognitions–mood, and hyperarousal symptoms. Suicidality was also significantly linked to
HRB, a finding that supports our conception of veterans’ risk
for early mortality as emergent from a constellation of risk
REEXPERIENCING
AVOIDANCE
NEGATIVE
COGNITIONS/
MOOD
HYPERAROUSAL
HRB
SUICIDALITY
0.47
0.21
0.50
-0.31
0.35
0.95
0.83
0.85 0.88
0.87
0.92
Figure 2. Factor structure and significant direct effects between PTSD, suicidality, and HRB latent factors.
Unidirectional arrows indicate regression relationships; bidirectional arrows indicate correlational relationships.
HRB ! high-risk behavior.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
PTSD, SUICIDALITY, AND RISK BEHAVIOR IN VETERANS 59
factors, including both overt suicidal behaviors and HRB associated with postcombat invincibility in the context of PTSD
symptoms.
Examination of total effects of PTSD symptom clusters on suicidality and HRB yields a complex picture of relationships between
symptoms and associated behavioral health risks. These findings lend
support to the notion that PTSD symptom presentation is highly
nuanced and associated with heterogeneous risk profiles. In these
data, both reexperiencing and negative cognitions–mood showed
small though positive and significant indirect effects on HRB through
the suicidality pathway, indicating that these symptom clusters account for a portion of the explanatory power of suicidality on HRB in
our model. In contrast, the negative indirect effect of avoidance on
HRB through the suicidality pathway is consistent with our conceptualization of avoidance symptoms as a buffer against behavioral
health risks driven by the other PTSD symptom clusters.
From a clinical perspective, these relationships support the view
that reexperiencing, negative cognitive–mood, and hyperarousal
symptoms ought to be the target of careful assessment because
they may be the primary drivers of PTSD-related risk for early
mortality in veterans. Hyperarousal symptoms in particular, even
absent the satisfaction of full clinical criteria for PTSD, ought to be
cause for concern because they may drive HRB not captured by
suicide screening instruments. In veterans, reexperiencing, negative cognitive–mood, or hyperarousal symptoms, even in the absence of additional indicators or the complete satisfaction of PTSD
diagnostic criteria, may warrant intervention.
The problem of veteran suicide in the context of PTSD symptoms continues to resist the best efforts of both researchers and
clinicians; more research is required to clarify relationships between PTSD symptoms and dimensions of risk in the veteran
population. Our findings support the view that PTSD symptom
presentation is complex, nuanced, and predictive of multiple dimensions of potentially lethal risk including but not limited to
suicidal behaviors. It is also important to note that approximately
50% of the current sample is over 50 years old. Although recent
evidence has shown that veterans ages 18 –29 are at the highest
risk for suicide (VHA, 2016) and popular media accounts often
depict suicide and risky behaviors as problems confined to
younger veteran cohorts, we argue that, in light of this study’s
findings, the constellation of risk factors for early mortality in
older veterans must take into account not only suicide risk behaviors but also HRB. The risk for early mortality posed by long-term
PTSD symptoms and associated risk behaviors remains underinvestigated in older veterans. Although researchers and institutional
actors have rightly focused on the mental and behavioral health
challenges faced by recent veterans, continued investigation of
these risks in older veterans remains imperative.
This study’s findings point to the need for more careful screening of veterans presenting with PTSD because nuances of symptomology may reflect different risk profiles; specific symptoms of
PTSD, such as reexperiencing or hyperarousal, may be more
salient predictors of suicidal behavior or HRB than is a simple
clinical cutoff point or satisfaction of full clinical diagnostic criteria. Expanding the understanding of the risks posed by specific
PTSD symptoms to include a more holistic picture of risk encompassing HRB in addition to suicidal behaviors is a critical component of effective intervention.
Limitations
Data for this study were cross-sectional, and causal relationships
among study variables cannot be assumed. In addition, this study’s
purposive, nonrandom sampling design may limit generalizability
to the larger population of military veterans, particularly in light of
the age of the sample. These data contained single-item indicators
for suicide ideation and plan; similarly, HRB indicators were
dichotomous and did not capture frequency of behaviors. Future
studies would benefit from inclusion of more robust indicators for
these constructs.
References
Adler, A. B., Bliese, P. D., McGurk, D., Hoge, C. W., & Castro, C. A.
(2009). Battlemind debriefing and battlemind training as early interventions with soldiers returning from Iraq: Randomization by platoon.
Journal of Consulting and Clinical Psychology, 77, 928 –940. http://dx
.doi.org/10.1037/a0016877
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric
Publishing.
Barr, N., Sullivan, K., Kintzle, S., & Castro, C. (2016). PTSD symptoms,
suicidality and non-suicidal risk to life behavior in a mixed sample of pre
and post 9/11 veterans. Social Work in Mental Health, 14, 465– 473.
http://dx.doi.org/10.1080/15332985.2015.1081666
Bell, J. B., & Nye, E. C. (2007). Specific symptoms predict suicidal
ideation in Vietnam combat veterans with chronic post-traumatic stress
disorder. Military Medicine, 172, 1144 –1147.
Castro, C. A., Kintzle, S., & Hassan, A. (2014). The state of the American
veteran: The Los Angeles County Veterans’ Survey. Los Angeles, CA:
Center for Innovation and Research on Military Veterans and Families.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2013). Applied multiple
regression/correlation analysis for the behavioral sciences. New York,
NY: Routledge.
Drescher, K. D., Rosen, C. S., Burling, T. A., & Foy, D. W. (2003). Causes
of death among male veterans who received residential treatment for
PTSD. Journal of Traumatic Stress, 16, 535–543. http://dx.doi.org/10
.1023/B:JOTS.0000004076.62793.79
Dursa, E. K., Reinhard, M. J., Barth, S. K., & Schneiderman, A. I. (2014).
Prevalence of a positive screen for PTSD among OEF/OIF and OEF/
OIF-era veterans in a large population-based cohort. Journal of Traumatic Stress, 27, 542–549. http://dx.doi.org/10.1002/jts.21956
Haugen, P. T., Evces, M., & Weiss, D. S. (2012). Treating posttraumatic
stress disorder in first responders: A systematic review. Clinical Psychology Review, 32, 370 –380. http://dx.doi.org/10.1016/j.cpr.2012.04
.001
Hooper, T. I., Debakey, S. F., Bellis, K. S., Kang, H. K., Cowan, D. N.,
Lincoln, A. E., & Gackstetter, G. D. (2006). Understanding the effect of
deployment on the risk of fatal motor vehicle crashes: A nested casecontrol study of fatalities in Gulf War era veterans, 1991–1995. Accident
Analysis & Prevention, 38, 518 –525. http://dx.doi.org/10.1016/j.aap
.2005.11.009
Kang, H. K., & Bullman, T. A. (1996). Mortality among U.S. veterans of
the Persian Gulf War. New England Journal of Medicine, 335, 1498 –
1504. http://dx.doi.org/10.1056/NEJM199611143352006
Kang, H. K., Bullman, T. A., Smolenski, D. J., Skopp, N. A., Gahm, G. A.,
& Reger, M. A. (2015). Suicide risk among 1.3 million veterans who
were on active duty during the Iraq and Afghanistan wars. Annals of
Epidemiology, 25, 96 –100.
Kessler, R. C., Berglund, P., Borges, G., Nock, M., & Wang, P. S. (2005).
Trends in suicide ideation, plans, gestures, and attempts in the United
States, 1990-1992 to 2001-2003. Journal of the American Medical
Association, 293, 2487–2495.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
60 BARR, KINTZLE, SULLIVAN, AND CASTRO
Killgore, W. D., Cotting, D. I., Thomas, J. L., Cox, A. L., McGurk, D., Vo,
A. H., . . . Hoge, C. W. (2008). Post-combat invincibility: Violent
combat experiences are associated with increased risk-taking propensity
following deployment. Journal of Psychiatric Research, 42, 1112–1121.
http://dx.doi.org/10.1016/j.jpsychires.2008.01.001
Li, C.-H. (2016). Confirmatory factor analysis with ordinal data: Comparing robust maximum likelihood and diagonally weighted least squares.
Behavior Research Methods, 48, 936 –949. http://dx.doi.org/10.3758/
s13428-015-0619-7
Little, T. D. (2013). Longitudinal structural equation modeling. New York,
NY: Guilford Press.
Maslowsky, J., Jager, J., & Hemken, D. (2015). Estimating and interpreting
latent variable interactions: A tutorial for applying the latent moderated
structural equations method. International Journal of Behavioral Development, 39, 87–96. https://doi.org/10.1177/0165025414552301
Maydeu-Olivares, A. (2017). Maximum likelihood estimation of structural
equation models for continuous data: Standard errors and goodness of
fit. Structural Equation Modeling, 24, 383–394. http://dx.doi.org/10
.1080/10705511.2016.1269606
McFall, M., & Cook, J. (2006). PTSD and health risk behavior. PTSD
Research Quarterly, 17, 1– 8.
Muthén, L. K., & Muthén, B. O. (2007). Mplus user’s guide. Los Angeles,
CA: Author.
Nock, M. K., Borges, G., Bromet, E. J., Cha, C. B., Kessler, R. C., & Lee,
S. (2008). Suicide and suicidal behavior. Epidemiologic Reviews, 30,
133–154.
Posner, K., Brown, G. K., Stanley, B., Brent, D. A., Yershova, K. V.,
Oquendo, M. A.,… Mann, J. J. (2011). The Columbia–Suicide Severity
Rating Scale: initial validity and internal consistency findings from three
multisite studies with adolescents and adults. American Journal of
Psychiatry, 168, 1266 –1277.
Sareen, J., Houlahan, T., Cox, B. J., & Asmundson, G. J. (2005). Anxiety
disorders associated with suicidal ideation and suicide attempts in the
National Comorbidity Survey. Journal of Nervous and Mental Disease,
193, 450 – 454. http://dx.doi.org/10.1097/01.nmd.0000168263.89652.6b
Veterans Health Administration. (2016). Suicide among veterans and other
Americans 2001–2014. Retrieved from http://www.mentalhealth.va.gov/
docs/2016suicidedatareport.pdf
Watanabe, K. K., & Kang, H. K. (1995). Military service in Vietnam and
the risk of death from trauma and selected cancers. Annals of Epidemiology, 5, 407– 412. http://dx.doi.org/10.1016/1047-2797(95)00039-A
Weathers, F., Ford, J. (1996). Psychometric properties of the PTSD Checklist (PCL-C, PCL-S, PCLM, PCL-PR). In B. H. Stamm (Ed.), Measurement of stress, trauma, and adaptation. Lutherville, MD: Sidran Press.
Received November 10, 2016
Revision received July 13, 2017
Accepted August 2, 2017 !
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
PTSD, SUICIDALITY, AND RISK BEHAVIOR IN VETERANS 61