Abstract

Personality Assessment Inventory (PAI) profiles were examined in 160 U.S. service members (SMs) following mild–severe traumatic brain injury (TBI). Participants who sustained a mild TBI had significantly higher PAI scores than those with moderate–severe TBI on eight of the nine clinical scales examined. A two-step cluster analysis identified four PAI profiles, heuristically labeled “High Distress”, “Moderate Distress”, “Somatic Distress,” and “No Distress”. Postconcussive and posttraumatic stress symptom severity was highest for the High Distress group, followed by the Somatic and Moderate Distress groups, and the No Distress group. Profile groups differed in age, ethnicity, rank, and TBI severity. Findings indicate that meaningful patterns of behavioral and personality characteristics can be detected in active duty military SMs following TBI, which may prove useful in selecting the most efficacious rehabilitation strategies.

Introduction

According to the Department of Defense (DoD), ∼300,000 military service members (SMs) have been diagnosed with traumatic brain injuries (TBIs) since the year 2000, with 26,561 new diagnoses of TBI in 2013 alone (Defense and Veterans Brain Injury Center, 2014). Statistics show that the median annual cost of VA care is nearly four times higher for TBI-diagnosed veterans compared with those without TBI (Taylor et al., 2012). In those with mild TBI (mTBI; 82.5% of the total TBI cases), a substantial minority report persistent symptoms and functional impairment during the year after injury (Hou et al., 2012; McMahon et al., 2014). In addition, the negative impact of moderate–severe TBI on quality of life through impaired physical and psychological functioning, as well as its potential connection with cognitive decline later in life, are significant areas of concern. Addressing symptoms and identifying effective assessment and treatment modalities among military personnel with TBI are top DoD and Veterans Health Affairs priorities (Otis, McGlinchey, Vasterling, & Kerns, 2011).

Posttraumatic Stress Disorder (PTSD) complicates TBI outcomes in post-deployment military populations (Hesdorffer, Rauch, & Tamminga, 2009; Pietrzak, Johnson, Goldstein, Malley, & Southwick, 2009). Both postconcussive and posttraumatic symptoms appear to contribute to level of overall global distress, consistent with the model of cumulative disadvantage or burden of adversity (Brenner, Vanderploeg, & Terrio, 2009). The cumulative disadvantage model (Merton, 1968), originally conceptualized to explain differences in career trajectories of those who did and did not receive early peer recognition, access to resources and opportunities for career advancement, has been applied to postconcussive symptoms (Ruff, Camenzuli, & Mueller, 1996). In this model, functional difficulties across physical, emotional, cognitive, psychosocial, vocational, financial, and recreational domains act as stressors that, in combination with premorbid factors, lead to increased disability and globally poor outcomes. As applied to a modern military sample, a poor outcome trajectory is associated with the concurrent presence of multiple disorders such as PTSD and mTBI (King, 2008). Long-term symptoms interact in an iatrogenic fashion and lead to maladaptive behaviors, additional stress-related conditions and poor outcome. Support for this model is provided by findings showing that recovery from PTSD over an extended period of 16 years is lowered from 69% to 48% among those who subsequently sustained an mTBI (Vanderploeg, Belanger, & Curtiss, 2009). Treatment implications for TBI include recommendations for early intervention, education, and expectations for recovery to improve the outcome trajectory. When comorbidities are identified, a stepwise treatment algorithm is proposed (Terrio et al., 2009) in which psychiatric issues are addressed first, followed by treatment of physical symptoms (headache, insomnia) and finally any remaining cognitive complaints. This algorithm is consistent with mTBI clinical practice guidelines (VA/DoD Clinical Practice Guideline: Management of Concussion/Mild Traumatic Brain Injury, 2009) and aimed at treatment of symptoms regardless of etiology and recovery of social roles to facilitate resilience (Seguin et al., 2007).

Efforts to use symptom ratings to discriminate persistent effects of TBI from effects of other ongoing conditions, including PTSD, depression, pain, and sleep disorders have been mostly unsuccessful, due in large part to the confounding of symptoms (Lew et al., 2008; Macera, Aralis, Macgregor, Rauh, & Galarneau, 2012; Morissette et al., 2011; Schneiderman, Braver, & Kang, 2008). As a consequence, many common self-report measures suffer from lack of specificity and weak empirical evidence of validity (Betthauser, Bahraini, Krengel, & Brenner, 2012). An alternative measure encompassing a variety of behaviors and symptoms is provided by the Personality Assessment Inventory (PAI; Morey, 1991), a well-validated and commonly used self-report rating covering a variety of clinical domains (Piotrowski, 2000). Including symptoms found in a variety of disorders in one instrument is a unique and beneficial characteristic offered by the PAI. It is particularly advantageous in the examination of outcome following TBI in military populations with high rates of comorbid conditions. Although the clinical scales of this instrument are based on symptoms commonly reported by individuals with Diagnostic and Statistical Manual diagnoses (i.e., somatic, anxiety, depression, and paranoia) the interpretation of scores is based on the profile shape and elevation of scales, rather than diagnostic specificity. Thus, it is suitable as an assessment tool that is consistent with the theory of cumulative disadvantage and the concept of treatment of symptoms, regardless of diagnostic etiology. The PAI is potentially useful in identifying distinct groups of patients on different recovery trajectories following TBI. The classification of groups has treatment implications based on the overall level and type of residual symptom patterns identified.

The PAI consists of three major factors, found in both the original mixed psychiatric standardization sample (Morey, 1991) and in a combined military and civilian TBI sample (Demakis et al., 2007). Factors represent general psychiatric distress (factor 1), antisocial features and substance use problems (factor 2), and mania (factor 3). Elevations of factor 1 scales occur more commonly among individuals with mTBI, whereas factor 2 scales tend to be found among those with more severe TBI (Kurtz, Shealy, & Putnam, 2007). Symptomatic distress among military personnel appears to occur in the post-deployment time frame or in the context of an injury, as suggested by a PAI study of deployed non-treatment seeking SMs who showed minimal subscale score differences compared with an age- and gender-matched sample from the original PAI standardization sample (Morey et al., 2011). Using a cluster analytic approach to identify groups or clusters of individuals based on PAI scale response profiles, Velikonja, Warriner, and Brum (2010) identified seven distinct groups of civilian brain injured patients (68% TBI) with consistent profile patterns, across split-sample subgroups, and various cluster methods.

Cluster analytical methodology was used in the present study to further examine common PAI personality profiles and extend findings to a sample of military SMs with TBI. This study aimed to advance knowledge about distinct types of emotional symptoms, psychological symptoms, personality characteristics, and maladaptive behaviors (such as alcohol and substance use) endorsed by military SMs with TBI. Demographic, injury related and symptom characteristics of the profile groups were compared. Results were related to previous studies of the PAI in military and civilian samples and treatment implications discussed.

Methods

Participants and Procedures

This study is based on a retrospective review of data from neuropsychological evaluations of SMs conducted at a large U.S. Military Treatment Facility and a major U.S. Military Medical Center. All procedures including recruitment, consent, and data acquisition were approved and monitored by the governing Institutional Review Board. Of 258 evaluations, two cases with unknown TBI severity and 56 cases missing key demographic variables (age, gender, education, and ethnicity) were excluded from analysis. In addition, cases with Trial 2 scores <45 on the test of memory malingering (n = 14) and those with invalid PAI validity scale scores (negative impression management T ≥ 93, positive impression management T ≥ 70, infrequency T ≥ 76, or inconsistency T ≥ 74) were excluded from this study (n = 26). These exclusions were intended to result in a sample with adequate evidence of both symptom and performance validity (Van Dyke, Millis, Axelrod, & Hanks, 2013). In total, 20% of the cases were excluded based on validity indicators, yielding a final study sample of 160 SMs with TBI. Based on research reporting that up to 40% of mTBI populations may show evidence of invalid test scores on neuropsychological evaluation (Lange, Iverson, Brooks, & Rennison, 2010), the proportion of excluded cases in our mixed severity TBI sample aligns with expectations.

All cases were referred for neuropsychological evaluation following TBI (n = 106 mild, n = 43 moderate, n = 11 severe). Moderate and severe TBI cases were combined into one group for analysis, to ensure adequate power of the findings. Evaluations were clinical in nature, but an unknown number may also have been used for determination of compensation and disability ratings. mTBI was characterized by a confused or disoriented state which lasted <24 h or loss of consciousness for up to 30 min and memory loss lasting <24 h. Moderate TBI was characterized by loss of consciousness for >30 min but <24 h and memory loss lasting >24 h but <7 days. Severe TBI was characterized by loss of consciousness for >24 h and memory loss for >7 days; and yielding normal or abnormal results. Presence and severity of TBI was confirmed by clinical interview with the evaluating neuropsychologist. Initial post-resuscitation GCS scores and structural brain imaging results were not used for severity classification, since these measures were unavailable. The most frequent cause of TBI was explosive blasts (53.8%), followed by motor vehicle accidents (18.1%), falls (12.5%), and “other” mechanisms (13.8%). Mechanism of injury was unknown for three individuals (1.8%). In addition to these injury characteristics, time since injury was calculated for each of the subjects in this study. Time since injury (mean = 30.4 weeks; SD = 37.7) was calculated as the difference between the most recent injury date and the date of administration of the PAI (see Table 1 for further details about the study sample).

Table 1.

Summary characteristics of the study sample (N = 160)

Demographic variable Mean (SDRange 
 Age 28.5 (6.8) 19–49 
 Education 12.7 (1.5) 8–18 
 N 
 Gender: Male/female 151 /9 94.4/5.6 
 Ethnicity: Caucasian/other 104 /56 65.0/35.0 
 Rank: E1–E4/E5+ 88 /72 55.0/45.0 
Injury variable Mean (SDRange 
 Time since injury (weeks) 30.4 (37.7) 1–303 
 Deployment number 1.6 (1.1) 0–9 
 N 
 Deployment location: OIF-OEF/other 126/34 78.8/21.2 
 LOC: ±/missing 84/64 /10 52.5/40.0/7.5 
 PTA: ±/missing 21/59 /61 15/42/43 
 TBI severity: mild/moderate–severe 106/54 66.3/33.7 
 Mechanism: blast/non-blast 54/74 33.7/46.2 
Symptom variable Mean (SDRange 
 PCL-M total score 43.2 (18.9) 17–85 
 NSI total score 34.0 (19.0) 0–76 
Demographic variable Mean (SDRange 
 Age 28.5 (6.8) 19–49 
 Education 12.7 (1.5) 8–18 
 N 
 Gender: Male/female 151 /9 94.4/5.6 
 Ethnicity: Caucasian/other 104 /56 65.0/35.0 
 Rank: E1–E4/E5+ 88 /72 55.0/45.0 
Injury variable Mean (SDRange 
 Time since injury (weeks) 30.4 (37.7) 1–303 
 Deployment number 1.6 (1.1) 0–9 
 N 
 Deployment location: OIF-OEF/other 126/34 78.8/21.2 
 LOC: ±/missing 84/64 /10 52.5/40.0/7.5 
 PTA: ±/missing 21/59 /61 15/42/43 
 TBI severity: mild/moderate–severe 106/54 66.3/33.7 
 Mechanism: blast/non-blast 54/74 33.7/46.2 
Symptom variable Mean (SDRange 
 PCL-M total score 43.2 (18.9) 17–85 
 NSI total score 34.0 (19.0) 0–76 

Notes: OIF = Operation Iraqi Freedom; OEF = Operation Enduring Freedom; TBI = traumatic brain injury; LOC = loss of consciousness; PTA = posttraumatic amnesia; PCL-M = Posttraumatic Stress Checklist Military version; NSI = Neurobehavioral Symptom Inventory.

Measures

The PAI is a well-validated 344-item self-report questionnaire that measures several facets of personality and psychopathology (Morey, 1991). Individual items are answered with one of four graded responses ranging from “false” to “very true” (with corresponding values of 0–3, respectively). The responses to items that comprise each scale and subscale are totaled to arrive at the raw score for that scale/subscale. These scores are then transformed to a standardized t score, with a mean of 50 and a SD of 10 based on Morey's original census sample of 1000 individuals. The PAI includes four validity scales designed to detect the endorsement of several unlikely/infrequently occurring items, inconsistency of item endorsement, and presentation of an overly positive or negative impression via item endorsements. In total, the PAI consists of 22 scales; 4 validity scales, 11 clinical scales, 5 treatment consideration scales, and 2 interpersonal style scales. The current cluster analysis of PAI variables was conducted on 9 of the 11 clinical scales. Alcohol and drug scales were not included since they grouped together as a single factor/component in an initial principle components analysis and had low correlations with other PAI scales. In addition, there was agreement among the neuropsychologists included as authors on this study that these items are often minimized by SMs who are not currently seeking substance abuse treatment and who have not previously been treated for abuse. The treatment consideration and interpersonal style scales were not included as a focus of the present study.

The Posttraumatic Stress Disorder Checklist-Military version (PCL-M) is a self-rated interval-level rating scale used to screen for PTSD (Blanchard, Jones-Alexander, Buckley, & Forneris, 1996; Weathers, Huska, & Keane, 1991). The PCL-M consists of 17 items, each designed to capture a diagnostically relevant symptom in criteria B, C, and D for PTSD in the Diagnostic and Statistical Manual of Mental Disorders, fourth edition text revision, (American Psychiatric Association, 2000). These three clusters are labeled re-experiencing (‘B’ items, 1–5), avoidance or numbing (‘C’ items, 6–12), and hyper-arousal (‘D’ items, 13–17). This self-reported measure requires the subject to rate how much he/she has been bothered by that symptom in the past month, using a 1 (not at all) to 5 (extremely) Likert-scale value. Scores are derived by summing the weighted frequencies for all items and can range from 17 to 85. The PCL-M has been validated for use in samples of military veterans (Wilkins, Lang, & Norman, 2011). For the current study, PCL-M raw scores were used to capture degree of PTSD symptom severity.

The Neurobehavioral Symptom Inventory (NSI) is a 22-item self-report inventory of common post-concussive sequelae (Cicerone & Kalmar, 1995) showing good psychometric properties in a Veteran population (King et al., 2012). Subjects are instructed to rate the presence/severity of each symptom on the NSI within the past 2 weeks on a five-item Likert scale ranging from 0 (None) to 4 (Very Severe). Summing the value of each of the 22 items derives a total score. For the purposes of this study, NSI total score was used for between-group comparisons.

The Test of Memory Malingering (TOMM) (Tombaugh, 1996, 1997) is a well-validated, commonly used performance validity test that employs a forced-choice recognition paradigm using 50 black-and-white line drawings. The test is composed of two recognition trials and a delayed retention trial. For this study, the TOMM was used for inclusion/exclusion screening of participants only, with those scoring <45 correct of 50 items on recognition Trial 2 being excluded from analysis.

Statistical Analysis

To facilitate sufficient statistical power, the sample was divided into two groups: mild (n = 106) and moderate–severe TBI (n = 54). The identification of distinct patterns of PAI subscale profiles was based on methodology described in more detail by Lange, Iverson, Senior, and Chelune (2002). This methodology aims to maximize the influence of profile shape and minimize the influence of profile elevation on the cluster solution through the use of deviation scores.

A two-step cluster analysis procedure (i.e., hierarchical and k-means analyses) was used to identify common profiles in the sample, based on deviation scores. The data was initially subjected to a hierarchical cluster analysis. The average linkage method was used as the cluster algorithm and Pearson correlation as the proximity measure since it is considered to be more sensitive to profile shape and less affected by magnitude of performance (Aldenderfer & Blashfield, 1984; Schinka & Vanderploeg, 1997). Scree plots were generated using the last 50 coefficients of the hierarchical cluster analysis agglomeration schedule. A preliminary cluster solution was determined from visual examination of the scree plots and dendrograms.

Using the visually determined preliminary cluster solutions based on the scree plot and dendrogram, graphic representations of the PAI profiles for each cluster were generated. In addition, graphs were generated for cluster solutions containing incremental numbers of clusters. These graphs were used to assist in visually verifying which solution was most distinct and meaningful. The two profiles in the n + 1 cluster solution that were combined to form one profile in the n cluster solution were examined across incremental numbers of clusters. A cluster solution was rejected when the two merging profiles provided similar interpretations, defined as a correlation greater than r = .45 between the two clusters. A preliminary hierarchical cluster solution was identified based on the maximum number of clusters that provided a meaningful solution.

The number of individuals belonging to each cluster in the preliminary hierarchical solution was examined. In accordance with published recommendations, any cluster with low group membership (<5% of the sample) was excluded from the final cluster solution (Everitt, 1993; Lange et al., 2002; Morris, Blashfield, & Satz, 1981).

Next, a k-means cluster analysis was conducted on the number of clusters identified from the hierarchical analysis, using the default nearest centroid clustering algorithm and Euclidean distance proximity measure. To establish internal validity, the mean PAI t-score profiles from each of the clusters identified by the hierarchical and k-means analyses were compared using a multi-profile multi-method correlation matrix. Internal validity was assessed based on differences in between-method and within-method profile correlations where greater differences indicate better internal validity. Only positive correlations were included, since negative correlations represent dissimilarity between profiles. The cluster solution was either accepted or rejected as a valid representation of inherent subgroups within the data (see Table 3).

Group membership in the final validated cluster solution was presented based upon results of the k-means analysis. This method was preferred over the hierarchical cluster analysis, which does not allow cases to be re-assigned across the clusters at successive steps (Aldenderfer & Blashfield, 1984). Thus, the profiles generated by the k-means analysis were considered to be most representative of the sample.

Traditional analyses of variance and covariance were used for between cluster profile comparisons of continuous variables and chi-square analyses were used for categorical variables.

Results

Description of the Sample

The study sample (n = 160) was >90% male, typical for a group of injured redeployed active duty military SMs (Table 1). The average age of those included in the sample was in the late 20s, with slightly over half of lower enlisted rank (E1–E4). Severity of TBI fell in the mild range for two thirds of the sample. TBI was associated with loss of consciousness in slightly over half of the cases. The majority of the sample was tested within 1 year following injury. The average number of combat deployments served by the participants at the time of evaluation was between one and two, with most being OIF/OEF related (78.7%).

PAI Clinical Scales and TBI

Table 2 shows summary statistics for T scores on the 9 PAI clinical scales included in the present analyses. Data are presented separately for mTBI and moderate-to-severe TBI subgroups, as well as across the entire sample. Participants with mTBI scored significantly higher than those with moderate-to-severe TBI (p < .001) on all PAI scales except the Antisocial Features scale (ANT). The Cohen's effect sizes ranged from d = 0.03 to d = 1.17.

Table 2.

Mean PAI T scores for the total sample and for subgroups with mTBI and moderate–severe TBI

 Total sample mTBI Mod/sev TBI
 
p d 
(n = 160) (n = 106) (n = 54) F 
SOM Somatic Complaints 64.82 68.6 57.41 27 <.001 0.87 
ANX Anxiety 60.33 65.51 50.15 48.1 <.001 1.17 
ARD Anx-Related Disorder 60.89 65.87 51.11 33.9 <.001 0.98 
DEP Depression 65.14 70.64 54.33 46 <.001 1.14 
MAN Mania 54.94 56.68 51.54 .003 0.50 
PAR Paranoia 60.47 63.65 54.22 22.1 <.001 0.79 
SCZ Schizophrenia 60.26 64.92 51.11 39.3 <.001 1.06 
BOR Borderline 59.21 63.58 50.65 42.1 <.001 1.1 
ANT Antisocial Features 57.86 57.74 58.09 0.04 .849 0.03 
 Total sample mTBI Mod/sev TBI
 
p d 
(n = 160) (n = 106) (n = 54) F 
SOM Somatic Complaints 64.82 68.6 57.41 27 <.001 0.87 
ANX Anxiety 60.33 65.51 50.15 48.1 <.001 1.17 
ARD Anx-Related Disorder 60.89 65.87 51.11 33.9 <.001 0.98 
DEP Depression 65.14 70.64 54.33 46 <.001 1.14 
MAN Mania 54.94 56.68 51.54 .003 0.50 
PAR Paranoia 60.47 63.65 54.22 22.1 <.001 0.79 
SCZ Schizophrenia 60.26 64.92 51.11 39.3 <.001 1.06 
BOR Borderline 59.21 63.58 50.65 42.1 <.001 1.1 
ANT Antisocial Features 57.86 57.74 58.09 0.04 .849 0.03 

Note: Figures in bold indicate statistically significant p values associated with between-groups comparisons.

Cluster Analysis of PAI Clinical Scales

Examination of the inverse scree plot from the initial hierarchical analysis indicated a displacement of the curve at 2, 4, and 8 cluster solutions. The dendrogram also supported the presence of 2, 4, or 8 cluster solutions. Examination of the progression of profiles from a two-cluster solution revealed a meaningful merge of profiles from two to three clusters (r = −.14), three to four clusters (r = .14), four to five clusters (r = .28), five to six clusters (r = .26), six to seven clusters (r = .28), seven to eight clusters (r = .28), eight to nine clusters (r = .33), but not nine to ten clusters (r = .50). Nine clusters were therefore identified as the preliminary hierarchical cluster solution. Only four of the nine preliminary clusters included group membership ≥5% of the total sample (i.e., n ≥ 8) and were thus retained. Sample sizes among the included clusters ranged from n = 11 to n = 70 and in the excluded clusters ranged from n = 3 to n = 7. A k-means cluster analysis was applied to the data with a four cluster solution specified in the analysis. Examination of the multi-profile multi-method correlation matrix (Table 3) revealed that three of the four clusters generated by the hierarchical analysis were highly correlated with three corresponding clusters generated by the k-means analysis (range: r = .953 to r = .997). The remaining cluster generated by the hierarchical analysis was moderately correlated with a corresponding cluster generated by the k-means analysis (r = .566). The correlations between non-corresponding clusters were low to moderate (range: r = .013 to r = .466), indicating a high degree of internal validity across cluster analytic techniques when applied to this sample. Based on examination of the mean PAI clinical subscale scores, the four clusters were labeled: High Distress, Somatic Distress, Moderate Distress and No Distress (see Fig. 1).

Table 3.

Multi-profile multi-method correlation matrix

 Hierarchical (H) profiles
 
k-means (K) profiles
 
H1 H2 H3 H4 K1 K2 K3 K4 
H1        
H2 0.219       
H3 −0.055 0.263      
H4 −0.856 0.011 0.374     
K1 0.995 0.214 −0.077 −0.880    
K2 0.214 0.953 0.100 0.013 0.209   
K3 0.466 0.061 0.566 −0.169 0.406 −0.122  
K4 −0.850 0.014 0.400 0.997 −0.876 −0.001 −0.132 
 Hierarchical (H) profiles
 
k-means (K) profiles
 
H1 H2 H3 H4 K1 K2 K3 K4 
H1        
H2 0.219       
H3 −0.055 0.263      
H4 −0.856 0.011 0.374     
K1 0.995 0.214 −0.077 −0.880    
K2 0.214 0.953 0.100 0.013 0.209   
K3 0.466 0.061 0.566 −0.169 0.406 −0.122  
K4 −0.850 0.014 0.400 0.997 −0.876 −0.001 −0.132 

Notes: N = 160 (Hierarchical: 70 High Distress, 28 Somatic Distress, 11 Moderate Distress, 27 No Distress; k-means: 51 High Distress, 35 Somatic Distress, 43 Moderate Distress, and 31 No Distress). Figures in bold indicate statistically significant correlations.

Fig. 1.

Cluster-analysis-based PAI Clinical Scale Profiles, where PAI = Personality Assessment Inventory; SOM = Somatic Complaints; ANX = Anxiety; ARD = Anxiety-Related Disorders; DEP = Depression; MAN = Mania; PAR = Paranoia; SCZ = Schizophrenia; BOR = Borderline Features; ANT = Antisocial Features.

Fig. 1.

Cluster-analysis-based PAI Clinical Scale Profiles, where PAI = Personality Assessment Inventory; SOM = Somatic Complaints; ANX = Anxiety; ARD = Anxiety-Related Disorders; DEP = Depression; MAN = Mania; PAR = Paranoia; SCZ = Schizophrenia; BOR = Borderline Features; ANT = Antisocial Features.

Description of the Cluster-derived PAI Profiles

Mean clinical scale PAI scores for the four cluster-derived profile groups are presented in Table 4. Individuals in the High Distress profile (n = 51, 32%) showed elevations on Somatic Complaints (SOM; mean T = 74), Anxiety (ANX; mean T = 74), Anxiety-Related Disorders (ARD; mean T = 75), Depression (DEP; mean T = 79), and Schizophrenia (SCZ; mean T = 71). The Somatic Distress profile (n = 35, 22%) included a single mild elevation in the Somatic Complaints scale (SOM; mean T = 69). The prototypical individual fitting the Moderate Distress profile had low-level elevations on several PAI scales including Depression (DEP; mean T = 66), Paranoia (PAR; mean T = 65), and Schizophrenia (SCZ; mean T = 64). Average T scores on the Borderline Features (BOR), Antisocial Features (ANT) and Anxiety-related Disorders (ARD) scales fell just >60 for this profile group. A single mild elevation in the ANT (mean T = 67) characterized the No Distress profile, with T scores falling within the normal range (T < 60) on other clinical scales of the PAI. Of note, the mean Mania (MAN) scale score was uniformly within the average range (T between 54 and 57) across profile clusters. This is consistent with PAI normative sample data, for which “elevations on the full scale of MAN tend to be rarer in clinical settings than are any of the other clinical scales of the PAI” (Morey, 2003, p. 91).

Table 4.

Mean PAI clinical scale T scores for the four cluster analysis-based profile groups

 High distress Somatic distress Moderate distress No distress 
(n = 51) (n = 35) (n = 43) (n = 31) 
SOM Somatic Complaints 73.80 69.34 59.93 51.74 
ANX Anxiety 74.49 53.26 58.47 47.58 
ARD Anxiety-Related Disorders 75.35 53.06 60.16 46.94 
DEP Depression 79.35 57.34 66.21 49.06 
MAN Mania 53.71 56.29 53.84 57.00 
PAR Paranoia 64.00 54.49 64.70 55.55 
SCZ Schizophrenia 70.67 51.71 63.77 47.94 
BOR Borderline Features 66.43 53.43 61.23 51.06 
ANT Antisocial Features 51.96 53.57 61.60 67.19 
 High distress Somatic distress Moderate distress No distress 
(n = 51) (n = 35) (n = 43) (n = 31) 
SOM Somatic Complaints 73.80 69.34 59.93 51.74 
ANX Anxiety 74.49 53.26 58.47 47.58 
ARD Anxiety-Related Disorders 75.35 53.06 60.16 46.94 
DEP Depression 79.35 57.34 66.21 49.06 
MAN Mania 53.71 56.29 53.84 57.00 
PAR Paranoia 64.00 54.49 64.70 55.55 
SCZ Schizophrenia 70.67 51.71 63.77 47.94 
BOR Borderline Features 66.43 53.43 61.23 51.06 
ANT Antisocial Features 51.96 53.57 61.60 67.19 

Demographic and Injury Comparisons Across Profile Groups

Comparisons of demographic and injury variables (Tables 5 and 6) revealed significant differences in age (p = .002, η2 = .091), ethnicity (p = .005), rank (p < .001), and TBI severity (p < .001) across the four personality profiles. Members of the High Distress group were significantly older than the Moderate Distress group (p = .004, d = 0.78) and No Distress group (p = .012, d = 0.69). Significant chi-square analysis results highlight that the High Distress group was composed of more high-ranking SMs and greater proportions of mTBI versus moderate or severe TBI than the other groups. Although between-profile differences in gender composition were not statistically significant, the High Distress profile was composed of somewhat more females than the other profiles. Individuals with the Somatic Distress profile were on average evaluated later postinjury, but with high variability across the group in time from injury to evaluation. Individuals fitting this profile were not distinguished by unique demographic characteristics compared with those included in other profile groups. Compared with the other groups, patients with the Moderate Distress profile were more likely to be Caucasian than other races/ethnicities. Proportionately more individuals with the No Distress profile experienced moderate–severe TBI, and were of lower enlisted rank.

Table 5.

Demographic characteristics for the four cluster analysis-based profile groups

 High distress Somatic distress Moderate distress No distress F p 
M (SDM (SDM (SDM (SD
Age 31.4 (6.4) 28.1 (6.8) 26.7 (5.5) 26.7 (7.6) 3.9 .002 
Education 12.8 (1.7) 12.6 (1.4) 12.7 (1.6) 12.6 (1.3) 0.25 .863 
Deployments 1.7 (1.0) 1.5 (1.0) 1.6 (1.0) 1.6 (1.6) 0.53 .748 
 N (%) N (%) N (%) N (%) p 
Gender 
 Male 46 (90.2) 34 (97.1) 41 (95.3) 30 (96.8) .562 
 Female 5 (9.8) 1 (2.9) 2 (4.7) 1 (3.2)   
Ethnicity 
 Caucasian 30 (58.8) 21 (60.0) 37 (86.0) 16 (51.6) .005 
 Other 21 (41.2) 14 (40.0) 6 (14.0) 15 (48.4)   
Deployment Location 
 OIF/OEF 39 (76.5) 29 (82.9) 34 (79.1) 24 (77.4) .924 
 Other 12 (23.5) 6 (17.1) 9 (20.9) 7 (22.6)   
Rank 
 E1–E4 16 (31.4) 20 (57.1) 26 (60.5) 26 (83.9) <.001 
 E5+ 35 (68.6) 15 (42.9) 17 (39.5) 5 (16.1)   
 High distress Somatic distress Moderate distress No distress F p 
M (SDM (SDM (SDM (SD
Age 31.4 (6.4) 28.1 (6.8) 26.7 (5.5) 26.7 (7.6) 3.9 .002 
Education 12.8 (1.7) 12.6 (1.4) 12.7 (1.6) 12.6 (1.3) 0.25 .863 
Deployments 1.7 (1.0) 1.5 (1.0) 1.6 (1.0) 1.6 (1.6) 0.53 .748 
 N (%) N (%) N (%) N (%) p 
Gender 
 Male 46 (90.2) 34 (97.1) 41 (95.3) 30 (96.8) .562 
 Female 5 (9.8) 1 (2.9) 2 (4.7) 1 (3.2)   
Ethnicity 
 Caucasian 30 (58.8) 21 (60.0) 37 (86.0) 16 (51.6) .005 
 Other 21 (41.2) 14 (40.0) 6 (14.0) 15 (48.4)   
Deployment Location 
 OIF/OEF 39 (76.5) 29 (82.9) 34 (79.1) 24 (77.4) .924 
 Other 12 (23.5) 6 (17.1) 9 (20.9) 7 (22.6)   
Rank 
 E1–E4 16 (31.4) 20 (57.1) 26 (60.5) 26 (83.9) <.001 
 E5+ 35 (68.6) 15 (42.9) 17 (39.5) 5 (16.1)   

Notes: Figures in bold indicate statistically significant p values associated with between groups comparisons. OIF = Operation Iraqi Freedom; OEF = Operation Enduring Freedom; TBI = traumatic brain injury; LOC = loss of consciousness; PTA = posttraumatic amnesia; PCL-M =Posttraumatic Stress Checklist Military Version; NSI = Neurobehavioral Symptom Inventory.

Table 6.

Injury-related characteristics and symptom ratings for the four cluster analysis-based profile groups

 High Distress Somatic Distress Moderate Distress No Distress F p 
M (SDM (SDM (SDM (SD
Time Since Injury (weeks) 24.0 (16.0) 45.9 (53.3) 27.8 (21.2) 27.5 (53.5) 1.69 .172 
NSI total score 48.9 (14.4) 31.0 (14.9) 33.5 (15.9) 13.7 (12.5) 24.21a <.001 
PCL-M total score 59.4 (13.3) 35.8 (14.8) 42.5 (16.7) 26.0 (11.9) 21.83a <.001 
  
 N (%) N (%) N (%) N (%) p  
LOC (N = 158) 
 Yes 26 (52.0) 16 (45.7) 24 (57.1) 18 (58.1) .701  
 No 20 (40.0) 16 (45.7) 15 (35.7) 13 (41.9)   
 Unknown 4 (8.0) 3 (8.6) 3 (7.2) 0 (0.0)   
PTA (N = 142) 
 Yes 4 (8.7) 7 (21.9) 6 (16.2) 4 (15.4) .557  
 No 17 (37.0) 14 (43.8) 17 (45.9) 11 (42.3)   
 Unknown 25 (54.3) 11 (34.3) 14 (37.9) 11 (42.3)   
TBI Severity 
 Mild 45 (88.2) 23 (65.7) 28 (65.1) 10 (32.3) <.001  
 Moderate/severe 6 (11.8) 12 (34.3) 15 (34.9) 21 (67.7)   
Mechanism of Injury 
 Blast 25 (49.0) 18 (51.4) 25 (58.1) 18 (58.1) .781 
 Non-blast 26 (51.0) 17 (48.6) 18 (41.9) 13 (41.9) 
 High Distress Somatic Distress Moderate Distress No Distress F p 
M (SDM (SDM (SDM (SD
Time Since Injury (weeks) 24.0 (16.0) 45.9 (53.3) 27.8 (21.2) 27.5 (53.5) 1.69 .172 
NSI total score 48.9 (14.4) 31.0 (14.9) 33.5 (15.9) 13.7 (12.5) 24.21a <.001 
PCL-M total score 59.4 (13.3) 35.8 (14.8) 42.5 (16.7) 26.0 (11.9) 21.83a <.001 
  
 N (%) N (%) N (%) N (%) p  
LOC (N = 158) 
 Yes 26 (52.0) 16 (45.7) 24 (57.1) 18 (58.1) .701  
 No 20 (40.0) 16 (45.7) 15 (35.7) 13 (41.9)   
 Unknown 4 (8.0) 3 (8.6) 3 (7.2) 0 (0.0)   
PTA (N = 142) 
 Yes 4 (8.7) 7 (21.9) 6 (16.2) 4 (15.4) .557  
 No 17 (37.0) 14 (43.8) 17 (45.9) 11 (42.3)   
 Unknown 25 (54.3) 11 (34.3) 14 (37.9) 11 (42.3)   
TBI Severity 
 Mild 45 (88.2) 23 (65.7) 28 (65.1) 10 (32.3) <.001  
 Moderate/severe 6 (11.8) 12 (34.3) 15 (34.9) 21 (67.7)   
Mechanism of Injury 
 Blast 25 (49.0) 18 (51.4) 25 (58.1) 18 (58.1) .781 
 Non-blast 26 (51.0) 17 (48.6) 18 (41.9) 13 (41.9) 

Notes: Figures in bold indicate statistically significant p values associated with between groups comparisons. OIF = Operation Iraqi Freedom; OEF = Operation Enduring Freedom; TBI = traumatic brain injury; LOC = loss of consciousness; PTA = posttraumatic amnesia; PCL-M = Posttraumatic Stress Checklist Military Version; NSI = Neurobehavioral Symptom Inventory.

aF-values represent ANCOVA, controlling for age, rank, ethnicity, and TBI severity.

Postconcussion/PTSD Symptom Comparisons Across Profile Groups

Controlling for age, ethnicity, rank, and TBI severity, a main effect of personality profile group on postconcussion symptom severity (NSI total score) was found (p < .001, η2 = .323) (Table 6). Bonferroni-corrected post hoc pairwise comparisons revealed significant differences (p < .00, d = 1.04–2.6) between all but the Somatic Distress and Moderate Distress groups. Postconcussion symptom endorsement generally followed a graded pattern with High Distress > Somatic Distress, Moderate Distress > No Distress. An almost identical pattern of results was found for PTSD symptoms (PCL-M total score), with a main effect of personality profile group (p < .001, η2 = .301) and Bonferroni-corrected post hoc analyses showing significant differences (p < .001, d = 0.73–2.64) between all but the Somatic Distress and Moderate Distress groups (see Table 4). PTSD symptom endorsement also followed a graded pattern with High Distress > Somatic Distress, Moderate Distress > No Distress.

Discussion

The purpose of this study was to examine symptom profiles in SMs who sustained TBI. Using a two-step cluster-analytic method of PAI clinical scale scores, four distinct profiles were detected in the current sample of active duty military SMs. We descriptively labeled these groups High Distress, Somatic Distress, Moderate Distress, and No Distress for heuristic purposes. Severity of endorsement of both postconcussive and posttraumatic stress symptoms differed across profiles. The current study is the first to perform cluster analysis of the PAI exclusively in military SMs with TBI. Findings indicate that valid and meaningful patterns of symptomatic and behavioral characteristics can be identified in this sample. Within the framework of burden of adversity, and taking into account the symptom pattern and personal characteristics of each group, specific treatment recommendations are proposed.

Based on the burden of adversity/cumulative disadvantage model, the High Distress group is experiencing a negative recovery trajectory, characterized by the presence of a variety of symptoms, poor functioning and high disability. This profile group, comprising approximately one third of the present sample, is easily identified by providers in the DoD and VA healthcare systems (Denning et al., 2014). These individuals are generally high healthcare utilizers, many of whom have functionally significant symptoms, multiple concurrent somatic injuries, and psychiatric disorders, including mood, anxiety, PTSD, acute stress and, substance use disorders (Meyer, Marion, Coronel, & Jaffee, 2010). In a redeployment military sample, individuals fitting this profile may have experienced recurrent combat-related traumatic events and/or multiple concussions. Although these trauma- and TBI-related data were not available to this investigation, the individuals in this group were somewhat older and higher ranking than those in the other profile groups, thereby increasing the time in service and likelihood of exposure to recurrent traumatic and concussive events. This group may also include proportionately more SMs who received traumatic amputations, burns, and/or orthopedic trauma, leading to protracted pain, sleep disturbances and other somatic symptoms. According to the cumulative disadvantage model, such chronic symptoms can become exacerbated, leading to other physical and psychological conditions and poor outcomes. The types of treatments warranted include multidisciplinary team approaches that address the range of symptoms present in these patients. As discussed by Brenner and colleagues (2009), early assessment and treatment of the symptoms regardless of etiology is indicated to reduce this burden of adversity. However, empirical support is lacking for specific treatment of individuals with this globally negative outcome at a more chronic stage post-injury.

The High Distress group is composed of proportionately more individuals who sustained mTBI. Although this may seem counterintuitive, it has been shown that mTBI is often associated with higher subjective distress and symptom report than moderate-to-severe TBI, due in part to impaired self-awareness among individuals with more severe TBI. This group also endorsed elevated levels of symptoms associated with ongoing psychiatric conditions such as PTSD, which have been shown to play a major role in elevating overall symptom ratings as well as negatively affecting outcomes following TBI (Hoge et al., 2008; Malec et al., 2007).

In order to reverse the chronic effects, multidisciplinary intensive treatment addressing the myriad of symptoms may be necessary for individuals in the high distress group. A major component of this type of treatment approach should include addressing attitudes and expectations, to empower and activate the individual to make significant positive changes in his/her overall level of function. Components of treatment could also address the clinically significant levels of somatic and emotional distress found among individuals in this group. Based on the elevations in the PAI clinical scales as well as the scores on the NSI and PCL-M symptom inventories, co-existing diagnoses of PTSD, other anxiety disorders, and/or a depressive disorder are likely. Standard pharmacotherapy and psychotherapy treatments for these disorders can be expected to benefit individuals in this highly distressed group. Effective treatment of these emotional components may secondarily reduce somatic and cognitive symptoms, lowering the overall level of distress and changing the recovery trajectory. Individuals in this high distress group would not be expected to experience a reduction in symptoms without behavioral health intervention.

The Somatic and Moderate Distress groups are not characterized by global distress and not clearly on a negative recovery trajectory. They may be seen as two different patterns of response to stress. The Somatic Distress group tends to respond in a way characteristic of somatoform disorders, focusing on physical symptoms and disability, while minimizing emotional distress. Recommended treatment goals for this group involves understanding, acceptance, and control of concerns about physical symptoms and functioning. A key aspect is reducing anxiety that drives this somatic overconcern. Lind, Delmar, and Nielsen (2014) have recently conceptualized treatment among individuals with somatoform type disorders in terms of controlling existential insecurity. They report beneficial effects of mindfulness group treatment for somatoform disorders including the perception and recognition by the individual that he/she is “really” ill, ability to achieve relaxation via mindfulness techniques, increased awareness of mind–body connections and improved ability to identify and express needs and feelings of distress by utilizing more active methods of coping. A recent review of pharmacotherapy treatment for somatoform disorders (Kleinstäuber et al., 2014) found insufficient evidence for specific medication treatments in this population. They also identified significant methodological concerns in the available studies, including sample bias, heterogeneity in the data, and small sample sizes. Nonetheless, antidepressants, atypical antidepressants, and anxiolytics are used in the somatoform clinical population and may be of benefit to some individuals in the Somatic Distress group, particularly those with SOM scale elevations in the clinical range (i.e., T > 70). In addition, pharmacotherapy for postconcussive somatic symptoms such as headaches is recommended for this group (Defense and Veterans Brain Injury Center, 2015; Theeler, Lucas, Riechers, & Ruff, 2013).

The Moderate Distress group is primarily distinguished by sub-clinical elevations on DEP and PAR. The prototype of this group is an individual who has become chronically dysthymic, frequently pessimistic, and interpersonally disconnected. These individuals tend to blame themselves for their problems but feel powerless to effect change. Treatments that are effective for depressive disorders are recommended for these individuals (Picardi & Gaetano, 2014). Although these individuals are not experiencing functional distress to a clinically significant degree, it is expected that they could benefit from common cognitive-behavioral and interpersonal treatments for depression. Individuals in this group need opportunities to experience success. Controlled and supportive opportunities for engagement in social and interpersonal activities are recommended.

Standard TBI treatments may be of benefit, particularly for individuals in the Somatic Distress and Moderate Distress groups. Although empirical evidence of their efficacy at a more chronic phase of recovery is lacking, the levels of distress present among these groups are not too severe to prohibit productive engagement in psychotherapeutic and psychoeducational interventions. These TBI-specific approaches include education regarding expectations for recovery and implementation of components of the stepwise treatment of TBI symptoms provided by Terrio and colleagues (2009) and incorporated into the current VA/DoD clinical practice guidelines for TBI care (VA/DoD Clinical Practice Guideline: Management of Concussion/Mild Traumatic Brain Injury, 2009). Both patient and family education regarding positive expectations for recovery along with addressing psychiatric symptoms, somatic symptoms such as headaches and self-care issues including sleep hygiene and appropriate diet and exercise are addressed. If cognitive symptoms such as problems with attention or memory are present, neuropsychological assessment is recommended to guide rehabilitative therapy goals. Often in the chronic stage post-TBI, cognitive inefficiencies are derived from nonspecific effects of emotional distress, pain, and fatigue. Once these factors are treated, there is often significant improvement in subjective and objective cognitive symptoms. In addition, monitoring symptoms over time is an important ongoing activity in order to avoid elevations in overall level of distress and development of a generalized negative recovery trajectory.

Individuals in the Low Distress profile group (who did not endorse elevated symptoms or concerns related to either physical or emotional health, but had sub-clinical elevation of the antisocial scale) appear to represent one of several distinct subgroups: those who are clinically recovered, those at risk for maladaptive behavior and those with moderate-to-severe TBI showing lack of deficit awareness. Clinical intervention for patients with this profile configuration therefore starts with effective screening/triage for subgroup membership. Individuals who have experienced symptomatic remission with ANT scores in the lower range for this group (i.e., mid-60T and lower) appear to represent a subgroup of clinically recovered individuals. Mild elevations in ANT (T scores between 60 and 70) are not uncommon among young adult males (Morey, 2003). However, post-deployment violence and other maladaptive characteristics among military SMs have been related to earlier history of antisocial behaviors (MacManus et al., 2012). For these individuals, interventions aimed at treatment of externally directed (i.e., aggression, violence) and internally directed (i.e., substance use disorders, self-injurious behaviors) negative behaviors are indicated. These individuals will generally be challenging in treatment, particularly those with a preponderance of externalizing behaviors, as this latter subgroup characteristically display relatively low levels of introspection and motivation for self-change, as well as difficulties in establishing interpersonal therapeutic rapport. Other individuals in the low distress profile group represent those who received moderate-to-severe TBI. Their low endorsement of problems and symptoms likely reflects lack of deficit awareness. Further examination of the item scores contributing to the subclinical elevations on ANT and the level of elevations among those with moderate-to-severe TBI would be informative, but was beyond the scope of this investigation. Importantly, individuals who had moderate-to-severe TBI would need to be evaluated for their level of awareness of their difficulties. Those with worse insight could be candidates for rehabilitative efforts, especially treatments that focus on increasing self-awareness.

As with any study, this investigation has strengths and weaknesses. Research has shown that compensation seeking in the context of mTBI elevates somatic, depressive, and anxiety scales on the PAI (Whiteside et al., 2012). A major strength of the present study is that individuals who failed either symptom validity tests (SVTs) or performance validity tests (PVTs) were omitted from the sample. Findings suggest that, even when these individuals are omitted, a large proportion of military TBI patients show elevations characteristic of a high distress profile on the PAI. While it remains possible that some of the individuals reported in this retrospective study were undergoing a military medical board (i.e., military disability/compensation-related evaluation), this subgroup endorsed a significantly elevated number of symptoms, without a corresponding failure of SVTs/PVTs. Another strength of this study is the selection of cases from two major military treatment facilities, one is a large Army deployment base and the other is a major military medical center. Personnel treated at these facilities are representative of OIF/OEF SMs with TBI. The use of the PAI in this study allowed us to examine symptoms and behavioral characteristics across multiple domains, an advantage over the use of symptom ratings limited to a single domain such as postconcussive, depressive or PTSD symptoms. A weakness of this study is the use of a self-rating instrument that might not correspond with interview or clinician-derived ratings. The broad range of time post-injury and inclusion of patients with multiple other clinical syndromes and disorders are both benefits and drawbacks of this study. A range of time post-injury allows generalizability of the results to patients ranging from the sub-acute to chronic stages post-injury. However, the sample size was not large enough to analyze subgroups by time since injury. Since symptom profile patterns may be dynamic and change with time, this is an important issue unanswered by the present study. Including all patients with TBI referred for evaluation without exclusions based on other clinical issues also enhances generalizability of the findings, but again, sample size limitations did not allow analysis of subgroups. For example, it is not known whether individuals with TBI plus PTSD show different PAI profiles than those with TBI and depression or TBI with no concurrent psychiatric disorders. Basic data regarding whether or not SMs are compensation seeking is important in interpreting whether patients report higher distress. However, due to the retrospective archival nature of this study, these data were not available.

In summary, present results suggest that several distinct chronic symptom profile patterns are exhibited by a heterogeneous group of military SMs with TBI. Four different profiles of symptom patterns based on clinical scales of the PAI were identified. These profiles appear to correspond with previous findings of PAI profiles in civilian TBI patients (Velikonja et al., 2010) and may thus represent characteristic postconcussive symptom patterns. Further research is needed to verify these findings and analyze PAI symptom patterns in larger separate groups of patients with mTBI and moderate-to-severe TBI. Longitudinal analysis of PAI profile changes across time post-injury is an area of potentially fruitful research focus. Finally, present results have implications for selecting treatment and rehabilitation strategies for TBI patients with differing symptom profiles and level of distress. Although one would not base treatment plans solely on the interpretation of a personality inventory, in cases of TBI where treatment is predominantly symptom based, an examination of the PAI symptom profile pattern may suggest a more strategic empirically based approach to care.

Funding

This work was supported by the Defense and Veterans Brain Injury Center and the Henry M. Jackson Foundation for the Advancement of Military Medicine and General Dynamics Information Technology.

Conflict of Interest

None declared.

Acknowledgements

Portions of these data were presented at the 41st Annual Meeting of the International Neuropsychological Society in Waikoloa, Hawaii: February 2013 and used in analyses reported in Vanderploeg et al. (2014).

The authors acknowledge the administrative, clinical and technical assistance of Dr. Alan Hopewell, Dr. Nancy Craig, Dr. Annie Chang, and Nicholas Elliotte from the TBI clinic, Darnall Army Medical Center, Ft Hood TX. We also acknowledge the administrative support of Dr. Amy Bowles from the Brain Injury Rehabilitation Service, Brooke Army Medical Center. We thank Ricardo Amador and Maren Cullen from the Defense and Veterans Brain Injury Center, Brooke Army Medical Center for their invaluable assistance in compiling and verifying the data used in this study.

The views expressed in this manuscript are those of the authors and do not reflect the official policy of the Department of the Army, Department of Defense, or the U.S. Government.

References

Aldenderfer
F. A.
Blashfield
R.
(
1984
).
Cluster analysis
 .
Beverly Hills, CA
:
Sage Publications
.
American Psychiatric Association
. (
2000
).
Diagnostic and statistical manual of mental disorders, (fourth edition—text revision)
 .
Washington, DC
:
Author
.
Betthauser
L. M.
Bahraini
N.
Krengel
M. H.
Brenner
L. A.
(
2012
).
Self-report measures to identify post traumatic stress disorder and/or mild traumatic brain injury and associated symptoms in military veterans of Operation Enduring Freedom (OEF)/Operation Iraqi Freedom (OIF)
.
Neuropsychology Review
 ,
22
(1)
,
35
53
.
Blanchard
E. B.
Jones-Alexander
J.
Buckley
T. C.
Forneris
C. A.
(
1996
).
Psychometric properties of the PTSD Checklist (PCL)
.
Behaviour Research and Therapy
 ,
34
(8)
,
669
673
.
Brenner
L. A.
Vanderploeg
R. D.
Terrio
H.
(
2009
).
Assessment and diagnosis of mild traumatic brain injury, posttraumatic stress disorder, and other polytrauma conditions: Burden of adversity hypothesis
.
Rehabilitation Psychology
 ,
54
(3)
,
239
246
.
Cicerone
K. D.
Kalmar
K.
(
1995
).
Persistent postconcussion syndrome: The structure of subjective complaints after mild traumatic brain injury
.
Journal of Head Trauma Rehabilitation
 ,
10
(3)
,
1
17
.
Defense and Veterans Brain Injury Center
. (
2014
).
DoD worldwide TBI numbers
. .
Defense and Veterans Brain Injury Center
. (
2015
).
Progressive return to activity following acute concussion/mild traumatic brain injury
.
Washington, DC
. .
Demakis
G. J.
Hammond
F.
Knotts
A.
Cooper
D. B.
Clement
P.
Kennedy
J.
et al
(
2007
).
The Personality Assessment Inventory in individuals with traumatic brain injury
.
Archives of Clinical Neuropsychology
 ,
22
(1)
,
123
130
.
Denning
L. A.
Meisnere
M.
Warner
K. E.
,
Committee on the Assessment of Resiliency and Prevention Programs for Mental and Behavioral Health in Service Members and Their Families
,
Board on the Health of Select Populations
,
Institute of Medicine
(
2014
).
Preventing psychological disorders in service members and their families: An assessment of programs
 .
Washington
,
DC
:
National Academies Press
.
Everitt
B. S.
(
1993
).
Cluster analysis
  (
3rd ed.
).
London
:
John Wiley & Sons
.
Hesdorffer
D. C.
Rauch
S. L.
Tamminga
C. A.
(
2009
).
Long-term psychiatric outcomes following traumatic brain injury: A review of the literature
.
Journal of Head Trauma Rehabilitation
 ,
24
(6)
,
452
459
.
Hoge
C. W.
McGurk
D.
Thomas
J. L.
Cox
A. L.
Engel
C. C.
Castro
C. A.
(
2008
).
Mild traumatic brain injury in U.S. Soldiers returning from Iraq
.
New England Journal of Medicine
 ,
358
(5)
,
453
463
.
Hou
R.
Moss-Morris
R.
Peveler
R.
Mogg
K.
Bradley
B. P.
Belli
A.
(
2012
).
When a minor head injury results in enduring symptoms: A prospective investigation of risk factors for postconcussional syndrome after mild traumatic brain injury
.
Journal of Neurology, Neurosurgery and Psychiatry
 ,
83
(2)
,
217
223
.
King
N. S
. (
2008
).
PTSD and traumatic brain injury: Folklore and fact?
Brain Injury
 ,
22
(1)
,
1
5
.
King
P. R.
Donnelly
K. T.
Donnelly
J. P.
Dunnam
M.
Warner
G.
Kittleson
C. J.
et al
(
2012
).
Psychometric study of the Neurobehavioral Symptom Inventory
.
Journal of Rehabilitation Research and Development
 ,
49
(6)
,
879
888
.
Kleinstäuber
M.
Witthöft
M.
Steffanowski
A.
van Marwijk
H.
Hiller
W.
Lambert
M. J.
(
2014
)
Pharmacological interventions for somatoform disorders in adults
.
The Cochrane Database of Systematic Reviews
 ,
11
,
CD010628
.
Kurtz
J. E.
Shealy
S. E.
Putnam
S. H.
(
2007
).
Another look at paradoxical severity effects in head injury with the Personality Assessment Inventory
.
Journal of Personality Assessment
 ,
88
(1)
,
66
73
.
Lange
R. T.
Iverson
G.
Senior
G.
Chelune
G. J.
(
2002
).
A primer on cluster analysis applications to cognitive rehabilitation research
.
Journal of Cognitive Rehabilitation
 ,
20
(1)
,
16
33
.
Lange
R. T.
Iverson
G. L.
Brooks
B. L.
Rennison
V. L.
(
2010
).
Influence of poor effort on self-reported symptoms and neurocognitive test performance following mild traumatic brain injury
.
Journal of Clinical and Experimental Neuropsychology
 ,
32
(9)
,
961
972
.
Lew
H. L.
Vanderploeg
R. D.
Moore
D. F.
Schwab
K.
Friedman
L.
Yesavage
J.
et al
(
2008
).
Overlap of mild TBI and mental health conditions in returning OIF/OEF service members and veterans
.
Journal of Rehabilitation Research and Development
 ,
45
(3)
,
xi
xvi
.
Lind
A. B.
Delmar
C.
Nielsen
K.
(
2014
).
Searching for existential security: A prospective qualitative study on the influence of mindfulness therapy on experienced stress and coping strategies among patients with somatoform disorders
.
Journal of Psychosomatic Research
 ,
77
(6)
,
516
521
.
Macera
C. A.
Aralis
H. J.
Macgregor
A. J.
Rauh
M. J.
Galarneau
M. R.
(
2012
).
Postdeployment symptom changes and traumatic brain injury and/or posttraumatic stress disorder in men
.
Journal of Rehabilitation Research and Development
 ,
49
(8)
,
1197
1208
.
MacManus
D.
Dean
K.
Iversen
A. C.
Hull
L.
Jones
N.
Fahy
T.
et al
(
2012
).
Impact of pre-enlistment antisocial behaviour on behavioural outcomes among U.K. military personnel
.
Social Psychiatry and Psychiatric Epidemiology
 ,
47
(8)
,
1353
1358
.
Malec
J. F.
Testa
J. A.
Rush
B. K.
Brown
A. W.
Moessner
A. M.
(
2007
).
Self-assessment of impairment, impaired self-awareness, and depression after traumatic brain injury
.
Journal of Head Trauma Rehabilitation
 ,
22
(3)
,
156
166
.
McMahon
P.
Hricik
A.
Yue
J. K.
Puccio
A. M.
Inoue
T.
Lingsma
H. F.
et al
(
2014
).
Symptomatology and functional outcome in mild traumatic brain injury: Results from the prospective TRACK-TBI study
.
Journal of Neurotrauma
 ,
31
(1)
,
26
33
.
Merton
R. K
. (
1968
).
The Matthew effect in science: The reward and communication systems of science are considered
.
Science
 ,
159
(3810)
,
56
63
.
Meyer
K. S.
Marion
D. W.
Coronel
H.
Jaffee
M. S.
(
2010
).
Combat-related traumatic brain injury and its implications to military healthcare
.
Psychiatric Clinics of North America
 ,
33
(4)
,
783
796
.
Morey
L. C.
(
1991
).
The Personality Assessment Inventory professional manual
 .
Odessa, FL
:
Psychological Assessment Resources
.
Morey
L. C.
(
2003
).
Essentials of PAI assessment
 .
New York
:
John Wiley
.
Morey
L. C.
Lowmaster
S. E.
Coldren
R. L.
Kelly
M. P.
Parish
R. V.
Russell
M. L.
(
2011
).
Personality Assessment Inventory profiles of deployed combat troops: An empirical investigation of normative performance
.
Psychological Assessment
 ,
23
(2)
,
456
462
.
Morissette
S. B.
Woodward
M.
Kimbrel
N. A.
Meyer
E. C.
Kruse
M. I.
Dolan
S.
et al
(
2011
).
Deployment-related TBI, persistent postconcussive symptoms, PTSD, and depression in OEF/OIF veterans
.
Rehabilitation Psychology
 ,
56
(4)
,
340
350
.
Morris
R.
Blashfield
R.
Satz
P.
(
1981
).
Neuropsychology and cluster analysis: Potentials and problems
.
Journal of Clinical Neuropsychology
 ,
3
(1)
,
79
99
.
Otis
J. D.
McGlinchey
R.
Vasterling
J. J.
Kerns
R. D.
(
2011
).
Complicating factors associated with mild traumatic brain injury: Impact on pain and posttraumatic stress disorder treatment
.
Journal of Clinical Psychology in Medical Settings
 ,
18
(2)
,
145
154
.
Picardi
A.
Gaetano
P.
(
2014
).
Psychotherapy of mood disorders
.
Clinical Practice and Epidemiology in Mental Health
 ,
10
,
140
158
.
Pietrzak
R. H.
Johnson
D. C.
Goldstein
M. B.
Malley
J. C.
Southwick
S. M.
(
2009
).
Posttraumatic stress disorder mediates the relationship between mild traumatic brain injury and health and psychosocial functioning in veterans of operations enduring freedom and Iraqi freedom
.
Journal of Nervous and Mental Disease
 ,
197
(10)
,
748
753
.
Piotrowski
C.
(
2000
).
How popular is the Personality Assessment Inventory in practice and training?
Psychological Reports
 ,
86
(1)
,
65
66
.
Ruff
R. M.
Camenzuli
L.
Mueller
J
. (
1996
).
Miserable minority: Emotional risk factors that influence the outcome of a mild traumatic brain injury
.
Brain Injury
 ,
10
(8)
,
551
565
.
Schinka
J. A.
Vanderploeg
R. D.
(
1997
).
Profile clusters in the WAIS-R standardization sample
.
Journal of the International Neuropsychological Society
 ,
3
(2)
,
120
127
.
Schneiderman
A. I.
Braver
E. R.
Kang
H. K.
(
2008
).
Understanding sequelae of injury mechanisms and mild traumatic brain injury incurred during the conflicts in Iraq and Afghanistan: Persistent postconcussive symptoms and posttraumatic stress disorder
.
American Journal of Epidemiology
 ,
167
(12)
,
1446
1452
.
Seguin
M.
Lesage
A.
Turecki
G.
Bouchard
M.
Chawky
N.
Tremblay
N.
et al
(
2007
).
Life trajectories and burden of adversity: Mapping the developmental profiles of suicide mortality
.
Psychological Medicine
 ,
37
(11)
,
1575
1583
.
Taylor
B. C.
Hagel
E. M.
Carlson
K. F.
Cifu
D. X.
Cutting
A.
Bidelspach
D. E.
et al
(
2012
).
Prevalence and costs of co-occurring traumatic brain injury with and without psychiatric disturbance and pain among Afghanistan and Iraq War Veteran V.A. users
.
Medical Care
 ,
50
(4)
,
342
346
.
Terrio
H.
Brenner
L. A.
Ivins
B. J.
Cho
J. M.
Helmick
K.
Schwab
K.
et al
(
2009
).
Traumatic brain injury screening: Preliminary findings in a US Army brigade combat team
.
Journal of Head Trauma Rehabilitation
 ,
24
(1)
,
14
23
.
Theeler
B.
Lucas
S.
Riechers
R. G.
II
Ruff
R. L.
(
2013
).
Post-traumatic headaches in civilians and military personnel: A comparative, clinical review
.
Headache
 ,
53
(6)
,
881
900
.
Tombaugh
T. N.
(
1996
).
Test of memory malingering
 .
Toronto, Canada
:
Multi-Health Systems
.
Tombaugh
T. N.
(
1997
).
The Test of Memory Malingering (TOMM): Normative data from cognitively intact and cognitively impaired individuals
.
Psychological Assessment
 ,
9
,
260
268
.
VA/DoD Clinical Practice Guideline: Management of Concussion/Mild Traumatic Brain Injury
. (
2009
). .
Vanderploeg
R. D.
Belanger
H. G.
Curtiss
G.
(
2009
).
Mild traumatic brain injury and posttraumatic stress disorder and their associations with health symptoms
.
Archives of Physical Medicine & Rehabilitation
 ,
90
(7)
,
1084
1093
.
Vanderploeg
R. D.
Cooper
D. B.
Belanger
H. G.
Donnell
A. J.
Kennedy
J. E.
Hopewell
C. A.
et al
(
2014
).
Screening for postdeployment conditions: Development and cross-validation of an embedded validity scale in the neurobehavioral symptom inventory
.
Journal of Head Trauma Rehabilitation
 ,
29
(1)
,
1
10
.
Van Dyke
S. A.
Millis
S. R.
Axelrod
B. N.
Hanks
R. A.
(
2013
).
Assessing effort: Differentiating performance and symptom validity
.
Clinical Neuropsychology
 ,
27
(8)
,
1234
1246
.
Velikonja
D.
Warriner
E.
Brum
C.
(
2010
).
Profiles of emotional and behavioral sequelae following acquired brain injury: Cluster analysis of the Personality Assessment Inventory
.
Journal of Clinical and Experimental Neuropsychology
 ,
32
(6)
,
610
621
.
Weathers
F. W.
Huska
J. A.
Keane
T. M.
(
1991
).
The PTSD Checklist Military version (PCL-M)
 .
Boston, MA
:
National Center for PTSD
.
Whiteside
D. M.
Galbreath
J.
Brown
M.
Turnbull
J
. (
2012
).
Differential response patterns on the Personality Assessment Inventory (PAI) in compensation-seeking and non-compensation-seeking mild traumatic brain injury patients
.
Journal of Clinical and Experimental Neuropsychology
 ,
34
(2)
,
172
182
.
Wilkins
K. C.
Lang
A. J.
Norman
S. B.
(
2011
).
Synthesis of the psychometric properties of the PTSD Checklist (PCL) Military, civilian, and specific versions
.
Depression and Anxiety
 ,
28
(7)
,
596
606
.