Factors Associated with Employment and Work Perception in Combat-Exposed Veterans

Abstract

Purpose/Objective:

The purpose of this study was twofold: (1) to evaluate whether perception of work ability is associated with employment status in a sample of combat-exposed Veterans, and (2) to determine whether the same sets of variables that are associated with employment status are also associated with perception of work ability.

Research Method/Design:

In this cross-sectional study, Veterans (N=83) underwent a neuropsychological assessment and completed questionnaires assessing demographic characteristics, combat-related experiences, and psychiatric and neurobehavioral/health-related symptoms. Primary outcomes of interest were employment status (unemployed vs. employed) and Veterans’ perception of whether their ability to work has declined due to an ongoing condition (yes vs. no).

Results:

A chi-square analysis revealed a significant relationship between perception of work ability and employment status. Additionally, psychiatric and neurobehavioral/health-related symptoms were associated with employment status and perception of work ability, whereas demographic characteristics (i.e., service-connected disability rating) and combat-related experiences (i.e., mTBI history) were only associated with perception of work ability. Objective cognitive functioning was not associated with employment status or perception of work ability.

Conclusions/Implications:

Although preliminary, results suggest that perception of work ability is an important factor to consider when evaluating employment-related outcomes in Veterans. Moreover, results indicate that while there is some overlap among the variables associated with employment status and perception of work ability, additional variables are linked with perception of work ability. Taken together, these findings suggest that perception of one’s ability to work and factors that influence it may be particularly important treatment targets in the Veteran population.

Combat-exposed Iraq and Afghanistan Veterans not only experience high rates of traumatic brain injury (TBI) and posttraumatic stress disorder (PTSD) (Fulton et al., 2015; O’Neil et al., 2013; Tanielian & Jaycox, 2008; Terrio et al., 2009), but they are at increased risk for developing other mental health disorders (Hoge, Auchterlonie, & Milliken, 2006; Milliken, Auchterlonie, & Hoge, 2007; Ramchand, Rudavsky, Grant, Tanielian, & Jaycox, 2015; Thomas et al., 2010), sustaining a wide range of physical health comorbidities (Eisen et al., 2012; Higgins et al., 2014; Pugh et al., 2014), and experiencing difficulties with psychosocial functioning and community reintegration (Beder, Coe, & Sommer, 2011; McGarity et al., 2017; Pietrzak, Goldstein, Malley, Johnson, & Southwick, 2009; Sayer, Carlson, & Frazier, 2014). Though not all military personnel experience chronic difficulties with combat-related injuries and adjustment after deployment, it is certainly not uncommon for returning service members to encounter functional impairments and reduced quality of life, especially following military discharge (Larson & Norman, 2014; Lippa et al., 2015; Pittman, Goldsmith, Lemmer, Kilmer, & Baker, 2012; Pugh et al., 2018; Sherman, Larsen, & Borden, 2015).

Perhaps one of the most difficult challenges Veterans face after separating from the military is finding and maintaining employment (Stern, 2017). Notably, rates of unemployment in the Veteran population are remarkably high (17%-80%) (Cohen, Suri, Amick, & Yan, 2013; Dillahunt-Aspillaga et al., 2017; Horton et al., 2013), and although the Veterans Health Administration has long provided vocational rehabilitation services to Veterans, including successful implementation of evidence-based supported employment programs (Carlson et al., 2018; Pogoda, Carlson, Gormley, & Resnick, 2018; Twamley et al., 2013), issues of unemployment remain a primary outcome of interest within the Department of Veterans Affairs (Dillahunt-Aspillaga & Powell-Cope, 2018). There are many psychological and physical health benefits associated with employment, as well as greater financial/economic stability (Benach et al., 2014; McKee-Ryan, Song, Wanberg, & Kinicki, 2005). Moreover, employment has been associated with both quality of life and satisfaction with life (Hawley, Armstrong, Czarnota, & Fields, 2016; Nakase-Richardson et al., 2017; O’Neill et al., 1998), and employment status is often used as a metric of successful transition from military to civilian life (Pogoda et al., 2016). Given the significance of employment, it is vital to increase understanding of the barriers associated with return to work in the Veteran population.

Importantly, a number of studies have previously evaluated factors associated with employment and employment-related outcomes (e.g., job status/position, work-role functioning, work-related problems) in Iraq and Afghanistan Veterans (Amara, Stolzmann, Iverson, & Pogoda, 2019; Cohen et al., 2013; Erbes, Kaler, Schult, Polusny, & Arbisi, 2011; Larson & Norman, 2014). Demographic characteristics such as age, gender, race/ethnicity, and years of education have been associated with work/occupational outcomes (Amara et al., 2019; Dillahunt-Aspillaga et al., 2017; Dillahunt-Aspillaga et al., 2018; Horton et al., 2013; Mortera, Kinirons, Simantov, & Klingbeil, 2018; Olsen, Hays, Orff, Jak, & Twamley, 2018; Pogoda et al., 2016), as have combat-related experiences such as combat exposure (Hamilton, Williams, & Washington, 2015; Larson & Norman, 2014) and TBI severity (Amara et al., 2019; Dillahunt-Aspillaga et al., 2018). Cognitive performance has also been explored as a factor associated with employment status and occupational functioning, but findings have been equivocal (Geuze, Vermetten, de Kloet, Hijman, & Westenberg, 2009; McGarity et al., 2017; Wrocklage et al., 2016). Even more common has been the examination of associations between psychiatric factors and other health-related symptoms (i.e., neurobehavioral symptoms, chronic pain, etc.) and employment, with these variables yielding the most consistent relationship with work-related outcomes (Amara et al., 2019; Amick et al., 2018; Cohen et al., 2013; Dillahunt-Aspillaga et al., 2018; Erbes et al., 2011; Hamilton et al., 2015; Horton et al., 2013; Larson & Norman, 2014; Mortera et al., 2018; Pogoda et al., 2016). Clearly, many factors have been associated with employment and occupational functioning in combat-exposed Veterans; however, these known variables do not fully account for the high rates of unemployment or difficulties with work-role functioning often observed in the Veteran population.

Beyond examining the abovementioned factors, another potentially relevant construct to consider in the context of employment and employment-related outcomes is self-efficacy. Broadly, self-efficacy refers to the belief or perception of one’s ability to accomplish tasks and/or achieve goals (Bandura, 1977, 1993). It has been proposed that individuals with low self-efficacy may shy away from tasks that are perceived to be difficult or challenging, whereas individuals with high self-efficacy eagerly approach difficult tasks and view them as “challenges to be mastered” (Bandura, 1993). While prior research has extensively examined self-efficacy as it relates to civilian employment (Carter, Nesbit, Badham, Parker, & Sung, 2018; Judge, Jackson, Shaw, Scott, & Rich, 2007; Stajkovic & Luthans, 1998), to date, this construct has not been readily explored in the Veteran population. In particular, exploring perceptions regarding work and/or one’s ability to work may provide additional insight into the employment challenges often faced by this population.

Although perceptions of work status or job functioning have not been specifically examined in Veterans, researchers have evaluated perceptions as they pertain to injury or illness within the context of TBI. Bahraini and colleagues (2018), for example, examined injury perceptions in post-9/11 Veterans and found that Veterans with TBI, relative to those with orthopedic injuries, had a more difficult time making sense of their injury or understanding their injury in the presence of greater symptoms of PTSD. In other words, more severe PTSD symptoms were associated with more negative injury perceptions in Veterans with TBI compared to those with orthopedic injuries. Another study evaluated associations between illness perceptions and post-concussive symptoms in a sample of civilian patients with mild TBI. Several dimensions of illness perception were evaluated, including beliefs about diagnosis and symptoms, illness duration, and illness consequences (i.e., what effects an illness has on physical, social, and psychological well-being). The authors found that patients’ perceptions about the consequences of their mild TBI significantly contributed to the maintenance of post-concussive symptoms following injury (Whittaker, Kemp, & House, 2007).

Finally, researchers have examined perceptions of cognitive problems and functional outcomes in the presence of PTSD. In a sample of Iraq and Afghanistan Veterans diagnosed with PTSD, Samuelson, Abadjian, et al. (2017) found an association between perception of cognitive problems and self-report of functional difficulties, including physical, emotional, and social functioning, as well as reintegration difficulties. Importantly, greater perception of cognitive difficulties was associated with poorer functional outcomes even though no objective cognitive deficits were identified on neuropsychological testing. Furthermore, objective cognitive functioning was unrelated to functional outcomes. The authors concluded that even in the absence of cognitive impairment, perception of cognitive difficulties has a strong influence on functional outcomes. In another study by the same group (Samuelson, Bartel, Valadez, & Jordan, 2017), the authors found that perception of cognitive difficulties was associated with poorer perceived quality of life in a largely non-Veteran sample who endorsed significant PTSD symptoms, again suggesting that perception (in this case, perception of cognitive problems) plays an important role in psychosocial outcomes.

Consideration of the above studies suggests that individuals’ perceptions of functioning may have a powerful influence on objective measures of functioning, and that evaluating perceptions in other contexts such as employment may offer valuable insights that can translate to important treatment targets or interventions for Veterans struggling to find and/or maintain employment following military separation. With this in mind, the purpose of this study was twofold: (1) to evaluate whether perception of a decline in one’s ability to work is associated with employment status in a sample of combat-exposed Veterans, and (2) to examine factors (including demographic characteristics, combat-associated experiences, neurocognitive functioning, psychiatric symptoms, and neurobehavioral/health-associated symptoms) associated with both employment status and perception of work ability. We hypothesized that perception of work ability, as well as psychiatric symptoms and neurobehavioral/health-associated symptoms, would be associated with employment status. As for factors related to perception of work ability, these analyses were exploratory and no a priori hypotheses were generated.

Method

Participants & Procedures

Participants included combat-exposed Iraq/Afghanistan Veterans primarily recruited from within a Veterans Affairs Medical Center as part of a larger Chronic Effects of Neurotrauma Consortium (CENC) project (Jurick et al., 2018; Merritt et al., 2019). Study procedures included structured interviews assessing combat-related experiences including TBI history and blast exposures as well as current psychiatric diagnoses, administration of self-report questionnaires, and a comprehensive neuropsychological assessment. All Veterans provided informed consent prior to research participation, and the study was approved by the local Veterans Affairs Institutional Review Board.

Inclusion criteria for the present study included being an Operation Enduring Freedom/Operation Iraqi Freedom/Operation New Dawn (OEF/OIF/OND) Veteran, having a history of combat exposure (defined by a total score of >17 on the Deployment Risk and Resiliency Inventory, Version 2, Section D [DRRI-2 (Vogt et al., 2013)]), and completing all study procedures (structured interviews, self-report questionnaires, and neuropsychological testing). Exclusion criteria were active psychosis or current mania, active substance dependence, diagnosis of dementia, and history of severe TBI as per VA/DoD guidelines (The Management of Concussion-mild Traumatic Brain Injury Working Group, 2016). Finally, in order to be included in the analyses pertaining to cognitive functioning, it was required that participants demonstrate acceptable scores on performance validity tests (PVTs; described below). Of the Veterans who initially consented to participate in the study (N=104), 21 were excluded due to not meeting full inclusion/exclusion criteria as outlined above, for a final sample of 83 Veterans. As for the analyses pertaining to the cognitive variables, 10 additional Veterans were excluded due to inadequate performance on PVTs and 2 Veterans were excluded due to missing cognitive test data, resulting in a sample of 71 participants for the cognitive analyses.

Measures

Demographic Characteristics

Participants completed a demographics form that gathered information pertaining to age, sex, race/ethnicity, education, employment, and service-connected disability rating. Regarding employment, Veterans were specifically asked the following questions: (1) “Are you currently employed?” and (2) “Has your ability to work declined due to an ongoing condition?”. Response options were “yes” or “no” for both questions. If participants selected “yes” to being currently employed, they were then asked to provide “hours per week” of employment. Veterans holding full or part-time positions were considered “employed” and those who indicated that they are not currently employed (“no” responses) were considered “unemployed.” Primary outcomes of interest were employment status (unemployed vs. employed) and Veterans’ perception of whether their ability to work has declined due to an ongoing condition (yes vs. no).

Combat-Related Experiences

As mentioned previously, the Deployment Risk and Resiliency Inventory, Version 2, Section D (DRRI-2) was administered to assess level of combat exposure (Vogt et al., 2013). The internal consistency of the DRRI-2 for the present sample was 0.93. To assess TBI history and blast exposure, the Virginia Commonwealth University (VCU) retrospective Concussion Diagnosis Interview-blast version (rCDI-B) and Concussion Diagnosis Interview-general version (rCDI-G) (Walker et al., 2015) were utilized. The VCU rCDI-B and VCU rCDI-G are structured interviews that gather information pertaining to potential concussive events, including the setting in which the injury occurred, date(s) of injury, and details on injury mechanism, including the nature and extent of any blast exposures. Additionally, the interview gathers details about event recollection (i.e., post-traumatic amnesia [PTA]), loss of consciousness (LOC), and associated symptoms—specifically symptoms that may indicate an alteration of consciousness (AOC). To determine whether a reported event qualified as a mild or moderate TBI, the VA/DoD Clinical Practice Guideline for the Management of Concussion-Mild Traumatic Brain Injury (The Management of Concussion-mild Traumatic Brain Injury Working Group, 2016) definition was applied. Based on responses to this diagnostic interview, the total number of lifetime TBIs and blast exposures were calculated for each participant.

Cognitive Measures

A comprehensive neuropsychological assessment was administered to all participants, assessing the broad domains of learning and memory, attention and processing speed, and executive functioning. Specific tests included the following: Brief Visuospatial Memory Test-Revised (BVMT-R) (Benedict, 1997); California Verbal Learning Test – Second Edition (CVLT-II) (Delis, Kramer, Kaplan, & Ober, 2000); selected subtests from the Delis-Kaplan Executive Function System (D-KEFS) (Delis, Kaplan, & Kramer, 2001), including Color Word Interference Test (CWIT) and Trail Making Test (TMT); selected subtests from the Wechsler Adult Intelligence Scale – Fourth Edition (WAIS-IV) (Wechsler, 2008), including Digit Span, Symbol Search, and Coding; and Wisconsin Card Sorting Test-64 Card Version (WCST-64) (Kongs, Thompson, Iverson, & Heaton, 2000). These particular measures were selected for the purpose of assessing a wide range of cognitive abilities and for their sound psychometric properties, including high test-retest reliability (e.g., 0.80-0.89) and good convergent validity (e.g., 0.65-0.80) (Benedict, 1997; Benedict, Schretlen, Groninger, Dobraski, & Shpritz, 1996; Delis et al., 2001; Delis et al., 2000; Kongs et al., 2000; Strauss, Sherman, & Spreen, 2006; Wechsler, 2008).

Raw scores from each test were transformed to standardized scores (e.g., z scores, T scores, scaled scores) using published normative data, adjusting for age, sex, and education as specified by each test manual. This transformation ensured that for all variables, higher scores reflect better performance. All standardized scores were then converted to the same metric (standard scores; M=100, SD=15) to facilitate the calculation of three domain-specific composite scores: (1) learning and memory (comprised of 5 variables including BVMT-R Total Recall, BVMT-R Delayed Recall, CVLT-II List A Total Recall, CVLT-II Short Delay Free Recall, and CVLT-II Long Delay Free Recall), (2) attention and processing speed (comprised of 11 variables including D-KEFS CWIT Color Naming, D-KEFS CWIT Word Reading, D-KEFS TMT Visual Scanning, D-KEFS TMT Number Sequencing, D-KEFS TMT Letter Sequencing, D-KEFS TMT Motor Speed, WAIS-IV Digit Span Forward, WAIS-IV Digit Span Backward, WAIS-IV Digit Span Sequencing, WAIS-IV Symbol Search, and WAIS-IV Coding) and (3) executive functioning (comprised of 5 variables including D-KEFS CWIT Inhibition, D-KEFS CWIT Inhibition Switching, D-KEFS TMT Number-Letter Switching, WCST-64 Perseverative Errors, and WCST-64 Total Errors).

In addition to the core neuropsychological battery listed above, the Wide Range Achievement Test 4 (WRAT4) Reading subtest (Wilkinson & Robertson, 2006) was used to assess pre-morbid intellectual functioning. Finally, performance validity was assessed using the following PVTs: Test of Memory Malingering (TOMM) Trial 2 and Retention Trial (Tombaugh, 1996), CVLT-II Forced Choice Recognition (Delis et al., 2000), and Green’s Medical Symptom Validity Test (MSVT) (Green, 2004). These are commonly-used measures of performance validity with well-established psychometric properties, including excellent internal consistency (e.g., 0.94-0.95) (Strauss et al., 2006). Standard cutoffs (according to manual-specific guidelines) were used to define suboptimal performance. Veterans failing one or more PVTs were excluded from the cognitive analyses.

Psychiatric Measures

The following self-report measures were used to assess symptoms of PTSD and depression, respectively: PTSD Checklist for DSM-5 (PCL-5) (Weathers et al., 2013) and Patient Health Questionniare-9 (PHQ-9) (Kroenke, Spitzer, & Williams, 2001). The internal consistency of the PCL-5 for the present sample was 0.94 and the PHQ-9 was 0.90, and these measures have been well validated in prior research (Blevins, Weathers, Davis, Witte, & Domino, 2015; Manea, Gilbody, & McMillan, 2012). A total score was calculated for each measure wherein higher scores indicate more severe psychiatric distress. Additionally, given the high prevalence of PTSD and depression in the Veteran population (Morissette et al., 2011; Vasterling et al., 2012), the Mini-International Neuropsychiatric Interview (MINI) 7.0 (Sheehan, 2014) was administered to assess current psychiatric disorders including PTSD and depression.

Neurobehavioral & Health-Related Measures

The Neurobehavioral Symptom Inventory (NSI) (Cicerone & Kalmar, 1995) was administered to assess vestibular, somatic, cognitive, and affective symptoms (Vanderploeg et al., 2015). The internal consistency of the NSI for the present sample was 0.95. Sleep quality and sleep disturbances were measured using the Pittsburgh Sleep Quality Index (PSQI) (Buysse, Reynolds III, Monk, Berman, & Kupfer, 1989). The internal consistency of the PSQI for the present sample was 0.82. Pain experiences were evaluated using the Short-Form McGill Pain Questionnaire (SF-MPQ) (Melzack, 1987) and pain catastrophizing was evaluated using the Pain Catastrophizing Scale (Sullivan, Bishop, & Pivik, 1995). The internal consistency of the SF-MPQ for the present sample was 0.92 and the Pain Catastrophizing Scale was 0.96. A total score was calculated for each measure; again, higher scores reflect more severe symptom reporting. Prior research has also shown that these neurobehavioral and health-related measures all have well established psychometric properties (Carpenter & Andrykowski, 1998; Hawker, Mian, Kendzerska, & French, 2011; King et al., 2012; Osman et al., 2000).

Approach to Data Analysis

To examine the association between employment status and perception of work ability, a chi-square analysis was used. Logistic regression was then used to determine how pre-specified sets of predictors, including demographic characteristics, combat-related experiences, cognitive measures, psychiatric symptoms, and neurobehavioral/health-related symptoms, were independently associated with the dichotomous outcomes of employment (all analyses were arranged to predict “no,” meaning unemployed) and perceived decline in work ability (all analyses were arranged to predict “yes,” meaning endorsement of a decline in work ability). Although there were some significant associations within sets of predictor variables for the logistic regression analyses, tolerance statistics indicated that multicollinearity was not an issue (tolerance = 0.31-0.87) (Midi, Sarkar, & Rana, 2010). SPSS (Version 25) was used for all analyses.

Results

Sample Characteristics

All participants (N=83) in the present study were OEF/OIF/OND Veterans with a history of combat exposure. Over one-third of the sample (38.6%) did not have a history of TBI (32 of 83 Veterans), 30.1% had a history of 1-2 TBIs (25 of 83), and 31.3% had a history of 3 or more TBIs (26 of 83). Of those with a positive TBI history, the majority (94.1%) had sustained mild TBIs (48 of 51), with only 5.9% of Veterans with TBI having experienced a moderate TBI (3 of 51). Furthermore, about half of the participants with a positive TBI history (51.0%) reported having experienced a blast-related TBI (26 of 51), 70.6% reported a TBI associated with LOC (36 of 51), and 92.2% reported a TBI associated with PTA (47 of 51). Veterans with a history of TBI were assessed, on average, 7.5 years following their most recent TBI (M=7.61, SD=5.73). With regard to psychiatric diagnoses, 49.4% of the sample met criteria for current PTSD (41 of 83) and 28.9% met criteria for current depression (24 of 83). Table 1 displays additional participant characteristics, including demographics, combat-related experiences, cognitive measures, psychiatric symptoms, and neurobehavioral/health-related symptoms.

Table 1.

Overall sample characteristics (N=83).

Variable	Mean (SD) / N (%)	Median	Range
Demographic Characteristics
Age	34.59 (6.41)	33.00	25 – 48
Sex (Male)	76 (91.6%)	--	--
Years of Education	15.12 (1.67)	15.00	12 – 18
WRAT-4 Reading (SS)	103.40 (10.39)	103.00	80 – 126
White	46 (55.4%)	--	--
Service-Connected Disability Rating	62.23 (35.99)	70.00	0 – 100
Combat-Related Experiences
DRRI-D-2 Total Score	38.67 (17.31)	32.00	18 – 92
# Blast Exposures	2.72 (3.47)	2.00	0 – 16
# TBIs	2.12 (2.95)	1.00	0 – 17
Cognitive Measures a
Learning/Memory Composite (SS)	99.70 (10.91)	99.94	66 – 124
Attention/PS Composite (SS)	99.76 (9.03)	101.24	77 – 119
EF Composite (SS)	99.80 (8.53)	100.16	82 – 120
Psychiatric Symptoms
PCL-5	27.35 (17.68)	27.00	0 – 65
PHQ-9	9.18 (6.54)	8.00	0 – 27
Neurobehavioral/Health-Related Symptoms
NSI	24.61 (15.83)	24.00	0 – 65
PSQI	10.63 (4.28)	11.00	2 – 20
SF-MPQ	12.75 (9.65)	11.00	0 – 45
Pain Catastrophizing	13.65 (12.70)	10.00	0 – 46

Abbreviations: SD = standard deviation; WRAT-4 = Wide Range Achievement Test 4; SS = standard score; DRRI-D-2 = Deployment Risk and Resiliency Inventory, Version 2, Section D; TBIs = traumatic brain injuries; PS = processing speed; EF = executive functioning; PCL-5 = PTSD Checklist for DSM-5; PHQ-9 = Patient Health Questionnaire-9; NSI = Neurobehavioral Symptom Inventory; PSQI = Pittsburgh Sleep Quality Index; SF-MPQ = Short-Form McGill Pain Questionnaire.

^aTen participants were excluded from the cognitive analyses due to performance validity test failure, and an additional two participants were not included in these analyses due to missing cognitive data, resulting in a sample of N=71.

Relationship between Perception of Work Ability & Employment Status

With regard to employment status, 54.2% of the sample were employed (45 of 83); of those employed, 73.3% reported full-time employment (33 of 45). As for perception of work ability, 65.1% of the sample reported a decline in their ability to work (54 of 83). A chi-square analysis revealed that a significantly greater proportion of unemployed relative to employed Veterans reported a decline in work ability (χ² (1, n = 83) = 3.91, p = .048, φ = 0.22). Specifically, among unemployed Veterans, 76.3% reported a decline in their ability to work (29 of 38); among employed Veterans, 55.6% reported a decline in their ability to work (25 of 45).

Predictors of Unemployment

Results showed that the set of demographic (χ² (6, n = 83) = 5.84, p = .442), combat (χ² (4, n = 83) = 6.99, p = .137), and cognitive (χ² (3, n = 71) = 2.06, p = .561) variables (when examined as independent models) were not significantly associated with unemployment. However, the psychiatric variables were significantly related to unemployment (χ² (2, n = 83) = 8.17, p = .017) and accounted for 12.5% of the variance (overall classification = 60.2%, unemployed = 42.1% and employed = 75.6%). Additionally, the neurobehavi oral/health-related variables also were significantly associated with unemployment (χ² (4, n = 83) = 17.41, p = .002) and accounted for 25.3% of the variance (overall classification = 74.7%, unemployed = 65.8% and employed = 82.2%). See Table 2.

Table 2.

Results of logistic regression predicting unemployment.

Variables	R2	B	SE	Wald’s χ2	OR [95% CI]	p
Demographic Characteristics	0.09
Age		0.04	0.05	0.01	1.00 [0.91, 1.10]	.934
Sex (Male vs. Female)		−1.59	1.15	1.92	0.20 [0.02, 1.93]	.166
Years of Education		−0.16	0.18	0.80	0.85 [0.60, 1.21]	.371
WRAT-4 Reading		0.01	0.03	0.06	1.01 [0.96, 1.06]	.802
White (White vs. Non-White)		0.03	0.49	0.00	1.03 [0.39, 2.71]	.949
Service-Connected Disability Rating		0.01	0.01	1.50	1.01 [1.00, 1.02]	.221
Combat-Related Experiences	0.11
DRRI-D-2 Total Score		−0.02	0.02	1.22	0.98 [0.95, 1.01]	.270
# Blast Exposures		0.12	0.08	2.42	1.13 [0.97, 1.31]	.120
# TBIs (0 vs. 1-2)		−1.09	0.61	3.15	0.34 [0.10, 1.12]	.076
# TBIs (0 vs. 3+)		−1.04	0.63	2.70	0.35 [0.10, 1.22]	.100
Cognitive Measures	0.04
Learning/Memory Composite		−0.03	0.02	1.29	0.97 [0.93, 1.02]	.255
Attention/PS Composite		0.00	0.04	0.00	1.00 [0.93, 1.08]	1.000
EF Composite		−0.01	0.03	0.17	0.99 [0.92, 1.05]	.677
Psychiatric Symptoms	0.13
PCL-5		−0.01	0.02	0.37	0.99 [0.95, 1.03]	.544
PHQ-9		0.12	0.06	4.90	1.13 [1.01, 1.26]	.027
Neurobehavioral/Health-Related Symptoms	0.25
NSI		0.01	0.03	0.06	1.01 [0.95, 1.06]	.801
PSQI		−0.02	0.09	0.03	0.98 [0.82, 1.18]	.862
SF-MPQ		0.06	0.05	1.32	1.06 [0.96, 1.17]	.250
Pain Catastrophizing Scale		0.04	0.03	2.14	1.05 [0.99, 1.11]	.143

Abbreviations: WRAT-4 = Wide Range Achievement Test 4; DRRI-D-2 = Deployment Risk and Resiliency Inventory, Version 2, Section D; TBIs = traumatic brain injuries; PS = processing speed; EF = executive functioning; PCL-5 = PTSD Checklist for DSM-5; PHQ-9 = Patient Health Questionnaire-9; NSI = Neurobehavioral Symptom Inventory; PSQI = Pittsburgh Sleep Quality Index; SF-MPQ = Short-Form McGill Pain Questionnaire.

Notes: R² = Nagelkerke R Square.

With regard to the model examining psychiatric variables and unemployment, greater depressive symptoms as measured by the PHQ-9 were significantly related to unemployment (p = .027) but PTSD symptoms as measured by the PCL-5 were not (p = .544; see table 2). As for the model examining whether neurobehavioral/health-related symptoms were associated with unemployment, none of the variables (NSI, PSQI, SF-MPQ, or Pain Catastrophizing) were independently related to unemployment (all p’s > .05; see Table 2).

Predictors of Perceived Decline in Work Ability

Results showed that the model examining the set of demographic characteristics was significantly associated with perceived decline in work ability (χ² (6, n = 83) = 18.31, p = .006) and accounted for 27.3% of the variance (overall classification = 69.9%, decline = 81.5% and no decline = 48.3%). Combat-related experiences also were significantly associated with perceived decline in work ability (χ² (4, n = 83) =10.76, p = .029) and accounted for 16.8% of the variance (overall classification = 63.9%, decline = 88.9% and no decline = 17.2%). Additionally, the models including the psychiatric variables (χ² (2, n = 83) = 21.33, p < .001) and the neurobehavioral/health-related variables (χ² (4, n = 83) = 39.02, p < .001) were significantly related to perceived decline in work ability, accounting for 31.2% (overall classification = 74.7%, decline = 83.3% and no decline = 58.6%) and 51.7% (overall classification = 80.7%, decline = 85.2.% and no decline = 72.4%) of the variance, respectively. In contrast, the model including the cognitive measures was not associated with perceived decline in work ability (χ² (3, n = 71) = 4.16, p = .245). See Table 3.

Table 3.

Results of logistic regression predicting perceived decline in work ability.

Variables	R2	B	SE	Wald’s χ2	OR [95% CI]	p
Demographic Characteristics	0.27
Age		0.01	0.06	0.02	1.01 [0.90, 1.13]	.901
Sex (Male vs. Female)		−0.18	0.93	0.04	0.84 [0.13, 5.20]	.847
Years of Education		−0.09	0.22	0.15	0.92 [0.60, 1.40]	.696
WRAT-4 Reading		0.02	0.03	0.35	1.02 [0.96, 1.08]	.555
White (White vs. Non-White)		−0.68	0.57	1.40	0.51 [0.17, 1.56]	.237
Service-Connected Disability Rating		0.03	0.01	11.73	1.03 [1.01, 1.05]	.001
Combat-Related Experiences	0.17
DRRI-D-2 Total Score		0.01	0.02	0.29	1.01 [0.98, 1.05]	.591
# Blast Exposures		−0.04	0.08	0.26	0.96 [0.82, 1.13]	.612
# TBIs (0 vs. 1-2)		−1.67	0.75	4.91	0.19 [0.04, 0.83]	.027
# TBIs (0 vs. 3+)		−1.94	0.78	6.25	0.14 [0.03, 0.66]	.012
Cognitive Measures	0.08
Learning/Memory Composite		−0.04	0.03	2.52	0.96 [0.91, 1.01]	.113
Attention/PS Composite		0.06	0.04	1.93	1.06 [0.98, 1.14]	.165
EF Composite		−0.04	0.04	1.47	0.96 [0.89, 1.03]	.226
Psychiatric Symptoms	0.31
PCL-5		0.05	0.02	5.22	1.06 [1.01, 1.10]	.022
PHQ-9		0.07	0.07	1.02	1.07 [0.94, 1.21]	.312
Neurobehavioral/Health-Related Symptoms	0.52
NSI		0.12	0.04	8.24	1.12 [1.04, 1.22]	.004
PSQI		−0.05	0.11	0.22	0.95 [0.77, 1.17]	.638
SF-MPQ		−0.08	0.07	1.34	0.93 [0.82, 1.05]	.247
Pain Catastrophizing Scale		0.11	0.04	6.71	1.12 [1.03, 1.22]	.010

Abbreviations: WRAT-4 = Wide Range Achievement Test 4; DRRI-D-2 = Deployment Risk and Resiliency Inventory, Version 2, Section D; TBIs = traumatic brain injuries; PS = processing speed; EF = executive functioning; PCL-5 = PTSD Checklist for DSM-5; PHQ-9 = Patient Health Questionnaire-9; NSI = Neurobehavioral Symptom Inventory; PSQI = Pittsburgh Sleep Quality Index; SF-MPQ = Short-Form McGill Pain Questionnaire.

Notes: R² = Nagelkerke R Square.

With regard to the model examining whether demographic characteristics were associated with perceived decline in work ability, service-connected disability rating was the only variable related to perception of work ability (p = .001; see table 3). Specifically, results showed that higher service-connected disability ratings were associated with increased odds of perceiving a decline in work ability. As for the model assessing variables associated with combat-related experiences, TBI history (0 vs. 1-2 TBIs and 0 vs. 3+ TBIs) was significantly related to perceived decline in work ability (p = .027 and p = .012, respectively; see Table 3). That is, Veterans with a TBI history were more likely to perceive a decline in work ability relative to Veterans without a history of TBI. Finally, within the psychiatric and neurobehavioral/health-related symptoms models, the PCL-5 (p = .022), NSI (p = .004), and pain catastrophizing (p = .010) were all significantly associated with perception of work ability (see Table 3), such that greater symptoms were associated with increased odds of perceiving a decline in work ability.

Discussion

Relationship between Perception of Work Ability & Employment Status

A key finding that emerged from this study was that perception of work ability was indeed associated with unemployment in our sample of combat-exposed Veterans. Although this is the first study, to our knowledge, to empirically examine the relationship between perception of one’s ability to work (a subjective rating) and employment status (an objective marker of functioning) in military Veterans, there is evidence within the broader psychology literature to suggest that perceptions of functioning, or self-efficacy, may have a powerful influence on a wide range of psychosocial and functional outcomes (Bahraini et al., 2018; Samuelson, Abadjian, et al., 2017; Samuelson, Bartel, et al., 2017; Whittaker et al., 2007). Moreover, perceptions and beliefs about illness and recovery have been associated with expectations regarding return to work and/or work participation in the context of somatic and psychiatric disorders (Hoving, Van der Meer, Volkova, & Frings-Dresen, 2010; Løvvik, ∅verland, Hysing, Broadbent, & Reme, 2014). Thus, our finding that perception of work ability is related to employment status is consistent with this broader body of literature and suggests that future studies within the Veteran population that target perceptions of employment and ability to work are warranted. Moving forward, it will be necessary for studies to utilize a more robust measure of perception of occupational functioning, and to further explore the directional nature of this relationship, but these preliminary findings indicate that perception of work ability plays an important role in actual employment status. Additionally, specifically examining self-efficacy and how that influences employment-related outcomes may be another important avenue of future research.

Predictors of Unemployment & Perceived Decline in Work Ability

Our results showed that while psychiatric and neurobehavioral/health-related symptoms were associated with both employment status and perception of work ability, additional variables were linked with perception of work ability. Specifically, the models assessing demographic characteristics and combat-related experiences were significantly related to perception of work ability but not employment status. Cognition (as measured by objective tests of learning and memory, attention and processing speed, and executive functions) was not associated with either employment status or perception of work ability.

As indicated above, a number of prior studies have examined variables associated with employment, with the most consistent finding being the link between psychiatric and neurobehavioral/health-related symptoms and employment status (Amara et al., 2019; Amick et al., 2018; Cohen et al., 2013; Dillahunt-Aspillaga et al., 2018; Erbes et al., 2011; Hamilton et al., 2015; Horton et al., 2013; Larson & Norman, 2014; Mortera et al., 2018; Pogoda et al., 2016), such that more severe symptomatology is associated with unemployment. Our results thus add to the growing body of literature supporting these associations—in particular, the relationship between elevated symptoms of depression and unemployment. However, the more novel findings to emerge from the present study were the variables associated with perception of work ability. Perhaps not surprisingly, both psychiatric and neurobehavioral/health-related symptoms were significantly associated with perceived decline in work ability. Specifically, PTSD symptoms, neurobehavioral symptoms, and pain catastrophizing were individually associated with perception, such that Veterans with greater symptoms and more catastrophizing reactions to pain were more likely to perceive that their work abilities had declined. These results provide empirical support for the notion that perception of functioning is influenced by mental health and other neurobehavioral/health-related symptoms in combat-exposed Veterans, and these findings suggest that perhaps symptom reduction may lead to improved perception of work ability and, in turn, actual employment. Although these results will need to be replicated using larger samples, this information has the potential to be used clinically when working with Veterans in psychotherapy to address both objective barriers to working as well as to challenge negative beliefs about ability to work.

As for demographic characteristics, none of the variables we examined were related to actual employment, but service-connected disability rating was significantly associated with perception of work ability. Notably, higher service-connected disability ratings were associated with a greater likelihood of perceiving a decline in work ability. While prior research has not evaluated service-connected disability rating as it pertains to perception of ability to work, previous studies have evaluated this variable as a predictor of employment status, with mixed results (Drew et al., 2001; Greenberg & Rosenheck, 2007; Olsen et al., 2018; Resnick & Rosenheck, 2008; Tsai & Rosenheck, 2013). For example, a recent study by Olsen et al. (2018) found that Veterans with higher service-connected disability ratings were less likely to be employed, suggesting that this particular group of Veterans may especially benefit from vocational rehabilitation programs that provide psychoeducation and offer skills-based training for returning to work in a civilian context. Notably, these results are consistent with general (i.e., non-military) disability statistics, as civilians with disabilities are similarly less likely to be employed relative to those without disabilities (U.S. Bureau of Labor Statistics, 2020). Taken together, there is evidence in both the military and civilian literature to suggest that an important relationship exists between disability and employment-related outcomes, including both actual employment and perception of functioning.

Although we did not find that service-connected disability rating was associated with employment status in our sample, results from our study suggest that it may be important to integrate information about perception of work ability into vocational rehabilitation programs and to specifically target Veterans’ negative beliefs regarding their ability to work. With that said, it is important for future research to evaluate the directionality of any observed associations between service-connected disability rating and perceived functioning. It may be the case that the relationship between these variables is bidirectional (i.e., receiving a higher disability rating may influence perceived functioning or reinforce having a poor perception of one’s functioning; alternatively, having a poor perception of functioning may influence one’s disability rating); thus, future research is needed to better understand the direction of this relationship, as well as evaluate additional variables that may moderate or mediate this relationship.

With regard to combat-related experiences, history of TBI was significantly associated with perception of work ability, whereas combat exposure and number of blast exposures were unrelated to perception, and no associations were found between combat-related experiences and employment status. These results are generally consistent with other studies that have examined the association between mTBI and unemployment, with no reliable relationship found between the two (Amara et al., 2019; Amick et al., 2018; Pogoda et al., 2016). Our results add to this literature by showing that history of mTBI also does not significantly contribute to actual employment status, but does influence perception of work ability. Although more research is needed to verify these results, this particular finding may also be useful clinically. Specifically, it may be beneficial for Veterans with a history of mTBI to learn that mTBI history/burden may impact perception more than objective abilities, and provides additional support for the importance of providing psychoeducation to Veterans with history of mTBI about the expected course of recovery following injury.

Finally, it is noteworthy that objective cognitive performance was unrelated to both employment status and perception of ability to work. Prior research examining the relationship between performance on neuropsychological tests and employment status has yielded inconsistent findings (Geuze et al., 2009; McGarity et al., 2017; Wrocklage et al., 2016), with results from our study supporting a lack of relationship between objective cognition and employment-related outcomes in combat-exposed Veterans. Clinically, these findings may be relevant for Veterans who believe that their cognitive functioning may be interfering with their ability to work.

Strengths, Limitations, and Future Directions

The present study had many strengths, including a well characterized sample of combat-exposed Veterans and a comprehensive assessment of several variables (i.e., demographic characteristics, combat-related experiences, cognition, and psychiatric and neurobehavioral/health-related symptoms) to determine which factors are most associated with employment status and perception of work ability. Furthermore, in addition to evaluating employment status, we examined perception of one’s ability to work—something that has not yet been readily examined in the Veteran population. This study, however, is not without its limitations. First, the data are cross-sectional, which limits our ability to make any directional or causal statements about the relationship between perception of work ability and employment status. The nature of the relationship could feasibly go both ways (i.e., Veterans developed a poor perception of their ability to work, which then led them to not seek employment, or the reverse could be true where Veterans found themselves unemployed after separating from the military, which then led to a negative perception of their ability to work), or the relationship could also be cyclical in nature, with both factors influencing one another simultaneously.

A second limitation of the study is that our primary variables of interest—employment status and perception of work ability—were measured using a single-item, dichotomous variable. Regarding the former (employment status), we assessed whether Veterans were currently employed or unemployed and did not evaluate reasons for unemployment or consider part-time versus full-time employment in our analyses. In order to more fully understand Veteran unemployment, future research could examine reasons for unemployment and consider part-time versus full-time employment in analyses. In the present study, approximately three-quarters (73.3%) of our employed Veteran sample reported full-time employment, with the remaining sample (26.6%) reporting part-time employment; thus, we were likely underpowered to be able to assess the significance of full-time versus part-time work in our sample. However, in follow-up analyses, we compared these groups on our variables of interest and found that Veterans who were employed full-time did not differ from Veterans who were employed part-time with regard to cognitive measures, psychiatric symptoms, and neurobehavioral/health-related symptoms. Full-time and part-time employees did differ, however, with respect to age, education, and number of blast exposures such that those who were employed full-time were older, had approximately one additional year of education, and experienced more blast exposures relative to Veterans who were employed part-time. Future research using larger samples would be better suited to examine relationships between predictor variables and those who are unemployed, employed part-time, and employed full-time, as these distinctions in employment status may reveal more nuanced findings, particularly in terms of perceived versus actual ability to maintain full-time employment. Finally, as for our perception variable, perceived decline in work ability was assessed with a “yes” or “no” response. Certainly, future research using a more robust and sensitive measure of perception of work functioning is needed. Nonetheless, the present study provides preliminary data to suggest that ongoing examination of perception of occupational functioning in this population is warranted.

As for the generalizability of our findings, we do not know whether our results would pertain to female Veterans, Veterans not exposed to combat, and those with more serious injuries (e.g., those with history of severe TBI) or other disabilities/polytrauma comorbidities. We also did not specifically explore any intersectional identities in this study. Additionally, caution should be taken when generalizing our results to civilians. Although research conducted on working adults in the civilian sector has found somewhat similar results regarding variables associated with perception of work ability (McGonagle, Fisher, Barnes-Farrell, & Grosch, 2015), more research is necessary to determine the extent to which results generalize to civilian samples.

Lastly, the number of Veterans meeting our inclusion/exclusion criteria did not allow for consideration of possible moderation or mediation effects of the predictor variables, perception of work ability, and employment status. Future research will need to examine these relationships, as it is conceivable that perception of work ability may mediate the relationship between psychiatric symptoms and employment status. It might also be interesting to evaluate perceptions more comprehensively both before and after deployment, as well as conduct longitudinal studies to better understand the nature of these associations.

Conclusions

Addressing Veteran unemployment has been a top priority within the Department of Veterans Affairs for many years (Dillahunt-Aspillaga & Powell-Cope, 2018). The present study adds to our understanding of the factors that are associated with employment-related outcomes in Veterans by showing that perception of work ability is related to employment status in this population. Moreover, our results showed that while psychiatric and neurobehavioral/health-related symptoms were associated with both employment status and perception of work ability, additional variables including service-connected disability rating and history of TBI were also significantly related to perception of work ability but not employment status. Objective cognitive functioning was unrelated to both employment status and perception of work ability. Altogether, these findings support the notion that perception of one’s ability to work and factors influencing it may be particularly important treatment targets in combat-exposed Veterans. Although these findings will need to be replicated using larger samples and more robust measures of employment and perception of work ability, perception appears to play an important role in employment-related outcomes.

Impact.

• This is one of the first studies to examine the relationship between perceptions of work ability and actual employment status in a sample of combat-exposed Veterans.

• Study findings indicate that work perception is related to actual employment status. Furthermore, psychiatric and neurobehavioral/health-related symptoms are associated with both employment status and perception of ability to work, but additional variables are related to perception, including demographic characteristics and combat-related experiences.

• Our findings suggest that symptom reduction and psychoeducation may lead to improved perception of work ability and, in turn, actual employment. The information from this study may ultimately be used clinically to address objective barriers to working, as well as negative beliefs about one’s ability to work.