Resilience Facilitates Adjustment Through Greater Psychological Flexibility among Iraq/Afghanistan War Veterans With and Without Mild Traumatic Brain Injury

Abstract

Objectives:

Although many Iraq/Afghanistan warzone veterans report few problems with adjustment, a substantial proportion report debilitating mental health symptoms and functional impairment, suggesting the influence of personal factors that may promote adjustment. A significant minority also incur warzone-related traumatic brain injury (TBI), the majority of which are of mild severity (mTBI). We tested direct and indirect pathways through which a resilient personality prototype predicts adjustment of warzone veterans with and without mTBI over time.

Method:

A sample of 264 war veterans (181 men) completed measures of lifetime and warzone-related TBIs, personality traits, psychological adjustment, quality of life, and functional impairment. Social support, coping and psychological flexibility were examined as mediators of the resilience-adjustment relationship. Instruments were administered at baseline, 4-, 8-, and 12-month assessments. Structural equation models accounted for combat exposure and response style.

Results:

Compared with a non-resilient personality prototype, a resilient prototype was directly associated with lower PTSD, depression, and functional disability, and higher quality of life at all time-points. Warzone mTBIs frequency was associated with higher scores on a measure of functional disability. Indirect effects via psychological flexibility were observed from personality to all outcomes, and from warzone-related mTBIs to PTSD, depression, and functional disability, at each time-point.

Conclusions:

Several characteristics differentiate veterans who are resilient from those who are less so. These findings reveal several factors through which a resilient personality prototype and the number of mTBIs may be associated with veteran adjustment. Psychological flexibility appears to be a critical modifiable factor in veteran adjustment.

Many veterans returning from deployment to the post-9/11 wars in Iraq and Afghanistan are resilient, reporting few problems with adjustment or distress (Bonanno et al., 2012; Isaacs et al., 2017). A substantial proportion report debilitating symptoms of posttraumatic stress disorder (PTSD; 12% to 23%) and depression (15%; Fulton et al., 2015; Seal, Bertenthal, Miner, Sen, & Marmar, 2007). A significant minority also sustain a traumatic brain injury (TBI) during deployment. An estimated 10% of Iraq and Afghanistan veterans using Veterans Health Administration services between 2009 and 2011 were classified as having incurred a TBI (Cifu et al., 2014). The majority of TBIs occurring among military personnel are mild (mTBI; 82.3% out of 383,947 cases; Defense and Veterans Brain Injury Center, 2019).

Certain symptoms of PTSD and depression are similar to those associated with post-concussive syndromes including fatigue, social withdrawal, sleep disturbance, somatic complaints, sadness, irritability, interpersonal sensitivity and difficulties concentrating (Bailie et al., 2016; Lange et al., 2014). Some studies find associations with distress and decreased quality of life among veterans with deployment-related mTBI (Brickell, Lange, & French, 2014; Dretsch, Silverberg, & Iverson, 2015). Critical and meta-analytic literature reviews argue that limited evidence substantiates mTBI as a risk factor for emotional problems (Carroll et al., 2014) suggesting that various psychological (e.g., preinjury mental health issues) and situational factors (e.g., litigation) account for much of the variance in adjustment problems post-mTBI (Broshek, De Marco, & Freeman, 2015; Panayioutou, Jackson, & Crowe, 2010; Rohling, Langhinrichsen-Rohling, & Axelrod, 2017). Studies utilizing omnibus measures of personality to monitor symptom exaggeration and distress among individuals with and without mTBI support these observations (Jones, Ingram, & Ben-Porath, 2012; Kennedy, Cooper, Reid, Tate, & Lange, 2015).

Despite the documented influence of individual and situational differences in adjustment of veterans with mTBI, the contribution of personality characteristics is understudied. Personality variables are often examined in this literature in relation to symptom exaggeration (e.g., Jurik et al., in press) or as indicators of distress that vary as a function of TBI severity (e.g., Kennedy et al., 2015). These approaches do little to inform our theoretical appreciation of personality factors that characterize veterans who may be more likely to exhibit optimal adjustment following deployment, regardless of their mTBI status.

To some extent, this gap has been addressed by recent cross-sectional studies that document inverse relationships between self-reported resilience and distress in mTBI samples (Losoi et al., 2015; Sullivan, Edmed, Allan, Smith, & Karlsson, 2015). However, these studies relied on face-valid measures derived from atheoretical conceptualizations of resilience that did not specify testable mechanisms through which resilience facilitated adjustment, a feature critical to developing clinical interventions. The lack of a comprehensive and testable theoretical model of resilience in this research undermines our ability to develop and provide psychological interventions that could promote resilience (Davydov, Stewart, Ritchie, & Chandieu, 2010).

In the present study, we rely on a theoretical model of resilience to test hypothesized relationships among psychological factors that differentially predict subsequent adjustment among warzone veterans with and without deployment-related mTBI. According to Block’s theory of personality development, ego control and ego resiliency develop from healthy attachments formed during infancy and maintained through childhood, and these factors provide an individual with capacities essential to effectively adapt to change, stress, and conflict (Block, 1993; Block & Block, 1980). Resilient individuals exhibit proactive behavior, pursue personally meaningful goals, and report greater flexibility, self-regulation, resourcefulness and active engagement with the environment under routine and stressful circumstances (Gramzow et al., 2004; Ong, Bergeman, & Boker, 2009). These capacities cultivate a positive cascade of rewarding emotional and behavioral experiences that, in turn, motivate further engagement with the environment and others, enhancing personal control, self-esteem, and confidence (Elliott, Barron, Stein, Wright, & Lowery, in press). Longitudinal research indicates resilient individuals are more socially competent as children and adults, have more rewarding interpersonal relationships, and are more likely to exhibit optimal physical and emotional health (Caspi, 2000; Dennissen, Asendorpf, & van Aken, 2008).

Applications of this model in studies of persons with debilitating, chronic health problems have found that resilient individuals have a proclivity for positive emotions (among persons with chronic pain; Ong, Zautra, & Reid, 2010), participate in desired activities (following upper limb loss; Walsh et al., 2016), and use more effective coping strategies (following spinal cord injury; Berry, Elliott, & Rivera, 2007). In previous work, we found a resilient personality prototype among war veterans predicted adjustment over time through associations with social support and psychological flexibility, and through an inverse relationship with avoidant coping (Elliott et al., 2015). These relationships were independent of a positive TBI screen and degree of combat exposure. Study limitations included a relatively low sample size (N = 127), inattention to quality of life, and no assessment of the number and severity of TBIs.

The current study extends our investigation of the ways in which a resilient personality prototype facilitates adjustment among war veterans with and without mTBI. We rectify the limitations of the previous work (Elliott et al., 2015) by including various indicators of TBI including lifetime number of TBIs, number of mTBIs incurred during post-9/11 warzone deployment, and length of loss of consciousness and post-traumatic amnesia as indicators of mTBI severity. We used a contextual model to test the positive effects of resilience on veteran adjustment, relying on an established measure of “Big Five” personality traits to identify a resilient personality prototype (Alessandri et al., 2014). We then examined the presumed beneficial associations of resilience with characteristics critical to veteran adjustment: social support, effective coping strategies, and psychological flexibility. We expected that resilience would predict important elements of veteran emotional adjustment (PTSD, depression) and functioning (quality of life, functional disability) over a 12-month period through its predicted effects on these mediators. In addition, we included a self-report measure of resilience as a potential mediator, to further elucidate the predictive properties of these popular measures when other psychological variables (e.g., personality traits, social support, coping) are taken into account. Combat exposure and socially desirable response tendencies served as covariates.

Method

Participants and Procedure

Procedures were approved by the local Institutional Review Board. Post-9/11 veterans enrolled in the Central Texas Veterans Health Care System (CTVHCS) were recruited to participate in a longitudinal study of predictors of post-deployment adjustment. Participants were recruited through advertisements and direct mailings to veterans enrolled at the healthcare system.

Telephone screenings were conducted to determine eligibility criteria: service in support of the Iraq or Afghanistan campaigns after September 11, 2001; stable on psychotropic medications or in psychotherapy if receiving such interventions; and not relocating out of the area within four months of consent. Initially eligible participants were scheduled for an in-person assessment. Informed written consent was obtained prior to the baseline interview during which participants completed a diagnostic interview and self-report measures. Participants were ineligible if they had suicidal or homicidal ideation warranting crisis intervention, or if they had an existing bipolar or psychotic disorder. Two follow-up assessments were mailed four and eight months later. At a 12-month follow-up, participants completed self-report measures and an in-person diagnostic interview conducted by master’s and doctoral level assessors. Diagnostic consensus was reached on all interviews under the supervision of a clinical psychologist.

A total of 345 veterans consented to participate. Thirty-six were excluded because of schizophrenia or bipolar disorder. Forty-three with moderate or severe TBI were also excluded. Two others were excluded due to missing personality data. The final sample for this study was 264 (average age = 38.94 years, SD = 9.94; average years of education = 14.09, SD = 2.09). The majority were male (n = 181; 68.6%) and identified as White/Caucasian (n = 147; African American, n = 90; other, n = 27). Fifty-two (19.7%) identified as Hispanic. Approximately 72% had a VA service-connected disability.

Independent Variables

Personality Prototypes.

The 60-item NEO Five-Factor Inventory (NEO–FFI; Costa & McCrae, 1992) was administered at baseline. The NEO-FFI is a shortened version of the Revised NEO Personality Inventory (NEO–PI–R; Costa & McCrae, 1992) for assessing five domains of adult personality: Neuroticism, Extraversion, Openness to Experience, Agreeableness, and Conscientiousness. The NEO–FFI factors show high correlations (.75 to .89) with the full-scale NEO–PI factors. Internal consistency reliabilities range from .68 (Agreeableness) to .86 (Neuroticism), and test–retest reliabilities from .79 (Extraversion, Openness) to .89 (Neuroticsm; Costa & McCrae, 1992). Raw scores were converted to gender-specific T scores based on adult norms (Costa & McCrae, 1992).

Studies of resilient personality prototypes usually apply clustering techniques to personality data collected from a variety of measurement methods including proxy reports, behavioral ratings, Q-sorts, and self-report measures of personality traits. Instruments that assess the “Big Five” traits are often used, as the resilient prototype is consistently represented by elevations on Agreeableness, Conscientiousness, Openness to Experience and Extraversion, and lower scores on Neuroticism (Alessandri et al., 2014; Farkas & Orosz, 2015). This pattern captures the ego resiliency and ego control dimensions described in the Block model (Block & Block, 1980). Two other replicable prototypes – Overcontrolled and Undercontrolled – appear in most studies of the prototypes (Asendorpf, Borkenau, Ostendorf, & Van Aken, 2001). Individuals with these characteristics are more likely than resilient individuals to experience problems with emotional adjustment and in interpersonal and social relationships, and social adjustment. For purposes of the present study we focus on the characteristics of the resilient prototype hypothesized to facilitate adjustment compared to those who are not resilient, and to isolate potentially malleable mechanisms that distinguish between resilient and non-resilient veterans (e.g., coping, psychological flexibility, social support).

A two-step cluster analysis was conducted to decide the number of personality types (clusters). The clusters were created from the Neuroticism, Extraversion, Openness to Experience, Agreeableness, and Conscientiousness subscale scores. Given 264 participants, the candidate number of clusters ranged from 1 (totally homogenous) to 264 (totally heterogeneous). The Ward’s method along with squared Euclidean distance measures were used to calculate the total within-cluster sums of squares for each cluster solution. The largest change in total within-cluster sums of squares happened when the number of clusters increased from 1 to 2. After the 2-cluster solution, the total within-cluster sums of squares decreased mildly. Hence, we decided on a 2-cluster result in step 1. The K-means analysis was employed to optimize the cluster classification by utilizing the initial cluster centers from step 1. The level of agreement was substantial: Cohen’s Kappa coefficient was .77 (Landis & Koch, 1977). The personality prototype served as an independent variable. Consistent with past research, the resilient prototype (n = 138) was characterized by the lower Neuroticism value, and higher values on the other four scales (see Figure 1). The non-resilient prototype (n = 126) had an elevated Neuroticism value and lower values on the other scales.

Figure 1.

Resilient (n=138) and non-resilient (n=126) personality prototypes among veterans based on NEO-FFI z scores.

Traumatic Brain Injury.

The structured interview developed by Vasterling and colleagues (Alosco et al., 2016) was used to assess the number, recency, type, and clinical sequelae associated with TBI experienced at any time. A positive screen was determined by a head injury that resulted in a loss of consciousness (LOC), an alteration of consciousness (e.g., dazed, confused, “seeing stars”), or posttraumatic amnesia (PTA). From this information we determined (a) the number of lifetime TBIs reported by the veteran, the number of TBIs incurred during post-9/11 warzone deployment, and the longest periods of LOC and PTA. The number of TBIs assessed in detail was limited to five. Length of LOC and PTA were rated for each TBI. Mild TBIs were determined by a LOC lasting less than 30 minutes and/or PTA lasting less than 24 hours (American Congress of Rehabilitation Medicine, 1993).

We coded the longest experience of LOC and PTA reported by each participant for our analyses. LOC was coded as either 0 (no LOC), 1 (less than 1 minute), 2 (1 to 15 minutes), or 3 (16 to 30 minutes). PTA was coded as 0 (no PTA occurred), 1 (less than 1 hour of PTA), or 2 (1 to 24 hours of PTA). A total of 164 (62%) veterans screened positive for at least one lifetime mTBI, and 118 (45%) screened positive for at least one mTBI during deployment (e.g., from a blast, vehicular accident, fall, bullet, fragment).

Covariates

Combat Exposure.

The 18-item Full Combat Experiences Scale (FCES; Hoge et al., 2004) was administered at baseline to assess level of combat exposure. The items describe combat events relevant to the Iraq and Afghanistan campaigns (e.g., being attacked or ambushed, handling human remains, incoming mortars). Items are rated on a 5-point scale, ranging from 0 (never) to 4 (10 or more times). The total score was used. Internal consistency was α = .93.

Social Desirability.

We included a ten-item version of the Marlowe-Crowne Social Desirability Scale (MC; Strahan & Gerbasi, 1972) at baseline to assess social desirability response tendencies. Each item was rated as 1 (true) or 0 (false). Higher scores indicate a greater tendency to respond in a socially desirable fashion. Recent evidence indicates the Marlowe-Crowne scale performs well compared to contemporary methods in detecting “faking bad” and “faking good” (Lambert, Arbuckle, & Holden, 2016). Internal consistency was α = .61.

Mediating Variables

Psychological flexibility.

The 7-item Acceptance and Action Questionnaire - II (AAQ–II; Bond et al., 2011) assessed psychological flexibility at baseline. Psychological flexibility reflects the ability to which a person may experience unwanted internal experiences without trying to change them, and commit to context-appropriate behaviors consistent with personal values (Hayes, Luoma, Bond, Masuda, & Lillis, 2006). Items are rated on a 7-point Likert-type scale ranging from 1 (never true) to 7 (always true). The AAQ–II has a single-factor structure and demonstrates acceptable internal consistency, test–retest reliability, and convergent associations with the original AAQ (Bond et al., 2011). Lower total scores indicate greater psychological flexibility, and predict greater PTSD symptom severity after accounting for established pre-, peri-, and post-trauma risk factors (Meyer et al., in press). Internal consistency was .93.

Coping.

The 28-item Brief COPE (B-COPE, Carver, 1997) assessed coping styles at baseline. Respondents are instructed to rate each item to indicate what they generally do and feel in stressful situations on a 4-point Likert-type scale that ranges from 0 (I usually don’t do this) to 3 (I usually do this a lot). A factor analysis of the B-COPE (Grosso et al., 2014) revealed two factor scores that we used for analysis: action-oriented and avoidant coping. Higher scores reflect a greater proclivity for that coping style. Internal coefficients were acceptable (.81, action-oriented coping; .70, avoidant coping).

Social support.

The 15-item Deployment Risk and Resilience Inventory Postdeployment Social Support scale (PDSS; King, King, Vogt, Knight, & Samper, 2006) assessed perceived availability of social support at baseline. Items are rated on a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). Validated for use with Iraq/Afghanistan veterans (Vogt, Proctor, King, King, & Vasterling, 2008), the PDSS measures the extent to which the respondent believes their friends and family provide emotional sustenance and instrumental assistance, and the extent to which they feel respected for their military service. Higher total scores indicate greater social support. Internal consistency was .87.

Self-reported resilience.

The Brief Resilience Scale (BRS; Smith et al., 2008) was administered at baseline. This six-item instrument assesses a unitary construct that signifies the ability to “…bounce back or recover from stress” (Smith et al., 2008; p. 199). Respondents are instructed to select how much they agree with each item on a Likert-type scale that ranges from 1 (strongly disagree) to 5 (strongly agree). Higher total scores reflect greater resilience. The BRS is significantly associated with other self-report measures of resilience (Windle, Bennett, & Noyes, 2011), but also appears to account for unique variance in adjustment apart from these other measures (Smith et al., 2008). Internal consistency was .89.

Outcome Variables

PTSD symptoms.

The 17-item PTSD Checklist–Military Version (PCL–M; Weathers, Litz, Herman, Huska, & Keane, 1993) was administered at each measurement occasion to assess military-related PTSD symptoms during the past month according to DSM-IV criteria. Respondents indicate the degree to which they have experienced each symptom on a scale from 1 (not at all) to 5 (extremely). The PCL-M has excellent internal consistency and validity (Blanchard, Jones-Alexander, Buckley, & Forneris, 1996; Forbes, Creamer, & Biddle, 2001). Internal consistency coefficients ranged from .96 to .97.

Depression symptoms.

The 21-item Beck Depression Inventory–II (BDI–II; Beck, Steer, & Brown, 1996) was administered at each time-point. Respondents rate on a scale that ranges from 0 (does not experience the symptom) to 3 (severe experience of the symptom) the symptoms they have experienced over the past two weeks. The total score was used. Internal consistency ranged from .95 to .96.

Quality of Life.

The 16-item Quality of Life Scale (QLS; Burckhardt, Woods, Schultz, & Ziebarth, 1989) assessed subjective quality of life/life satisfaction at each measurement occasion. Respondents are instructed to rate their degree of satisfaction at the present time with an activity or a relationship on a scale that ranges from 1 (delighted) to 7 (terrible). The QLS exhibits good reliability and validity (Burckhardt & Anderson, 2003). Internal consistency ranged from .93 to .95.

Functional Disability.

The 36-item self-report version of the World Health Organization Disability Assessment Scale 2.0 (WHODAS; Üstün, 2010; Üstün et al., 2010) assessed functional disability at each time-point. Behaviorally anchored items, rated on a scale from 0 (none) to 4 (extreme/cannot do), assess function across 7 domains (understanding and communicating, getting around, getting along with people, life activities, work and school, participation in society, self-care) over the past 30 days. Respondents who have not worked or attending school in the past 30 days are instructed to skip those items. To obtain a single indicator of overall functional disability for all respondents, we computed an average item response score for subsequent analyses. Internal consistency ranged from .96 to .97.

Statistical Analyses

Preliminary analyses.

The associations among personality type (resilient vs. non-resilient), gender, race, and service-connected disability variables were estimated. Differences between personality prototypes on all other study variables were examined (Table 1) using SPSS 22.0.

Table 1.

Means and Standard Deviations Among Study Variables by Personality Prototypes (N = 264)

	Personality Prototype
	Resilient (n = 138)		Non-Resilient (n = 126)
Measure	M	SD	M	SD	p
TBI
LOC	0.48	0.82	0.55	0.84	.519
PTA	0.17	0.48	0.20	0.51	.663
NTBI-Lifetime	1.23	1.36	1.53	1.57	.096
NTBI (warzone)	0.65	1.09	0.95	1.26	.034
Mediators
BRS	23.59	3.66	17.12	4.07	<.001
PDSS	58.64	10.03	49.58	9.55	<.001
B-COPE Action	10.57	4.26	9.58	3.37	.042
B-COPE Avoidant	1.59	1.63	3.42	2.09	<.001
AAQ-II	13.27	6.63	25.67	8.26	<.001
Covariates (Baseline)
MC	6.39	1.89	5.18	2.25	<.001
FCES	17.05	13.98	20.36	14.58	.068
Outcomes (Baseline)
PCL-M	27.94	14.21	46.70	15.21	<.001
BDI-II	6.85	8.22	22.46	15.21	<.001
QLS	86.87	13.25	65.98	15.76	<.001
WHODAS 2.0	17.43	17.86	45.25	24.09	<.001
Outcomes (4-months)
PCL-M	28.48	13.46	48.73	17.72	<.001
BDI-II	9.81	9.61	25.81	13.28	<.001
QLS	85.11	17.29	60.01	15.40	<.001
WHODAS 2.0	21.63	20.98	52.93	25.50	<.001
Outcomes (8-months)
PCL-M	30.95	16.06	50.70	17.61	<.001
BDI-II	10.71	10.77	26.72	13.45	<.001
QLS	83.39	17.65	61.57	17.67	<.001
WHODAS 2.0	24.14	22.95	54.31	25.78	<.001
Outcomes (12-months)
PCL-M	27.45	13.59	44.90	17.55	<.001
BDI-II	7.81	9.74	21.26	12.52	<.001
QLS	86.72	15.75	65.65	15.67	<.001
WHODAS 2.0	19.03	20.70	44.94	26.52	<.001

Note. WHODAS 2.0 scores are the sum of all items. Mean scores differences were examined by conducting independent sample t-tests. LOC = Worst loss of consciousness rating coded as 0 (no LOC), 1 (less than 1 minute), 2 (1 to 15 minutes), or 3 (16 to 30 minutes); PTA = Worst posttraumatic amnesia rating coded as 0 (no PTA occurred), 1 (less than 1 hour of PTA), or 2 (1 to 24 hours of PTA; NTBI = Lifetime number of traumatic brain injuries including civilian and military; NTBI (warzone) = Number of TBIs during post-9/11 warzone deployment; BRS = Brief Resilience Scale; PDSS = Post-deployment Social Support Scale; B-COPE Action = Brief COPE Active coping factor score; B-COPE Avoidant = Brief COPE Avoidant coping factor score; AAQ-II = Acceptance and Action Questionnaire-II; MC = Marlowe-Crowne Social Desirability Scale; FCES = Full Combat Exposure Scale; PCL-M = PTSD checklist-military version; BDI-II = Beck Depression Inventory-II; QLS = Quality of Life Scale; WHODAS = World Health Organization Disability Assessment Schedule 2.0.

Path Model.

Path analyses under the structural equation modeling (SEM) framework were utilized to test the hypothesized model (Figure 2) using Mplus 7.1 (Muthén & Muthén, 1998–2012). Personality prototypes and the TBI variables were positioned as exogenous variables to directly predict the four outcome variables (PCL-M, BDI-II, QLS, and WHODAS) across time-points (baseline, 4-, 8-, and 12-months). The effects from the predictors to the outcomes were assumed to be mediated by social support (PDSS), action-oriented and avoidant coping (Brief COPE), psychological flexibility (AAQ-II), and self-reported resilience (BRS). Combat exposure and social desirability were treated as covariates and were assumed to have direct effects on all other study variables. Mediators were allowed to correlate with each other. Outcomes were allowed to correlated with each other over time.

Figure 2.

Saturated a priori model of personality prototypes, TBI and mediating characteristics predicting adjustment over 12-months.

Note: The covariates were assumed to have direct effects on TBI variables, personality type, mediators and outcomes. Mediators were allowed to correlate with each other. Outcomes were allowed to correlated with each other over time. FCES = Full Combat Exposure Scale; MC = Marlowe-Crowne Social Desirability Scale; NTBI = Lifetime number of traumatic brain injuries including civilian and military; PTA = Worst posttraumatic amnesia rating; LOC = Worst loss of consciousness rating; NTBI (warzone) = Number of TBIs during post-9/11 warzone deployment; BRS = Brief Resilience Scale; PDSS = Postdeployment Social Support Scale; AAQ-II = Acceptance and Action Questionnaire-II; B-COPE Action = Brief COPE Active coping factor score; B-COPE Avoidant = Brief COPE Avoidant coping factor score; PCL-M = PTSD Checklist-Military Version; BDI-II = Beck Depression Inventory-II; QLS = Quality of Life Scale; WHODAS = World Health Organization Disability Assessment Scale 2.0.

The path model was evaluated with the following criteria: χ² test ≥ .05, comparative fit index (CFI) and Tucker-Lewis index (TLI) ≥ .95 (Hu & Bentler, 1999; Yu, 2002), root-mean-square error of approximation (RMSEA) ≤ .08 (Browne & Cudeck, 1992), and standardized root-mean-square residual (SRMR) ≤ .08 (Hu & Bentler, 1999). Missing data were accommodated with the maximum likelihood estimation method to utilize all available information. The overall R² values were estimated for each outcome across time. Effect sizes were interpreted using the cutoffs of .02, .15, and .35 as small, medium, and large, respectively (Cohen, 1995). Direct and indirect effects were examined, and 1000 bootstrapping samples were produced to estimate the 95% confidence interval for the indirect effects (Preacher & Hayes, 2008).

Results

Ninety percent of the participants completed the assessments at each measurement occasion. Preliminary χ² tests of independence found that personality type was not associated with gender [χ²(1) = 1.36, ns], race/ethnicity [χ²(3) = 2.89, ns], or receipt of service-connected disability [χ²(1) = 2.68, ns].

Independent sample t-test results showed that personality prototypes were associated with number of warzone-related mTBIs (p < .05): Veterans with a non-resilient prototype had more mTBIs. All mediators were associated with personality prototype (p values < .05). Veterans with a resilient prototype reported more social support, more action-oriented coping, greater psychological flexibility, higher self-perceived resilience and less avoidant coping. The resilient group also had significantly higher social desirability scores. Combat exposure was not significantly associated with personality prototype. Personality prototypes were associated with the four outcome variables at each time-point (p values < .001). Veterans with a resilient prototype reported lower PTSD and depression symptoms and functional disability, and higher quality of life than veterans with a non-resilient prototype. The average BDI-II scores for the non-resilient group were indicative of “moderate depression” at every time-point (scores of 20 to 28; Beck et al., 1996). Similarly, the average PCL-M scores for the non-resilient group approached the recommended cut-off score of 50 that is often used to indicate a positive screen for PTSD among military personnel (and consistently exceeded a cut-off score of 38 that has good sensitivity with clinician diagnoses of PTSD among Iraq/Afghanistan veterans; Tsai, Pietrzak, Hoff, & Harpaz-Rotem, 2016). While firm interpretive guidelines for the WHODAS 2.0 in this population are needed (Konecky, Meyer, Marx, Kimbrel, Morissette, 2014), one recent study that combined several independent veteran samples identified a cutoff score on the WHODAS 2.0 that may differentiate between veterans with and without PTSD-related functional impairment (Bovin et al., 2019). The non-resilient and resilient groups in the current study scored well above and below this cutoff score, respectively, providing some evidence that these groups exhibited clinically meaningful differences in functional disability.

Path Analysis

In the saturated model, two direct effects -- total number of lifetime mTBIs to self-reported resilience and total number of lifetime mTBIs to active coping -- were non-significant and had effect sizes close to zero. Hence, these two paths were fixed to zero in the corrected model. The corrected model had excellent fit: χ²(2) = 0.019, p = 0.990; CFI = 1.00; TLI = 1.00; RMSEA = .000 (90% CI: [.000, .000]); SRMR = .001.

Direct effects.

We analyzed three types of direct effects: from predictors to mediators; from mediators to outcome variables; from predictors to outcome variables, as well as the indirect effects from predictors to outcomes through mediators. All effects accounted for combat exposure and social desirability and are described using standardized path coefficients. Statistically significant (p < .05) direct effects are depicted in Figures 3a to 3d for the four outcome variables.

Figure 3a.

Path model of personality prototype, traumatic brain injury variables, and mediating characteristics predicting PTSD Checklist-Military Version (PCL-M) scores across four time points. R² = 52.4% – 72.4%

Note: Only statistically significant direct effects denoted by using standardized path coefficients were shown. The covariates were assumed to have direct effects on TBI variables, personality type, mediators and outcomes. Mediators were allowed to correlate with each other. Outcomes were allowed to correlate with each other over time. FCES = Full Combat Exposure Scale; MC = Marlowe-Crowne Social Desirability Scale; NTBI = Lifetime number of traumatic brain injuries including civilian and military; PTA = Worst posttraumatic amnesia rating; LOC = Worst loss of consciousness rating; NTBI (warzone) = Number of TBIs during post-9/11 warzone deployment; BRS = Brief Resilience Scale; PDSS = Postdeployment Social Support Scale; AAQ-II = Acceptance and Action Questionnaire-II; B-COPE Action = Brief COPE Active coping factor score; B-COPE Avoidant = Brief COPE Avoidant coping factor score.

Figure 3d.

Path model of personality prototype, traumatic brain injury variables, and mediating characteristics predicting World Health Organization Disability Assessment Schedule 2.0 (WHODAS) scores across four time points. R² = 40.8% – 58.1%

Note: Only statistically significant direct effects denoted by using standardized path coefficients are shown. The covariates were assumed to have direct effects on TBI variables, personality type, mediators and outcomes. Mediators were allowed to correlate with each other. Outcomes were allowed to correlate with each other over time. FCES = Full Combat Exposure Scale; MC = Marlowe-Crowne Social Desirability Scale; NTBI = Lifetime number of traumatic brain injuries including civilian and military; PTA = Worst posttraumatic amnesia rating; LOC = Worst loss of consciousness rating; NTBI (warzone) = Number of TBIs during post-9/11 warzone deployment; BRS = Brief Resilience Scale; PDSS = Postdeployment Social Support Scale; AAQ-II = Acceptance and Action Questionnaire-II; B-COPE Action = Brief COPE Active coping factor score; B-COPE Avoidant = Brief COPE Avoidant coping factor score.

Personality prototype was associated with all mediators except action-oriented coping. Compared with the resilient group, the non-resilient prototype was associated with lower self-reported resilience, social support, and psychological flexibility, and with greater avoidant coping (all p’s < .05). In contrast, only two significant effects were observed among the TBI variables: the number of warzone-related mTBIs was negatively associated with psychological flexibility and social support (p’s < .05).

The resilient personality prototype predicted lower PTSD and depression symptoms and functional disability and higher quality of life over time (p’s < .05). The number of lifetime TBIs, and the worst PTA and LOC ratings were not associated with the four outcome variables at any time with two exceptions: the worst PTA predicted 8-month PTSD and the worst LOC predicted baseline QOL. The number of warzone-related mTBIs was associated with greater functional disability at each time point.

Several mediators were directly predictive of outcome variables. As depicted in Figure 3a, psychological flexibility was associated with lower PTSD symptoms at all four time-points (p’s < .01) and social support was associated with lower PTSD at three time-points (p’s < .01). Avoidant coping was associated with greater PTSD at three time-points (p’s < .05). Psychological flexibility was consistently negatively associated with depression over time, and avoidant coping was consistently positively associated with depression over time (p’s < .01; Figure 3b). Social support was related to lower depression at baseline and 4-months (p’s < .05). Higher self-reported resilience (as measured by the BRS) and psychological flexibility predicted higher quality of life across time (p’s < .05; Figure 3c). Social support was positively associated with quality of life at baseline, 4-, and 12-month follow-up (p’s < .05). Psychological inflexibility also predicted greater functional disability at all time-points (p’s < .05; Figure 3d). Lower social support predicted greater functional disability at baseline, 4-, and 12-months.

Figure 3b.

Path model of personality prototype, traumatic brain injury variables, and mediating characteristics predicting Beck Depression Inventory-II (BDI-II) scores across four time points. R² = 47.4% – 68.7%

Note: Only statistically significant direct effects denoted by using standardized path coefficients are shown. The covariates were assumed to have direct effects on TBI variables, personality type, mediators and outcomes. Mediators were allowed to correlate with each other. Outcomes were allowed to correlate with each other over time. FCES = Full Combat Exposure Scale; MC = Marlowe-Crowne Social Desirability Scale; NTBI = Lifetime number of traumatic brain injuries including civilian and military; PTA = Worst posttraumatic amnesia rating; LOC = Worst loss of consciousness rating; NTBI (warzone) = Number of TBIs during post-9/11 warzone deployment; BRS = Brief Resilience Scale; PDSS = Postdeployment Social Support Scale; AAQ-II = Acceptance and Action Questionnaire-II; B-COPE Action = Brief COPE Active coping factor score; B-COPE Avoidant = Brief COPE Avoidant coping factor score.

Figure 3c.

Path model of personality prototype, traumatic brain injury variables, and mediating characteristics predicting Quality of Life Scale (QLS) scores across four time points. R² = 39.2% – 59.3%

Note: Only statistically significant direct effects denoted by using standardized path coefficients are shown. The covariates were assumed to have direct effects on TBI variables, personality type, mediators and outcomes. Mediators were allowed to correlate with each other. Outcomes were allowed to correlate with each other over time. FCES = Full Combat Exposure Scale; MC = Marlowe-Crowne Social Desirability Scale; NTBI = Lifetime number of traumatic brain injuries including civilian and military; PTA = Worst posttraumatic amnesia rating; LOC = Worst loss of consciousness rating; NTBI (warzone) = Number of TBIs during post-9/11 warzone deployment; BRS = Brief Resilience Scale; PDSS = Postdeployment Social Support Scale; AAQ-II = Acceptance and Action Questionnaire-II; B-COPE Action = Brief COPE Active coping factor score; B-COPE Avoidant = Brief COPE Avoidant coping factor score.

Indirect effects.

Psychological flexibility partially mediated the association between personality prototype and PTSD across time (p’s < .001). Social support and avoidant coping each partially mediated the association between personality and PTSD at three time-points (p’s < .01). Compared with the resilient prototype, the non-resilient group had higher PTSD scores through lower psychological inflexibility, lower social support, and greater use of avoidant coping. Number of warzone-related mTBIs was associated with higher PTSD through its association with psychological inflexibility across time (p’s < .05).

The association between the non-resilient prototype and higher depression was partially mediated by both avoidant coping and psychological inflexibility across time (all p’s < .01). Higher depression scores in the non-resilient group were also partially mediated by lower social support at baseline and 4-months (both p’s < .05). Psychological inflexibility mediated the effect of number of warzone-related mTBI on greater depression scores at each time point (all p’s < .05).

Partial mediation of psychological flexibility on the personality-quality of life relationship occurred at all time-points (p’s < .05), and by social support and self-reported resilience at three time-points (p’s < .05). A resilient prototype predicted higher QOL through its beneficial associations with psychological flexibility, higher social support and greater self-perceptions of personal resilience. Lower psychological flexibility mediated the relationship between greater number of warzone-related mTBIs and lower QOL at baseline and 4-months (p’s < .05).

Psychological inflexibility also partially mediated the personality-functional disability relationship at every measurement occasion (p’s < .01). Social support partially mediated this relationship at three time-points (p’s < .05). The non-resilient prototype reported greater disability through lower psychological flexibility and social support. Avoidant coping partially mediated the effect of personality prototype on functional disability at baseline and 12-months (p’s < .01). Similarly, greater psychological inflexibility partially mediated the effects of a greater number of warzone-related mTBIs on greater functional disability across time-points (all p’s < .05).

R² values indicative of large effect sizes were observed for each outcome variable in the final model. The model accounted for 52% to 72% of the variance in PTSD across each measurement, 47% to 69% of the variance in depression across time, 39% to 59% of the variance in QOL, and 41% to 58% of the variance in functional disability.

Discussion

Consistent with the Block conceptualization of ego resilience and control, veterans with a resilient personality prototype exhibited fewer problems with PTSD, depression, and functional disability, and reported greater QOL over the course of a year, compared to veterans who were not resilient. This “person-centered” approach to personality research provides support for studying “normal,” non-pathological personality factors in veteran adjustment, as our findings reveal valuable insights into several behavioral mechanisms that account for differences in veteran adjustment. Consistent with our theoretical expectations, resilience facilitated adjustment on all outcome variables at all time points through its beneficial association with psychological flexibility. Indirect effects through higher social support and less avoidant coping and in a few cases, with a measure of perceived abilities to “bounce back” in time of stress, reflect characteristics that typify the resilient prototype: Enduring capacities for personal and social relationships and self-regulation that foster a sense of confidence, self-esteem, and self-efficacy (Farkas & Orosz, 2015).

The consistency of the direct effects from the resilient prototype to each outcome variable independent of the mediating variables suggest that other mechanisms associated with the prototype may also facilitate veteran adjustment (e.g., capacities for positive emotions, active engagement with the environment). These particular mechanisms are part of the constellation of adaptive qualities associated with ego resiliency and control that are measured, in part, by the separate traits that constitute the resilient prototype. Higher Extraversion is associated with sociability and positive emotions, and higher Conscientiousness has been associated with active engagement and goal-directed activity among persons with disabling conditions (Elliott et al., in press). These two traits are also associated with the higher-order personality factors of stability (Conscientiousness) and plasticity (Extraversion) that are represented in the resilient prototype (DeYoung, 2006; Farkas & Orosz, 2015).

Moreover, the results of this study indicate that the effects of resilience, social support, and psychological flexibility on optimal adjustment occurred regardless of mTBI history. Of the TBI variables in the model, the number of warzone-related mTBIs had a direct effect on functional disability, and its association with PTSD, depression and functional disability was mediated by its relationship with psychological inflexibility. Although research has found multiple mTBIs to be associated with veteran distress (Dretsch et al., 2015; Spira, Lathan, Bleiberg, & Tsao, 2014), critical observers argue that the research concerning the “cumulative effects of multiple mTBIs” is equivocal (Rohling et al., 2017; p. 160). Our findings do not resolve these issues, but perhaps imply that the negative effects of multiple mTBIs may appear when studied in the context of additional factors such as combat exposure, personality characteristics, and specific behaviors. Thus, our results may reflect the “burden of adversity” in which multiple factors, including premorbid factors, personal history (including combat exposure), and comorbidities, affect symptom reporting among veterans with deployment-related mTBI and PTSD (Brenner, Vanderploeg, & Terrio, 2009).

Brenner et al. (2009) recommended the use of evidence-based interventions to treat veteran distress regardless of etiology, and recent work suggests that cognitive-behavioral interventions may be more effective in alleviating post-concussive symptoms associated with mTBI than cognitive rehabilitation strategies (Vanderploeg, Belanger, Curtiss, Bowles, & Cooper, 2019). Our results suggest that there are pathways common to both personality prototypes and mTBIs incurred during warzone-deployment that impact symptom reports of PTSD, depression, and functional impairment. Of particular interest is the critical role of psychological flexibility in our model as a significant mediator and predictor of veteran adjustment. Arguably a transdiagnostic concept relevant to an array of psychopathological conditions, psychological flexibility is a central mechanism of change in Acceptance and Commitment therapy (ACT; Hayes et al., 2006), and its properties parallel those associated with ego-resiliency (Kashdan & Rottenberg, 2010). Rather than focus on symptom reduction, ACT intends to increase psychological flexibility by helping a person to remain present in the moment, accept uncomfortable, distressing internal and external experiences, and persist in activities and behaviors consistent with their values and goals, without avoiding the internal stimuli or the contextual demands (Hayes et al., 2006). The theoretical properties of psychological flexibility and its pivotal role in ACT distinguish it from cognitive flexibility, a component of executive function recognized in neuropsychological research and in TBI rehabilitation (Whiting, Deane, Simpson, McLeod, & Ciarrochi, 2017). Moreover, the evidence to date – including the results of the present study – suggests that it warrants consideration in psychological interventions with veterans with mTBI specifically, (Whiting et al., 2017), and polytrauma, generally (e.g., Cook et al., 2015). Clinical interventions that address interpersonal skills and social competence may prevent a downward spiral in relationship quality and well-being (Diener et al., 2017) and in the process, promote social and interpersonal ties and corresponding experiences of positive emotion that characterize personal resilience (Kok & Fredrickson, 2014).

There are several limitations to this study. The sample was predominately male and Caucasian, and enrolled in a specific VA healthcare system, which may limit generalizability. Predeployment personality data were not available; we cannot readily dismiss concerns about possible personality changes following warzone service. The indicators of PTSD and depression reflect symptom reports, and these should not be interpreted as actual diagnostic conditions. Interpretive guidelines for use of the WHODAS 2.0 are needed across populations, including in war veterans, to facilitate making strong inferences about the relative degree of functional impairment reported by the non-resilient group (Konecky et al., 2014). Our model included mediators that were assessed at the time we obtained personality data. Consequently, the observed relationships among the variables assessed at baseline reflect atemporal associations, precluding strong inferences about mediation (Winer et al., 2016).

Our reliance on the “person-centered” use of personality prototypes varies from the traditional trait-based approach, and clustering prototypes can, at times, result in sample-specific configurations (Donnellan & Robins, 2010). To a great extent, this issue was circumvented by restricting our theoretical and clinical focus to two prototypes -- resilient and non-resilient – rather than the three that are usually studied in the relevant literature. Controversies about the incremental and predictive validity of the prototypes versus the traditional use of specific traits are contentious and unresolved. In the current study, we were invested in a theoretical model from which we could conceptualize, anticipate, and test relationships that would have implications for psychological interventions. In contrast, the use of isolated traits in clinical research sometimes provides “…little information about the internal psychological mechanisms…that could be the target of change” in psychological interventions (Shadel, 2010; p. 336). In addition, the person-centered approach that guided our study helped us avoid the tautology that plagues the use of self-report, face-valid measures of resilience in studies of distress. Longitudinal research finds that responses to these instruments fluctuate with changes in emotional well-being and social support over time (positive or negative; Laird et al., in press) and the predictive value of these instruments diminishes in contextual models that account for stable personality factors and other behavioral and social characteristics (Elliott et al., 2015).

Our findings provide additional evidence that a resilient personality prototype serves a protective function, endowing a veteran with several adaptive characteristics that facilitate their adjustment. Veterans who have personality characteristics associated with a non-resilient prototype lack personal and social resources that, in turn, undermines their psychological adjustment and quality of life, and contributes to impairment in everyday roles and activities. Additional research is needed to demonstrate the effectiveness of interventions that target the mechanisms of change identified in this study – psychological flexibility and social support, in particular – in promoting veteran resilience.

Table 2.

Standardized Path Coefficients and 95% Confidence Interval (CI) of Indirect Effects of Personality Subtypes and Number of mTBIs Incurred During Deployment on the Outcome Variables Over Time

	Baseline		4-months		8-months		12-months
	Estimate	CI	Estimate	CI	Estimate	CI	Estimate	CI
	PCL-M
Non-Resilient → BRS →	−0.08	[−0.20, 0.05]	0.07	[−0.04, 0.19]	0.06	[−0.03, 0.15]	0.05	[−0.05, 0.14]
Non-Resilient → PDSS →	0.02	[−0.07, 0.11]	0.07*	[0.01, 0.13]	0.05*	[0.00, 0.10]	0.05*	[0.00, 0.10]
Non-Resilient → B-COPE Action →	0.02	[−0.01, 0.04]	−0.01	[−0.02, 0.01]	0.00	[−0.01, 0.02]	−0.01	[−0.03, 0.01]
Non-Resilient → B-COPE Avoidant →	0.09*	[0.02, 0.17]	0.02	[−0.04, 0.08]	0.05*	[0.00, 0.11]	0.06*	[0.01, 0.11]
Non-Resilient → AAQ-II →	0.32**	[0.18, 0.45]	0.18**	[0.08, 0.28]	0.17**	[0.08, 0.26]	0.19**	[0.09, 0.28]
NTBI(warzone) → PDSS →	0.01	[−0.03, 0.04]	0.03	[−0.01, 0.07]	0.02	[−0.01, 0.05]	0.02	[−0.01, 0.05]
NTBI(warzone)→ AAQ-II →	0.09*	[0.01, 0.17]	0.05*	[0.00, 0.10]	0.04*	[0.00, 0.09]	0.05*	[0.00, 0.10]
	BDI-II
Non-Resilient → BRS →	0.01	[−0.07, 0.09]	0.07	[−0.02, 0.15]	0.08	[−0.02, 0.17]	0.03	[−0.08, 0.13]
Non-Resilient → PDSS →	0.05*	[0.01, 0.09]	0.06*	[0.01, 0.11]	0.02	[−0.03, 0.07]	0.05	[−0.01, 0.10]
Non-Resilient → B-COPE Action →	0.01	[−0.01, 0.03]	0.01	[−0.01, 0.02]	0.01	[−0.01, 0.03]	0.00	[−0.01, 0.02]
Non-Resilient → B-COPE Avoidant →	0.09**	[0.04, 0.14]	0.07*	[0.02, 0.12]	0.07*	[0.01, 0.13]	0.09*	[0.03, 0.16]
Non-Resilient → AAQ-II →	0.27**	[0.19, 0.36]	0.20**	[0.12, 0.29]	0.16**	[0.07, 0.25]	0.17**	[0.06, 0.27]
NTBI(warzone) → PDSS →	0.02	[−0.01, 0.04]	0.02	[−0.01, 0.05]	0.01	[−0.02, 0.03]	0.02	[−0.01, 0.05]
NTBI(warzone) → AAQ-II →	0.08*	[0.00, 0.15]	0.06*	[0.00, 0.11]	0.04*	[0.00, 0.09]	0.05*	[0.00, 0.10]
	QLS
Non-Resilient → BRS →	−0.09*	[−0.16, −0.02]	−0.10*	[−0.18, −0.02]	−0.09	[−0.19, 0.00]	−0.11*	[−0.22, −0.01]
Non-Resilient → PDSS →	−0.07*	[−0.12, −0.02]	−0.08*	[−0.13, −0.02]	−0.03	[−0.09, 0.03]	−0.06*	[−0.13, −0.00]
Non-Resilient → B-COPE Action →	−0.02	[−0.04, 0.01]	−0.01	[−0.03, 0.01]	−0.01	[−0.03, 0.01]	−0.01	[−0.02, 0.01]
Non-Resilient → B-COPE Avoidant →	−0.02	[−0.07, 0.02]	−0.04	[−0.08, 0.01]	−0.04	[−0.09, 0.02]	−0.04	[−0.09, 0.02]
Non-Resilient → AAQ-II →	−0.23**	[−0.32, −0.15]	−0.15**	[−0.22, −0.07]	−0.09*	[−0.19, −0.01]	−0.10*	[−0.20, −0.00]
NTBI(warzone) → PDSS →	−0.03	[−0.06, 0.01]	−0.03	[−0.07, 0.01]	−0.01	[−0.04, 0.02]	−0.02	[−0.06, 0.02]
NTBI(warzone) → AAQ-II →	−0.07*	[−0.13, −0.00]	−0.04*	[−0.08, −0.00]	−0.03	[−0.06, −0.0H	−0.03	[−0.07, 0.0H
	WHODAS
Non-Resilient → BRS →	0.08*	[0.00, 0.17]	0.08	[−0.01, 0.17]	0.08	[−0.02, 0.19]	0.01	[−0.08, 0.09]
Non-Resilient → PDSS →	0.05*	[0.00, 0.10]	0.07*	[0.02, 0.13]	0.03	[−0.03, 0.10]	0.05*	[0.00, 0.11]
Non-Resilient → B-COPE Action →	0.01	[−0.01, 0.02]	0.01	[−0.01, 0.03]	0.00	[−0.02, 0.02]	0.01	[−0.01, 0.02]
Non-Resilient → B-COPE Avoidant →	0.08*	[0.03, 0.13]	0.04	[−0.02, 0.09]	0.03	[−0.02, 0.09]	0.11**	[0.04, 0.17]
Non-Resilient → AAQ-II →	0.16**	[0.08, 0.24]	0.16**	[0.07, 0.26]	0.12*	[0.02, 0.23]	0.16**	[0.06, 0.26]
NTBI(warzone) → PDSS →	0.02	[−0.01, 0.05]	0.03	[−0.01, 0.06]	0.01	[−0.02, 0.04]	0.02	[−0.02, 0.06]
NTBI(warzone) → AAQ-II →	0.04*	[0.00, 0.09]	0.05*	[0.00, 0.09]	0.03*	[0.00, 0.07]	0.04*	[0.00, 0.09]

Note. Resilient personality prototype is the reference group. PCL-M = PTSD Checklist-Military Version; BDI-II = Beck Depression Inventory-II; QLS = Quality of Life Scale; WHODAS = World Health Organization Disability Assessment Schedule 2.0; BRS = Brief Resilience Scale; PDSS = Post-deployment Social Support Scale; AAQ-II = Acceptance and Action Questionnaire-II; NTBI (warzone) = Number of TBIs during post-9/11 warzone deployment.

^*p < .05,

^**p < .001

Impact.

• The present study indicates that the beneficial effects of a resilient personality prototype occur among warzone veterans with and without mTBI.

• Several of the characteristics associated with a resilient personality prototype can be addressed in psychological interventions to facilitate psychological adjustment of veterans who are experiencing difficulties with depression, post-traumatic stress symptoms, and quality of life.

• Of these characteristics, psychological flexibility appears to be a particularly important mechanism through which resilience promotes adjustment, and psychological flexibility is a central element of Acceptance and Commitment Therapy.